JP7308740B2 - Water absorption state evaluation device and water absorption state evaluation method - Google Patents

Description

TECHNICAL FIELD The present invention relates to a water absorption state evaluation device and a water absorption state evaluation method for evaluating the water absorption state of a group of target rice submerged in water.

For example, in sake brewing, the processing steps for raw rice, such as rice polishing, water absorption, and steaming, are important processes that influence the subsequent brewing process. In the process of water absorption in the production of daiginjo-shu, so-called limited water absorption, in which the soaking time is shortened to prevent the raw rice from absorbing too much water, is sometimes carried out. Since this limited water absorption requires adjustment of the water absorption in units of 1%, it is extremely important to grasp the water absorption state of the raw material rice in detail in real time in order to improve the quality of the alcoholic beverages to be produced.

It is known that the change in water absorption of raw rice varies greatly depending on the condition of raw rice (variety, rice polishing ratio, moisture content, etc.) and soaking conditions (time, temperature, concentration, etc.). As a method for evaluating the change in water absorption, it was common to measure the weight of raw rice before and after soaking with a weighing machine and calculate the amount of water absorption from the difference in weight before and after soaking. However, with such a method, it is not possible to accurately grasp changes in water absorption over time, water absorption distribution, and the like. Therefore, at the actual sake brewing site, the water absorption distribution is estimated from the change in the color and shape of the raw material rice by visual observation, and the change in water absorption is estimated by combining the above-mentioned calculation results of the water absorption amount. Therefore, there is a demand for the development of a method that can be evaluated in real time.

By the way, the applicant of the present application has proposed a water content determination method described in Patent Document 1 as a method for determining the water absorption amount and water absorption distribution of rice to be cooked by image observation. According to this water content state determination method, it is possible to optically monitor the rice to be cooked that is placed in a state of being immersed in water, and to measure the water content related information regarding the water content of the rice to be cooked over time. , it is possible to evaluate the water absorption amount and water absorption distribution of the rice to be cooked based on the moisture content related information. Specifically, the change in the size of the rice to be cooked due to water absorption of the rice to be cooked is measured over time, and the size of the rice to be cooked after water absorption is compared with the size of the rice to be cooked before water absorption. Since the rate of change (expansion) in thickness can be taken as the expansion rate, the water absorption amount and water absorption distribution of the rice to be cooked can be evaluated based on this expansion rate.

JP 2016-105074 A

By the way, the water absorption state of sake rice is an important factor that influences the taste and aroma of sake. At actual brewing sites, skilled workers evaluate the water absorption state of sake rice based on many years of experience. It is considered desirable to reproduce the evaluation based on the experience of the person.

The present invention has been made in view of the above circumstances, and its object is to provide a water absorption state evaluation apparatus and a water absorption state evaluation method that can reproduce the evaluation of the water absorption state of rice to be measured based on the experience of an expert. do.

The characteristic configuration of the water absorption state evaluation device according to the present invention for achieving the above object is as follows:
an image data acquisition unit that acquires image data obtained by photographing a group of target rice immersed in water;
an image processing unit capable of executing a predetermined process on the image data acquired by the image data acquiring unit;
A functional unit that determines an index related to the water absorption state of the rice clump to be measured, which calculates an area-weighted average brightness determined based on image data of the rice clump to be measured, and calculates the calculated area an index determination unit that determines the weighted average brightness as one of the indices;
a water absorption state evaluation unit that evaluates the water absorption state of the rice group to be measured based on the index determined by the index determination unit;
The area-weighted average brightness is the average brightness of the portion occupied by the rice mass to be measured in the image data of the rice mass to be measured,
The image processing unit is capable of executing a gradation conversion process for converting the image data acquired by the image data acquisition unit into image data having a predetermined gradation,
The area-weighted average brightness is an index determined based on the image data obtained by photographing the rice group to be measured with a predetermined gradation, and the area occupied by each rice to be measured is added to the average brightness of the portion occupied by each rice to be measured is obtained by dividing the sum of the products obtained by multiplying the number of pixels by the number of all the pixels in the portion occupied by the rice mass to be measured.
Further, the characteristic configuration of the water absorption state evaluation method according to the present invention for achieving the above object is as follows:
an image data acquisition step of acquiring image data obtained by photographing a group of target rice immersed in water;
an image processing step capable of executing a predetermined process on the image data obtained in the image data obtaining step;
A step of determining an index related to the water absorption state of the rice mass to be measured, wherein an area-weighted average luminance determined based on image data obtained by photographing the rice mass to be measured is calculated, and the calculated area-weighted an index determination step of determining the average luminance as one of the indices;
a water absorption state evaluation step of evaluating the water absorption state of the rice group to be measured based on the index determined in the index determination step;
The area-weighted average brightness is the average of the brightness of the portion occupied by the rice mass to be measured in the image data obtained by photographing the rice mass to be measured,
The image processing step is capable of executing gradation image processing for converting the image data acquired in the image data acquisition step into image data of a predetermined gradation,
The area-weighted average brightness is an index determined based on image data obtained by photographing the rice group to be measured with a predetermined gradation. is obtained by dividing the sum of the products multiplied by the number of pixels of , by the number of all pixels in the portion occupied by the rice mass to be measured.

According to the apparatus for evaluating water absorption state and the method for evaluating water absorption state according to the present invention, first, image data of a group of target rice immersed in water is acquired. Then, the acquired image data is subjected to predetermined processing in the image processing section.

Here, the inventors of the present application conducted intensive research on a method that can reproduce the evaluation of the water absorption state of sake rice based on the experience of experts. We found that the evaluation can be reproduced. That is, the inventor of the present application found that the average brightness of the portion occupied by the rice group to be measured weighted by the area of the portion occupied by the rice group to be measured (area-weighted average brightness) in the image of the rice group to be measured is obtained by a skilled person. We obtained new knowledge that it is an important index for reproducing the evaluation by

The water absorption state evaluation device and the water absorption state evaluation method according to the present invention have been made based on such knowledge. is determined as one of the indices.

Thereafter, the water absorption state evaluation unit evaluates the water absorption state of the rice group to be measured based on the index (area-weighted average luminance) determined by the index determination unit.

As described above, according to the water absorption state evaluation device and the water absorption state evaluation method according to the present invention, when evaluating the water absorption state of the rice group to be measured, by using the area-weighted average luminance as one of the indices, the skilled person can It is possible to reproduce the evaluation of the water absorption state of the rice group to be measured based on the experience of

In addition, according to the above characteristic configuration, the sum of the products of the average brightness of the portion occupied by each measurement target rice multiplied by the number of pixels in the portion occupied by each measurement target rice is the sum of all pixels in the portion occupied by the measurement target rice group. The area-weighted average brightness as an index is obtained by dividing by a number. By using this area-weighted average luminance as an index, it is possible to reproduce the evaluation of an expert. Note that the extracted image data of predetermined gradation is not particularly limited, and for example, extracted image data of 256 gradations and extracted image data of 16 gradations can be exemplified.

The characteristic configuration of the water absorption state evaluation device according to the present invention is as follows :
an image data acquisition unit that acquires image data obtained by photographing a group of target rice immersed in water;
an image processing unit capable of executing a predetermined process on the image data acquired by the image data acquiring unit;
A functional unit that determines an index related to the water absorption state of the rice clump to be measured, which calculates an area-weighted average brightness determined based on image data of the rice clump to be measured, and calculates the calculated area an index determination unit that determines the weighted average brightness as one of the indices;
a water absorption state evaluation unit that evaluates the water absorption state of the rice group to be measured based on the index determined by the index determination unit;
The area-weighted average brightness is the average brightness of the portion occupied by the rice mass to be measured in the image data of the rice mass to be measured,
The image processing unit is capable of executing a gradation conversion process for converting the image data acquired by the image data acquisition unit into image data having a predetermined gradation,
The area-weighted average brightness is an index determined based on image data of the rice group to be measured with a predetermined gradation, and the sum of the brightness of each pixel in the portion occupied by the rice to be measured is the measurement target. It is divided by the number of all pixels in the portion occupied by rice clusters.

According to the apparatus for evaluating the state of water absorption according to the present invention, first, image data obtained by photographing a group of target rice immersed in water is acquired. Then, the acquired image data is subjected to predetermined processing in the image processing section.
Here, the inventors of the present application conducted intensive research on a method that can reproduce the evaluation of the water absorption state of sake rice based on the experience of experts. We found that the evaluation can be reproduced. That is, the inventor of the present application found that the average brightness of the portion occupied by the rice group to be measured weighted by the area of the portion occupied by the rice group to be measured (area-weighted average brightness) in the image of the rice group to be measured is obtained by a skilled person. We obtained new knowledge that it is an important index for reproducing the evaluation by
The apparatus for evaluating the state of water absorption according to the present invention has been made based on such knowledge. Determine as one.
Thereafter, the water absorption state evaluation unit evaluates the water absorption state of the rice group to be measured based on the index (area-weighted average luminance) determined by the index determination unit.
As described above, according to the water absorption state evaluation apparatus according to the present invention, when evaluating the water absorption state of the rice group to be measured, by using the area-weighted average luminance as one of the indexes, the measurement based on the experience of the expert It is possible to reproduce the evaluation of the water absorption state of the target rice group.
Further, according to the above characteristic configuration, the sum of the brightness of each pixel in the portion occupied by the rice to be measured is divided by the number of all the pixels in the portion occupied by the rice to be measured to obtain the area-weighted average brightness as an index. . By using this area-weighted average luminance as an index, it is possible to reproduce the evaluation of an expert. Note that the extracted image data of predetermined gradation is not particularly limited, and for example, extracted image data of 256 gradations and extracted image data of 16 gradations can be exemplified.

The characteristic configuration of the water absorption state evaluation device according to the present invention is as follows:
an image data acquisition unit that acquires image data obtained by photographing a group of target rice immersed in water;
Based on the image data acquired by the image data acquisition unit, an extraction process for extracting an area assumed to be one grain of rice to be measured, and the image data of the rice group to be measured after the extraction process, an image processing unit capable of at least executing image generation processing for generating extracted image data including the region extracted by the extraction processing;
A functional unit that determines an index related to the water absorption state of the rice clump to be measured, which calculates an area-weighted average brightness determined based on image data of the rice clump to be measured, and performs the image processing. A high-luminance portion ratio determined based on the extracted image data generated by the image generation processing is calculated in the section, and the calculated area-weighted average luminance and the high-luminance portion ratio are determined as the index. an index determination unit to
a water absorption state evaluation unit that evaluates the water absorption state of the rice group to be measured based on the index determined by the index determination unit;
The area-weighted average brightness is the average brightness of the portion occupied by the rice mass to be measured in the image data of the rice mass to be measured,
The ratio of the high luminance portion is the ratio of the portion occupied by the high luminance portion to the portion occupied by the rice to be measured in the extracted image data.

According to the apparatus for evaluating the state of water absorption according to the present invention, first, image data obtained by photographing a group of target rice immersed in water is obtained. Then, the acquired image data is subjected to predetermined processing in the image processing section.
Here, the inventors of the present application conducted extensive research on a method that can reproduce the evaluation of the water absorption state of sake rice based on the experience of experts. It was also found that the evaluation by That is, the inventor of the present application has determined the high-luminance portion ratio, which is the ratio of the portion occupied by the high-luminance portion in the portion occupied by the rice to be measured in the image including the region assumed to be one grain of rice to be measured, and the area weighting A new finding was obtained that the average luminance is an important index for reproducing the evaluation by an expert.

The water absorption state evaluation apparatus according to the present invention is based on such knowledge, and in the index determination unit, in addition to the area-weighted average brightness determined based on the image data of the rice cluster to be measured, image processing A high-brightness portion ratio determined based on the extracted image data generated by image generation processing in the portion is determined as one of the indices.
After that, the water absorption state evaluation unit evaluates the water absorption state of the rice group to be measured based on the indices (ratio of high luminance portions and area-weighted average luminance) determined by the index determination unit.
As described above, according to the water absorption state evaluation apparatus according to the present invention, when evaluating the water absorption state of a rice group to be measured, by using the high luminance portion ratio and the area weighted average luminance as indexes, It is possible to reproduce the evaluation of the water absorption state of the rice group to be measured.

Further, a further characteristic configuration of the water absorption state evaluation device according to the present invention is that the image processing section can execute a gradation conversion process of converting the image data acquired by the image data acquisition section into image data of a predetermined gradation. and
The high brightness portion ratio is an index determined based on the extracted image data of a predetermined gradation, and is the number of pixels having a brightness higher than a predetermined threshold among the pixels in the portion occupied by the rice to be measured. , divided by the number of all pixels in the portion occupied by the rice to be measured.

According to the above characteristic configuration, among the pixels in the portion occupied by the rice to be measured, the index is obtained by dividing the number of pixels whose luminance is higher than a predetermined threshold by the number of all pixels in the portion occupied by the rice to be measured. high luminance ratio. By using this high-brightness portion ratio as one of the indices, it is possible to reproduce the evaluation of an expert. Note that the extracted image data of predetermined gradation is not particularly limited, and for example, extracted image data of 256 gradations and extracted image data of 16 gradations can be exemplified.

Further, a further characteristic configuration of the water absorption state evaluation apparatus according to the present invention is that the extracted image data of predetermined gradation is extracted image data of 256 gradations,
Said threshold lies in that it is set within the range of 90-150.

When the extracted image data of predetermined gradation is the extracted image data of 256 gradations, the predetermined threshold value is appropriately set according to the type and average shape of the rice to be measured. By setting a threshold within the range of ~150 and using the ratio of high brightness areas obtained by dividing the number of pixels with brightness higher than the threshold by the number of all pixels as an index, it is possible to reproduce the evaluation of an expert. The inventor of the present application has experimentally confirmed that this is possible.

Further, a further characteristic configuration of the water absorption state evaluation device according to the present invention is that, when the index determination unit determines the high-brightness portion ratio as the index, in the extracted image data, The high luminance portion ratio is determined based only on the extracted image data in which the number of pixels in the portion occupied by the rice to be measured is within a predetermined range.

The extraction processing performed by the image processing unit is processing capable of extracting a region that is assumed to be one grain of measurement target rice from image data obtained by photographing a measurement target rice group. However, depending on the state of the rice to be measured in the image of the rice group to be measured and the positional relationship between the rice to be measured, the unintended area can be assumed to be a single normal grain of rice to be measured. It may be extracted. For example, if two grains of rice to be measured overlap and it is not possible to extract only one grain of rice to be measured, it is assumed that the two grains of rice to be measured are one grain of rice to be measured. If the extracted image data including the extracted area is generated, the image data showing two grains of rice to be measured will be obtained as the extracted image data showing one grain of rice to be measured. . In addition, if there is rice subject to measurement that has become extremely smaller than usual due to damage, etc., the image data showing the rice subject to measurement that has become extremely small in size (abnormal rice subject to measurement) is one grain of measurement subject. It is obtained as extracted image data showing rice. Then, based on the extracted image data showing a single grain of rice that is not the normal rice to be measured, the ratio of the high-luminance portion is determined, and when this ratio of the high-luminance portion is used as an index, the water absorption due to the experience of the expert There is a possibility that the state evaluation cannot be reproduced with high accuracy.
However, according to the above characteristic configuration, the high-brightness portion ratio is determined based only on extracted image data in which the number of pixels in the portion occupied by the rice to be measured in the extracted image data is within a predetermined range. Therefore, the water absorption state of the rice group to be measured can be evaluated with high accuracy. The above-mentioned predetermined range may be any range as long as the number of pixels in the portion occupied by one normal grain of rice to be measured can be taken, and it varies according to the magnification etc. when imaging the cluster of rice to be measured. be.

Further, a further characteristic configuration of the water absorption state evaluation device according to the present invention is
The index determination unit calculates a crack ratio determined based on the extracted image data generated by the image generation process in the image processing unit, and calculates the area-weighted average luminance, the high-luminance portion ratio, and the Determine the crack rate as the index,
The cracking rate is the percentage of the rice to be measured that is determined to have cracks based on the extracted image data, among the rice group to be measured.

Here, the inventors of the present application conducted intensive research on a method that can reproduce the evaluation of the water absorption state of sake rice based on the experience of experts. It was also found that the evaluation by an expert could be reproduced. That is, the inventor of the present application determined that cracks were generated based on an image containing an area assumed to be one grain of rice to be measured, out of a group of rice to be measured. A new knowledge was obtained that the ratio, the area-weighted average luminance, and the high-luminance portion ratio are important indexes for reproducing the evaluation by a skilled person.

The water absorption state evaluation apparatus according to the present invention is based on such knowledge, and in the index determination unit, in addition to the area-weighted average brightness determined based on the image data of the rice cluster to be measured, image processing A crack ratio and a high-brightness portion ratio determined based on the extracted image data generated by the image generation processing are determined as one of the indices.
After that, the water absorption state evaluation unit evaluates the water absorption state of the rice group to be measured based on the index (crack rate, area-weighted average luminance, and high luminance portion ratio) determined by the index determination unit.
As described above, according to the water absorption state evaluation apparatus according to the present invention, when evaluating the water absorption state of the rice group to be measured, by using the crack rate, the area-weighted average luminance, and the high luminance portion ratio as indicators, the skilled person can It is possible to reproduce the evaluation of the water absorption state of the rice group to be measured based on the experience of
The term "crack" in the present application is a concept that includes both cracks along the long axis direction of the rice to be measured (longitudinal cracks) and cracks along the short axis direction of the rice to be measured (transverse cracks).

The characteristic configuration of the water absorption state evaluation device according to the present invention is as follows:
an image data acquisition unit that acquires image data obtained by photographing a group of target rice immersed in water;
Based on the image data acquired by the image data acquisition unit, an extraction process for extracting an area assumed to be one grain of rice to be measured, and the image data of the rice group to be measured after the extraction process, an image processing unit capable of at least executing image generation processing for generating extracted image data including the region extracted by the extraction processing;
A functional unit that determines an index related to the water absorption state of the rice clump to be measured, which calculates an area-weighted average brightness determined based on image data of the rice clump to be measured, and performs the image processing. an index determination unit that calculates a crack ratio determined based on the extracted image data generated by the image generation process in the unit, and determines the calculated area-weighted average luminance and the crack ratio as the index ; and,
a water absorption state evaluation unit that evaluates the water absorption state of the rice group to be measured based on the index determined by the index determination unit;
The area-weighted average brightness is the average brightness of the portion occupied by the rice mass to be measured in the image data of the rice mass to be measured,
The cracking rate is the percentage of the rice to be measured that is determined to have cracks based on the extracted image data, among the rice group to be measured.

According to the apparatus for evaluating the state of water absorption according to the present invention, first, image data obtained by photographing a group of target rice immersed in water is obtained. Then, the acquired image data is subjected to predetermined processing in the image processing section.
Here, the inventors of the present application conducted intensive research on a method that can reproduce the evaluation of the water absorption state of sake rice based on the experience of experts. was also found to be reproducible. That is, the inventor of the present application determined that cracks were generated based on an image containing an area assumed to be one grain of rice to be measured, out of a group of rice to be measured. A new knowledge was obtained that the ratio and the area-weighted average luminance are important indexes for reproducing the evaluation by a skilled person.

The water absorption state evaluation apparatus according to the present invention is based on such knowledge, and in the index determination unit, in addition to the area-weighted average brightness determined based on the image data of the rice cluster to be measured, image processing A crack rate determined based on the extracted image data generated by the image generation process is determined as one of the indices.
Thereafter, the water absorption state evaluation unit evaluates the water absorption state of the rice clump to be measured based on the indices (crack rate and area-weighted average brightness) determined by the index determination unit.
As described above, according to the water absorption state evaluation apparatus according to the present invention, when evaluating the water absorption state of the rice group to be measured, by using the crack rate and the area-weighted average luminance as indices, the measurement based on the experience of the expert It is possible to reproduce the evaluation of the water absorption state of the target rice group.
The term "crack" in the present application is a concept that includes both cracks along the long axis direction of the rice to be measured (longitudinal cracks) and cracks along the short axis direction of the rice to be measured (transverse cracks).

Further, a further characteristic configuration of the water absorption state evaluation apparatus according to the present invention is that the cracking rate is an index determined based on the extracted image data, and the accumulated training data of the rice group to be measured It is the ratio of the rice to be measured that is determined to have cracks based on the extracted image data using the execution result of machine learning.

According to the above characteristic configuration, the percentage of the rice to be measured that was determined to have cracks based on the extracted image data using the execution result of machine learning of the accumulated training data among the groups of rice to be measured is be an indicator. By using the result of machine learning execution of accumulated training data as an index, it is possible to identify cracked rice for measurement in a short time and consistently, and determine the cracking rate as an index. be able to. Moreover, by using the crack rate determined in this manner as an index, it is possible to reproduce the evaluation of the expert.

Further, a further characteristic configuration of the water absorption state evaluation apparatus according to the present invention is that the index determination unit determines the crack ratio as the index, out of the extracted image data, the measurement in the extracted image data The crack rate is determined based only on the extracted image data in which the number of pixels in the portion occupied by the target rice is within a predetermined range.

As described above, the extraction process performed by the image processing unit is a process that can extract an area that is assumed to be a grain of rice to be measured from the image data of the rice group to be measured. In some cases, an area where no measurement is performed is extracted as an area assumed to be a normal grain of rice to be measured. If the crack rate is determined based on the extracted image data showing a single grain of rice that is not the target rice to be measured, and if this crack rate is used as an index, the water absorption state of the rice group to be measured cannot be evaluated with high accuracy. There is fear.
However, according to the above-described characteristic configuration, the crack rate is determined based only on extracted image data in which the number of pixels in the portion occupied by the rice to be measured in the extracted image data is within a predetermined range. The water absorption state of the rice group to be measured can be evaluated with high accuracy. The above-mentioned predetermined range may be any range as long as the number of pixels in the portion occupied by one normal grain of rice to be measured can be taken, and it varies according to the magnification etc. when imaging the cluster of rice to be measured. be.

Further, a further characteristic configuration of the water absorption state evaluation device according to the present invention is
The image processing unit is capable of executing a gradation conversion process for converting the image data acquired by the image data acquisition unit into image data having a predetermined gradation,
The area-weighted average brightness is an index determined based on the image data obtained by photographing the rice group to be measured with a predetermined gradation, and the area occupied by each rice to be measured is added to the average brightness of the portion occupied by each rice to be measured is obtained by dividing the sum of the products obtained by multiplying the number of pixels by the number of all the pixels in the portion occupied by the rice mass to be measured.

According to the above characteristic configuration, the sum of the average brightness of the portion occupied by each rice to be measured multiplied by the number of pixels in the portion occupied by each rice to be measured is the number of all pixels in the portion occupied by the rice group to be measured. The result obtained by dividing is the area-weighted average luminance as an index. By using this area-weighted average luminance as an index, it is possible to reproduce the evaluation of an expert. Note that the extracted image data of predetermined gradation is not particularly limited, and for example, extracted image data of 256 gradations and extracted image data of 16 gradations can be exemplified.

Further, a further characteristic configuration of the water absorption state evaluation device according to the present invention is
The image processing unit is capable of executing a gradation conversion process for converting the image data acquired by the image data acquisition unit into image data having a predetermined gradation,
The area-weighted average brightness is an index determined based on image data of the rice group to be measured with a predetermined gradation, and the sum of the brightness of each pixel in the portion occupied by the rice to be measured is the measurement target. It is divided by the number of all pixels in the portion occupied by rice clusters.

According to the above characteristic configuration, the area-weighted average brightness as an index is obtained by dividing the sum of the brightness of each pixel in the portion occupied by the rice to be measured by the number of all pixels in the portion occupied by the rice to be measured. By using this area-weighted average luminance as an index, it is possible to reproduce the evaluation of an expert. Note that the extracted image data of predetermined gradation is not particularly limited, and for example, extracted image data of 256 gradations and extracted image data of 16 gradations can be exemplified.

Further, a further characteristic configuration of the water absorption state evaluation device according to the present invention is that the water absorption state evaluation unit evaluates the measurement object based on an evaluation model previously constructed by regression analysis using the index and evaluation results by a skilled person. The point is to evaluate the water absorption state of the rice group.
According to the above characteristic configuration, it is possible to reproduce the evaluation of the water absorption state of the rice group to be measured based on the experience of an expert using an evaluation model constructed by a relatively simple method.

The characteristic configuration of the water absorption state evaluation device according to the present invention is as follows:
an image data acquisition unit that acquires image data obtained by photographing a group of target rice immersed in water;
an image processing unit capable of executing a predetermined process on the image data acquired by the image data acquiring unit;
A functional unit that determines an index related to the water absorption state of the rice clump to be measured, which calculates an area-weighted average brightness determined based on image data of the rice clump to be measured, and calculates the calculated area an index determination unit that determines the weighted average brightness as one of the indices;
a water absorption state evaluation unit that evaluates the water absorption state of the rice group to be measured based on the index determined by the index determination unit;
The area-weighted average brightness is the average brightness of the portion occupied by the rice mass to be measured in the image data of the rice mass to be measured,
The water absorption state evaluation unit evaluates the water absorption state of the rice group to be measured based on an evaluation model previously constructed by regression analysis using the index and evaluation results by experts.
According to the apparatus for evaluating the state of water absorption according to the present invention, first, image data obtained by photographing a group of target rice immersed in water is obtained. Then, the acquired image data is subjected to predetermined processing in the image processing section.
Here, the inventors of the present application conducted intensive research on a method that can reproduce the evaluation of the water absorption state of sake rice based on the experience of experts. We found that the evaluation can be reproduced. That is, the inventor of the present application found that the average brightness of the portion occupied by the rice group to be measured weighted by the area of the portion occupied by the rice group to be measured (area-weighted average brightness) in the image of the rice group to be measured is obtained by a skilled person. We obtained new knowledge that it is an important index for reproducing the evaluation by
The apparatus for evaluating the state of water absorption according to the present invention has been made based on such knowledge. Determine as one.
Thereafter, the water absorption state evaluation unit evaluates the water absorption state of the rice group to be measured based on the index (area-weighted average luminance) determined by the index determination unit.
As described above, according to the water absorption state evaluation apparatus according to the present invention, when evaluating the water absorption state of the rice group to be measured, by using the area-weighted average luminance as one of the indexes, the measurement based on the experience of the expert It is possible to reproduce the evaluation of the water absorption state of the target rice group.
Moreover, according to the above-described characteristic configuration, it is possible to reproduce the evaluation of the water absorption state of the rice group to be measured based on the experience of an expert using an evaluation model constructed by a relatively simple method.

It is a figure which shows the structure of the water absorption state evaluation apparatus which concerns on this embodiment. It is a flowchart explaining the process of the water absorption state evaluation method. FIG. 4 is a diagram illustrating an example of image processing; FIG. 10 is a diagram for explaining a method of selecting an area to be excluded from areas assumed to be one grain of sake rice; FIG. 10 is a diagram for explaining a method of determining a luminance threshold for determining a high luminance portion; It is a figure for demonstrating the method to determine a high-luminance part ratio. It is a figure for demonstrating the method to determine a high-luminance part ratio. 7 is a graph showing the relationship between the ratio of high-brightness areas and the water absorption rate judged by an expert. It is a graph which shows the relationship between a crack rate and the water absorption rate which the expert judged. 7 is a graph showing the relationship between area-weighted average luminance and water absorption determined by an expert.

Hereinafter, a water absorption state evaluation device 1 according to an embodiment of the present invention will be described with reference to the drawings. In this embodiment, the rice to be measured is sake rice, and the water absorption state of the sake rice group (measurement rice group) is evaluated based on the high luminance portion ratio, the crack rate, and the area-weighted average luminance. explain.

As shown in FIG. 1, the water absorption state evaluation device 1 according to this embodiment includes a photographing device 2, an image data acquisition unit 10, an image processing unit 11, an index determination unit 12, and a water absorption state evaluation unit 13. It also has a storage device 3 for storing information to be handled. The water absorption state evaluation device 1 is realized using one or a plurality of computer devices having an information arithmetic processing function, an information input/output function, an information storage function, and the like. In this case, a program for realizing the function of the image data acquisition unit 10, the function of the image processing unit 11, the function of the index determination unit 12, and the function of the water absorption state evaluation unit 13 in the computer is installed in the computer. You should keep it.

The storage device 3 of the present embodiment includes an image data storage unit 3a for storing image data captured by the imaging device 2 and image data processed by the image processing unit 11, and a water absorption state evaluation unit 13 for evaluating the water absorption state of the sake rice group. and a model storage unit 3b for storing an evaluation model used when evaluating.

In this embodiment, the photographing device 2 is provided in the water absorption state evaluation device 1, but a photographing device may be provided separately from the water absorption state evaluation device 1. FIG.

FIG. 2 is a flow chart explaining each step included in the water absorption state evaluation method according to the present embodiment. In the water absorption state evaluation method according to the present embodiment, first, a step (photographing step) of photographing a group of sake rice placed in a state of being immersed in water (photographing step) is performed (step #10). Next, in step #11, the image data acquisition unit 10 acquires image data of the group of brewer's rice that has been placed in a state of being immersed in water, photographed by the photographing device 2 (image data acquisition step). The image data acquired by the image data acquisition section 10 may be stored in the image data storage section 3 a of the storage device 3 . For example, image data A shown in FIG. 3 is an example of image data of a cluster of sake rice acquired by the image data acquisition unit 10 .

Next, in step #12 of FIG. 2, the image processing unit 11 converts the image data of the group of sake rice acquired by the image data acquiring unit 10 into image data with a predetermined gradation (256 gradations in this example). A conversion process (gradation conversion process) is performed, and in step #13, noise is removed from the image data of 256 gradations, and noise is removed from the image data of the sake rice group (noise removal process). ). For example, image data B shown in FIG. 3 is an example of image data of a group of sake rice with 256 gradations from which noise has been removed by the image processing unit 11 . The noise-removed image data of the sake rice group is stored in the image data storage section 3 a of the storage device 3 . In this embodiment, the image processing step is a step of executing the tone conversion processing, the noise removal processing, the extraction processing described later, and the image generation processing.

Next, in step #14 of FIG. 2, the image processing unit 11 performs an extraction process for extracting a region assumed to be a grain of sake rice using a so-called Watershed algorithm (extraction process). Specifically, in the present embodiment, after performing binarization processing on the noise-removed 256-gradation image data of the group of sake rice, after extracting the background region and the foreground region, Extract unknown regions that are not in the foreground. After that, label marker processing is performed on the reliable foreground, markers are applied to the image data of the sake rice group in 256 levels using the Watershed algorithm, and contours on the markers are detected. Each grain of sake rice is extracted. For example, image data C shown in FIG. 3 is image data obtained by extracting a background area, image data D is image data obtained by extracting a foreground area, image data E is image data obtained by extracting an unknown area, and image data F is image data obtained by extracting an unknown area. 4 is an example of image data of a group of brewer's rice after extraction processing.

Here, when a region assumed to be one grain of sake rice is extracted by the extraction process, for example, regions containing two or more grains of sake rice, such as regions A and B in FIG. It may be extracted as a region assumed to be a single grain of sake rice. When a region assumed to be a grain of sake rice is normally extracted, it is considered that there is a certain degree of correlation between the area (the number of pixels) and the perimeter of the region. For example, when two or more grains of sake rice are included, as in regions A and B in FIG. A certain degree of correlation cannot be seen between them.

Therefore, in the present embodiment, an area that has a predetermined relationship between the area (number of pixels) of the extracted area and the perimeter of the extracted area is assumed to be a grain of sake rice. , regions (regions A and B) that do not have a predetermined relationship are excluded from regions that are assumed to be a single grain of sake rice. Specifically, the area where the relationship between the area (the number of pixels) and the perimeter is in the shaded part in FIG. 4 is assumed to be a grain of sake rice, and the shaded part are excluded from the area assumed to be a grain of sake rice (area A and area B).

Next, at step #15 in FIG. 2, the image processing unit 11 executes a process of generating extracted image data including the extracted region based on the image data of the group of brewer's rice after the extraction process (image generation process ). Specifically, in the image data of the group of brewer's rice after the extraction process, extracting image data including the extracted region is generated by generating image data that includes only the region that is assumed to be one grain of brewer's rice. do. If there are a plurality of extracted regions in the image data of the group of sake rice, extracted image data including each region is generated for each extracted region. In addition, in the present embodiment, for regions excluded from the region assumed to be one grain of sake rice, extracted image data including these regions is not generated, but is not limited to this. do not have.

Next, in step #16 in FIG. 2, the index determination unit 12, which is a functional unit that determines an index related to the water absorption state of the sake rice cluster (measurement target rice group), determines an index related to the water absorption state of the sake rice cluster ( High brightness portion ratio, crack rate and area-weighted average brightness) are determined (index determination step). In the water absorption state evaluation device 1 according to the present embodiment, the index determination unit 12 is composed of a high brightness portion ratio determination unit 12a, a crack rate determination unit 12b, and an area-weighted average brightness determination unit 12c.

Here, in the image processing unit 11, it is possible to extract an area assumed to be a single grain of sake rice from the image data obtained by photographing the group of sake rice, and generate extracted image data including the extracted area. is. However, depending on the state of the sake rice and the positional relationship between the two grains of sake rice, two grains of sake rice may overlap, making it impossible to properly extract just one grain of sake rice, resulting in two grains of sake rice being photographed. There are cases where extracted image data is generated, and extracted image data showing the sake rice whose size has become extremely small due to damage or the like is extracted, and the like. Then, based on the extracted image data in which it cannot be said that a single normal grain of sake rice is shown, the ratio of high-brightness portions and the ratio of cracks are determined, and when these are used as indicators, it depends on the experience of a skilled person. There is a possibility that the evaluation of the water absorption state cannot be reproduced with high accuracy.

Therefore, when determining the high-brightness portion ratio and the crack ratio, the index determination unit 12 sets the number of pixels in the portion occupied by sake rice in the extraction image data generated by the image processing unit 11 to a predetermined number. Based only on the extracted image data within the range, the high-brightness portion ratio and the crack ratio are determined. Specifically, in the present embodiment, the high-brightness portion ratio determination unit 12a and the cracking ratio determination unit 12b determine the high-brightness portion ratio using extracted image data in which the number of pixels in the portion occupied by sake rice is 7000 to 9999. and determine the crack rate. The predetermined range is not limited to 7000 to 9999, but may be 6000 to 9999, or any range as long as the number of pixels in the portion occupied by one normal grain of rice to be measured can be taken. Therefore, it can be appropriately set according to the variation according to the magnification or the like when imaging the rice clumps to be measured.

Based on the extracted image data generated by the image processing unit 11, the high-brightness portion ratio determination unit 12a determines the ratio of the high-brightness portion to the portion occupied by sake rice in the extracted image data (high-brightness portion ratio ). Specifically, in this embodiment, the high-brightness portion ratio determination unit 12a determines the number of pixels whose brightness is higher than a predetermined threshold among the pixels in the portion occupied by sake rice. , is determined as the high luminance portion ratio. The predetermined threshold value can be appropriately set according to the type and average shape of rice to be measured, and is preferably set within the range of 90 to 150, for example.

Here, the predetermined threshold for the brightness will be described. When evaluating the water absorption state of a group of sake rice (a group of rice to be measured) using the ratio of high-brightness portions as an index, the inventors of the present application have determined how high-brightness portions should be regarded as high-brightness portions based on the experience of experts. Intensive research was repeated to see if the evaluation of the water absorption state of the sake rice group could be reproduced. Specifically, for 40 pieces of soaked sake rice, when the boundary of brightness judged to be a high brightness portion was examined, a histogram as shown in FIG. 5 was obtained, with a mode value of 136 and an average value of 138.1. Met. Therefore, when the extracted image data is 256-gradation image data, the inventor of the present application sets the threshold within the range of 134 to 142 to determine the high-luminance portion ratio. It was judged to be particularly significant as an index when reproducing the evaluation of the water absorption state of the sake rice group based on experience. The applicant of the present application has found that when the rice to be measured is cooked rice, the cooked rice tends to be darker overall in the extracted image data of 256 gradations due to the thickness compared to the sake rice (that is, Since the brightness of each pixel tends to decrease), by setting the threshold to about 100 and determining the high brightness portion ratio, it is an index for reproducing the evaluation of the water absorption state of the rice group based on the experience of experts. It is confirmed that it is significant as

FIGS. 6 and 7 show an example in which the number of pixels having a luminance of 137 or more divided by the total number of pixels in the luminance histogram of the extracted image data is used as the high luminance portion ratio. In the present embodiment, when a plurality of pieces of extracted image data are generated, the high-brightness portion ratio determination unit 12a calculates and averages the high-brightness portion ratios for each of them to obtain the final high-brightness portion ratio. When a plurality of pieces of extracted image data are generated, it is possible to appropriately set how many pieces of the extracted image data for which the high-luminance portion ratio is to be calculated.

Based on the extracted image data generated by the image processing unit 11, the cracking rate determination unit 12b determines the proportion of the sake rice with cracks (breaking rate) in the group of sake rice. Specifically, in the present embodiment, the cracking rate determination unit 12b determines that cracks have occurred in the sake rice group based on the extracted image data using the machine learning execution results of the accumulated training data. The percent of the brewed sake rice is determined as the split rate. As a method of determining the occurrence of cracks using the results of machine learning of accumulated training data, for example, an algorithm such as CNN (Convolutional Neural Network) is used to determine the shape of sake rice (i.e., whether there is a crack ) can be determined, and the percentage of sake rice with cracks in the group of sake rice can be determined as the crack rate. Table 1 shows the accuracy rate of crack recognition when a crack determination test was conducted using machine learning execution results of accumulated training data for a plurality of samples. As can be seen from Table 1, the recognition correct rate shows a high numerical value of 92.9% on average, which is a level sufficient for practical use. The "number of grains of rice" in Table 1 is the number of pieces of sake rice judged to have cracks by human visual inspection.

Based on the noise-removed image data of the sake rice group generated in the image processing unit 11, the area-weighted average luminance determining unit 12c calculates the average luminance (area weighted average luminance). Specifically, in the present embodiment, the area-weighted average luminance determining unit 12c multiplies the average luminance of the portion occupied by each sake rice by the number of pixels in the portion occupied by each sake rice, and calculates the sum of the products. Determine the area-weighted average luminance divided by the number of all pixels in the part. In other words, when three sake rice exist in the noise-removed image data of a group of sake rice, the average brightness of the portion occupied by each sake rice is calculated, and the calculated average brightness is added to the portion occupied by each sake rice. are multiplied by the number of pixels in , and the resulting values for each sake rice are added together. Then, the added value is divided by the number of pixels in the portion occupied by the three sake rice, and the obtained value is determined as the area-weighted average luminance.

Incidentally, the inventor of the present application determined each of the above indices for a plurality of samples (rice group for sake brewing), judged the water absorption rate by an expert, and compared each of these indices with the water absorption rate as the water absorption state judged by the expert. A test was conducted to confirm the degree of correlation between them. 8 to 10 are graphs showing the results, FIG. 8 shows the relationship between the ratio of the high-luminance portion and the water absorption rate judged by the expert, FIG. 9 shows the relationship between the crack rate and the water absorption rate judged by the expert, FIG. 10 is a graph showing the relationship between the area-weighted average luminance and the water absorption rate judged by an expert.

As can be seen from FIGS. 8 to 10, each index shows a high correlation with the water absorption rate judged by an expert, and in particular, the high luminance portion ratio and area-weighted average luminance show a high correlation. From this, it is considered that each of the above indices is a useful index in reproducing the evaluation of the water absorption state based on the experience of experts, and the high luminance portion ratio and the area-weighted average luminance are particularly useful indexes. .

Next, in step #17 of FIG. 2, the water absorption state evaluation unit 13 evaluates the water absorption state of the sake rice group based on the indices (ratio of high-brightness areas, crack rate, and area-weighted average brightness) determined by the index determination unit 12. is evaluated (water absorption state evaluation step). Specifically, in the present embodiment, the water absorption state evaluation unit 13 applies the index determined by the index determination unit 12 to a pre-built evaluation model to calculate a predicted water absorption rate that indicates the water absorption state of the sake rice group.

Here, the evaluation model in the present embodiment uses the high luminance portion ratio, the crack ratio, and the area-weighted average luminance for multiple samples (sake rice group), and the evaluation results of an expert (master brewer) for these multiple samples. 1 is a regression equation (Equation 1) that is constructed in advance by regression analysis. In the following Equation 1, "index 1" is the high luminance portion ratio, "index 2" is the area-weighted average luminance, "index 3" is the crack rate, and a, b, c, and d are partial regression coefficients. be.
(Number 1)
Predicted water absorption (%) = (a x index 1 + b x index 2 + c x index 3 + d) x 100

In this embodiment, the ratio of high-brightness portions, the crack rate, and the area-weighted average brightness of nine samples are determined, and an expert judges the water absorption rate of these nine samples. A regression equation is obtained in advance by multiple regression analysis using the brightness portion ratio, crack rate, area-weighted average brightness, and water absorption, and this regression equation is stored in the model storage unit 3b as an evaluation model. Then, the water absorption state evaluation unit 13 determines the sake rice group based on the regression equation stored in the model storage unit 3b and the index (high luminance portion ratio, crack ratio, and area-weighted average luminance) determined by the index determination unit 12. Evaluate the water absorption state (predicted water absorption rate). In this regression equation, a is 11.117, b is -2.776, c is -54.839, and d is 352.135.

Table 2 is a table summarizing the results of verifying to what extent the predicted water absorption calculated based on the above regression equation is reproduced with the water absorption determined by an expert for nine groups of sake rice. As shown in Table 2, the average deviation between the predicted water absorption rate and the water absorption rate judged by the expert was 7.74, which is sufficiently small, and the evaluation by the expert was sufficiently reproduced.

As described above, according to the water absorption state evaluation device 1 and the water absorption state evaluation method, a group of sake rice (measurement target rice group) composed of sake rice (measurement target rice) immersed in water is photographed, and this photographed Based on the image data and the extracted image data including the area assumed to be one grain of sake rice in this image data, indices related to the water absorption state of sake rice (high brightness part ratio, crack rate and area weighting average luminance) can be determined. The ratio of high-brightness areas, crack rate, and area-weighted average brightness are important indicators for reproducing the evaluation of the water absorption state by experts, and these indicators are used when evaluating the water absorption state of sake rice. Therefore, it is possible to reproduce the evaluation of the water absorption state of sake rice based on the experience of experts.

[Another embodiment]
[1] In the above embodiment, the water absorption state evaluation unit 13 evaluates the water absorption state of the sake rice based on the three indexes of high luminance portion ratio, crack rate, and area-weighted average luminance. It is not limited to this. For example, the water absorption state evaluation unit 13 may use only the area-weighted average luminance as an index, or may use the area-weighted average luminance and the high luminance portion ratio or crack ratio as indexes. When the water absorption state evaluation unit 13 evaluates the water absorption state based on one or two indices, the index determination unit 12 may determine the above three indices. Only the indicators that are needed may be determined.

[2] In the above embodiment, the sum of the average brightness of the portion occupied by each sake rice multiplied by the number of pixels in the portion occupied by each sake rice is divided by the number of all pixels in the portion occupied by the sake rice group. Although it is determined as the area-weighted average luminance, it is not limited to this, and the area-weighted average luminance may be determined by other methods. For example, the area-weighted average brightness may be determined by dividing the sum of the brightness of each pixel in the portion occupied by sake rice by the number of all pixels in the portion occupied by the group of sake rice. For example, in the noise-removed image data of a group of sake rice, the number of pixels in the portion occupied by sake rice is 20, of which 4 pixels have a brightness of 40, 7 pixels have a brightness of 130, and 9 pixels have a brightness of 130. When the luminance is 150, the area-weighted average luminance is 121, which is the calculation result of the following formula (Equation 2).
(Number 2)
(4×40+7×130+9×150)/20=121

[3] In the above embodiment, the ratio of the high-luminance portion is obtained by dividing the number of pixels whose brightness is higher than a predetermined threshold among the pixels in the portion occupied by sake rice by the number of all pixels in the portion occupied by sake rice. , and out of the sake rice group, the percentage of sake rice judged to have cracks based on the extracted image data using the machine learning execution results of the accumulated training data is determined as the crack rate. However, it is not limited to this, and other methods may be used to determine the ratio of high-brightness portions and the ratio of cracks.

[4] In the above-described embodiment, the index determination unit 12 determines the portion of the extraction image data generated by the image processing unit 11 that is occupied by the sake rice. The high-brightness portion ratio and crack ratio are determined based only on the extracted image data in which the number of pixels of is within a predetermined range. , the ratio of high-brightness portions and the ratio of cracks may be determined.

[5] In the above embodiment, the water absorption state evaluation unit 13 uses a regression equation obtained in advance by multiple regression analysis as an evaluation model to evaluate the water absorption state of the sake rice group. However, the water absorption state of the group of sake rice may be evaluated by another method based on the index determined by the index determination unit 12 .

[6] In the above embodiment, the rice to be measured is sake rice, and the water absorption state of the sake rice group (measurement rice group) is evaluated based on the high luminance portion ratio, crack rate, and area-weighted average luminance. Although described as an example, it is not limited to this, and the rice to be measured may be cooked rice, and the water absorption state of the cooked rice group may be evaluated based on the above three indices.

[7] The configurations disclosed in the above embodiments (including other embodiments) can be applied in combination with the configurations disclosed in other embodiments unless there is a contradiction. The embodiments disclosed in the document are exemplifications, and the embodiments of the present invention are not limited thereto, and can be modified as appropriate without departing from the object of the present invention.

INDUSTRIAL APPLICABILITY The present invention can be applied to apparatuses and methods for evaluating the water absorption state of rice to be measured.

1 water absorption state evaluation device 10 image data acquisition unit 11 image processing unit 12 index determination unit 13 water absorption state evaluation unit

Claims (15)

an image data acquisition unit that acquires image data obtained by photographing a group of target rice immersed in water;
an image processing unit capable of executing a predetermined process on the image data acquired by the image data acquiring unit;
A functional unit that determines an index related to the water absorption state of the rice clump to be measured, which calculates an area-weighted average brightness determined based on image data of the rice clump to be measured, and calculates the calculated area an index determination unit that determines the weighted average brightness as one of the indices;
a water absorption state evaluation unit that evaluates the water absorption state of the rice group to be measured based on the index determined by the index determination unit;
The area-weighted average brightness is the average brightness of the portion occupied by the rice mass to be measured in the image data of the rice mass to be measured,
The image processing unit is capable of executing a gradation conversion process for converting the image data acquired by the image data acquisition unit into image data having a predetermined gradation,
The area-weighted average brightness is an index determined based on the image data obtained by photographing the rice group to be measured with a predetermined gradation, and the area occupied by each rice to be measured is added to the average brightness of the portion occupied by each rice to be measured The water absorption state evaluation device, which is obtained by dividing the sum of the products obtained by multiplying the number of pixels by the number of all pixels in the portion occupied by the rice group to be measured. an image data acquisition unit that acquires image data obtained by photographing a group of target rice immersed in water;
an image processing unit capable of executing a predetermined process on the image data acquired by the image data acquiring unit;
A functional unit that determines an index related to the water absorption state of the rice clump to be measured, which calculates an area-weighted average brightness determined based on image data of the rice clump to be measured, and calculates the calculated area an index determination unit that determines the weighted average brightness as one of the indices;
a water absorption state evaluation unit that evaluates the water absorption state of the rice group to be measured based on the index determined by the index determination unit;
The area-weighted average brightness is the average brightness of the portion occupied by the rice mass to be measured in the image data of the rice mass to be measured,
The image processing unit is capable of executing a gradation conversion process for converting the image data acquired by the image data acquisition unit into image data having a predetermined gradation,
The area-weighted average brightness is an index determined based on image data of the rice group to be measured with a predetermined gradation, and the sum of the brightness of each pixel in the portion occupied by the rice to be measured is the measurement target. A water absorption state evaluation device that is obtained by dividing by the number of all pixels in the portion occupied by rice clumps. an image data acquisition unit that acquires image data obtained by photographing a group of target rice immersed in water;
Based on the image data acquired by the image data acquisition unit, an extraction process for extracting an area assumed to be one grain of rice to be measured, and the image data of the rice group to be measured after the extraction process, an image processing unit capable of at least executing image generation processing for generating extracted image data including the region extracted by the extraction processing;
A functional unit that determines an index related to the water absorption state of the rice clump to be measured, which calculates an area-weighted average brightness determined based on image data of the rice clump to be measured, and performs the image processing. A high-luminance portion ratio determined based on the extracted image data generated by the image generation processing is calculated in the section, and the calculated area-weighted average luminance and the high-luminance portion ratio are determined as the index. an index determination unit to
a water absorption state evaluation unit that evaluates the water absorption state of the rice group to be measured based on the index determined by the index determination unit;
The area-weighted average brightness is the average brightness of the portion occupied by the rice mass to be measured in the image data of the rice mass to be measured,
The water absorption state evaluation apparatus, wherein the ratio of the high brightness portion is the ratio of the portion occupied by the high brightness portion to the portion occupied by the rice to be measured in the extracted image data. The image processing unit is capable of executing a gradation conversion process for converting the image data acquired by the image data acquisition unit into image data having a predetermined gradation,
The high brightness portion ratio is an index determined based on the extracted image data of a predetermined gradation, and is the number of pixels having a brightness higher than a predetermined threshold among the pixels in the portion occupied by the rice to be measured. , by the number of all pixels in the portion occupied by the rice to be measured. the predetermined gradation extracted image data is 256 gradation extracted image data,
5. The water absorption state evaluation device according to claim 4 , wherein the threshold is set within a range of 90-150. When determining the ratio of the high-luminance portion as the index, the index determination unit determines that the number of pixels in the portion of the extracted image data occupied by the rice to be measured is within a predetermined range. 6. The water absorption state evaluation device according to any one of claims 3 to 5 , wherein the ratio of the high luminance portion is determined based only on the extracted image data. The index determination unit calculates a crack ratio determined based on the extracted image data generated by the image generation process in the image processing unit, and calculates the area-weighted average luminance, the high-luminance portion ratio, and the Determine the crack rate as the index,
7. The cracking rate according to any one of claims 3 to 6, wherein the percentage of rice to be measured is the ratio of the rice to be measured that is determined to have cracks based on the extracted image data, among the group of rice to be measured. Water absorption state evaluation device. an image data acquisition unit that acquires image data obtained by photographing a group of target rice immersed in water;
Based on the image data acquired by the image data acquisition unit, an extraction process for extracting an area assumed to be one grain of rice to be measured, and the image data of the rice group to be measured after the extraction process, an image processing unit capable of at least executing image generation processing for generating extracted image data including the region extracted by the extraction processing;
A functional unit that determines an index related to the water absorption state of the rice clump to be measured, which calculates an area-weighted average brightness determined based on image data of the rice clump to be measured, and performs the image processing. an index determination unit that calculates a crack ratio determined based on the extracted image data generated by the image generation process in the unit, and determines the calculated area-weighted average luminance and the crack ratio as the index ; and,
a water absorption state evaluation unit that evaluates the water absorption state of the rice group to be measured based on the index determined by the index determination unit;
The area-weighted average brightness is the average brightness of the portion occupied by the rice mass to be measured in the image data of the rice mass to be measured,
The water absorption state evaluation device, wherein the cracking rate is the ratio of the rice to be measured that is determined to have cracks based on the extracted image data, among the rice groups to be measured. The cracking rate is an index determined based on the extracted image data, and the cracking rate is based on the extracted image data using the execution result of machine learning of the accumulated training data of the rice group to be measured. 9. The water absorption state evaluation device according to claim 7 or 8, wherein it is the ratio of the rice to be measured that is determined to have occurred. When determining the crack rate as the index, the index determination unit determines the extracted image data in which the number of pixels in the portion occupied by the rice to be measured in the extracted image data is within a predetermined range. The water absorption state evaluation device according to any one of claims 7 to 9 , wherein the crack rate is determined based only on image data. The image processing unit is capable of executing a gradation conversion process for converting the image data acquired by the image data acquisition unit into image data having a predetermined gradation,
The area-weighted average brightness is an index determined based on the image data obtained by photographing the rice group to be measured with a predetermined gradation, and the area occupied by each rice to be measured is added to the average brightness of the portion occupied by each rice to be measured The water absorption state evaluation device according to any one of claims 3 to 10, wherein the total number of pixels multiplied by the number of pixels is divided by the number of all pixels in the portion occupied by the rice mass to be measured. The image processing unit is capable of executing a gradation conversion process for converting the image data acquired by the image data acquisition unit into image data having a predetermined gradation,
The area-weighted average brightness is an index determined based on image data of the rice group to be measured with a predetermined gradation, and the sum of the brightness of each pixel in the portion occupied by the rice to be measured is the measurement target. The water absorption state evaluation device according to any one of claims 3 to 10, which is obtained by dividing by the number of all pixels in the portion occupied by rice clumps. 13. Any one of claims 1 to 12 , wherein the water absorption state evaluation unit evaluates the water absorption state of the rice group to be measured based on an evaluation model previously constructed by regression analysis using the index and evaluation results by an expert. 3. The water absorption state evaluation device according to the item. an image data acquisition unit that acquires image data obtained by photographing a group of target rice immersed in water;
an image processing unit capable of executing a predetermined process on the image data acquired by the image data acquiring unit;
A functional unit that determines an index related to the water absorption state of the rice clump to be measured, which calculates an area-weighted average brightness determined based on image data of the rice clump to be measured, and calculates the calculated area an index determination unit that determines the weighted average brightness as one of the indices;
a water absorption state evaluation unit that evaluates the water absorption state of the rice group to be measured based on the index determined by the index determination unit;
The area-weighted average brightness is the average brightness of the portion occupied by the rice mass to be measured in the image data of the rice mass to be measured,
The water absorption state evaluation unit is a water absorption state evaluation device that evaluates the water absorption state of the rice group to be measured based on an evaluation model previously constructed by regression analysis using the index and evaluation results by a skilled person. an image data acquisition step of acquiring image data obtained by photographing a group of target rice immersed in water;
an image processing step capable of executing a predetermined process on the image data obtained in the image data obtaining step;
A step of determining an index related to the water absorption state of the rice mass to be measured, wherein an area-weighted average luminance determined based on image data obtained by photographing the rice mass to be measured is calculated, and the calculated area-weighted an index determination step of determining the average luminance as one of the indices;
a water absorption state evaluation step of evaluating the water absorption state of the rice group to be measured based on the index determined in the index determination step;
The area-weighted average brightness is the average brightness of the portion occupied by the rice mass to be measured in the image data of the rice mass to be measured,
The image processing step is capable of executing gradation image processing for converting the image data acquired in the image data acquisition step into image data of a predetermined gradation,
The area-weighted average brightness is an index determined based on the image data obtained by photographing the rice group to be measured with a predetermined gradation, and the area occupied by each rice to be measured is added to the average brightness of the portion occupied by each rice to be measured A method for evaluating the state of water absorption, which is obtained by dividing the sum of the products obtained by multiplying the number of pixels by the number of all pixels in the portion occupied by the rice group to be measured.

Images