JP7535874B2 - Information processing device, information processing program, and information processing method - Google Patents

Description

The present invention relates to an information processing device, an information processing program, and an information processing method.

Traditionally, in waste incineration facilities, garbage with a high moisture content, such as kitchen waste, compacted garbage located at the bottom of the garbage pit, and sludge, is low-quality garbage with a low calorific value, which causes unstable combustion.

The absorption spectrum of water in the near-infrared wavelength range shows peaks around 970 nm, 1190 nm, 1450 nm, and 1940 nm.

Patent Document 1 proposes a method of imaging a garbage pit to obtain a spectral image of the water absorption band and a spectral image of the water non-absorption band, and determining the moisture content of the surface layer of the garbage by using the difference in the skewness of the brightness value histograms of each spectral image as a moisture index.

JP 2019-158256 A

However, when using spectroscopic images, an infrared camera and an infrared irradiation device are required, taking into account the amount of light.

In addition, when using spectroscopic images, images are taken with a fixed camera, so if there are cliffs or unevenness (i.e. areas where light does not reach) in the garbage pit, or if an object other than garbage, such as a crane, appears in the image, that area becomes a blind spot and the moisture content cannot be measured accurately.

The present invention has been made in consideration of the above points. The object of the present invention is to provide a technology that can solve at least one of the above problems associated with methods that use spectroscopic images.

An information processing device according to a first aspect of the present invention includes:
a data acquisition unit that acquires measurement data from a laser level sensor that irradiates an infrared laser onto the surface of the waste stored in the waste pit and measures the reflection intensity of the surface of the waste;
an estimation unit that uses a trained model that has learned by machine learning the relationship between the measurement data of the surface of the garbage previously acquired by the data acquisition unit and the value indicating the characteristic of the garbage or the label classifying the quality of the garbage, and estimates and outputs the value indicating the characteristic of the garbage or the label classifying the quality of the garbage, using the measurement data of the surface of the garbage newly acquired by the data acquisition unit as input;
Equipped with.

According to this embodiment, a laser-type level sensor that directly measures the infrared laser emitted by itself as a light source is used, which is economical because there is no need to prepare an infrared camera and an infrared irradiation device separately. Also, when an infrared irradiation device serves as a light source and irradiates infrared rays, if there are irregularities on the surface of the garbage stored in the garbage pit, a shadow will be cast on the side opposite the light source, creating a blind spot, making it impossible to measure the moisture content accurately, and in order to prevent the shadow from being cast, it is necessary to install multiple infrared irradiation devices around the camera. However, according to this embodiment, a laser-type level sensor that directly measures the infrared laser emitted by itself as a light source is used, so there is no blind spot caused by a shadow on the side opposite the light source, and the value indicating the characteristics of the garbage or the label classifying the quality of the garbage can be more accurately estimated.

An information processing device according to a second aspect of the present invention is the information processing device according to the first aspect,
The laser level sensor irradiates an infrared laser onto the surface of the garbage stored in the garbage pit and measures the reflection intensity and height of the garbage surface.

According to this embodiment, both the reflection intensity and height of the garbage surface can be measured simultaneously using a laser level sensor, and there is no need to install a separate height measuring device to measure the height of the garbage surface, which is economical. In addition, by using both the reflection intensity and height of the garbage surface as input and using a trained model to estimate a value indicating the characteristics of the garbage or a label classifying the quality of the garbage, more accurate estimation is possible.

An information processing device according to a third aspect of the present invention is the information processing device according to the first or second aspect,
The estimation unit divides the measurement data of the surface of the garbage acquired by the data acquisition unit into multiple blocks, inputs the divided measurement data into the trained model, and estimates and outputs values indicating the characteristics of the garbage or labels classifying the quality of the garbage on a block-by-block basis.

An information processing device according to a fourth aspect of the present invention is an information processing device according to any one of the first to third aspects,
an instruction unit that transmits an operation instruction to a crane control device that controls a crane that mixes or transports the waste, or transmits an operation instruction to a combustion control device that controls the combustion of the waste, based on a value indicating the characteristics of the waste or a label that classifies the quality of the waste estimated by the estimation unit;
It further comprises:

An information processing device according to a fifth aspect of the present invention is an information processing device according to any one of the first to fourth aspects,
The laser level sensor is installed on the crane girder.

According to this aspect, even if a 2D type laser level sensor is used that measures the area below the sensor in two dimensions, it is possible to measure the entire area within the garbage pit by taking measurements each time the crane moves.

An information processing device according to a sixth aspect of the present invention is an information processing device according to any one of the first to fifth aspects,
The waste characteristic is at least one of weight, density, moisture content, and heating value.

An information processing device according to a seventh aspect of the present invention is an information processing device according to any one of the first to sixth aspects,
The machine learning algorithm is at least one of maximum likelihood classification, Boltzmann machines, neural networks, support vector machines, Bayesian networks, sparse regression, decision trees, statistical inference using random forests, reinforcement learning, and deep learning.

A waste treatment plant according to an eighth aspect of the present invention comprises:
A laser level sensor that irradiates an infrared laser onto the surface of the garbage stored in the garbage pit and measures the reflection intensity of the surface of the garbage;
An information processing device according to any one of the first to seventh aspects;
Equipped with.

An information processing program according to a ninth aspect of the present invention,
Computer,
a data acquisition unit that acquires measurement data from a laser level sensor that irradiates an infrared laser onto the surface of the waste stored in the waste pit and measures the reflection intensity of the surface of the waste;
an estimation unit that uses a trained model that has learned by machine learning the relationship between the measurement data of the surface of the garbage previously acquired by the data acquisition unit and the value indicating the characteristic of the garbage or the label classifying the quality of the garbage, and estimates and outputs the value indicating the characteristic of the garbage or the label classifying the quality of the garbage, using the measurement data of the surface of the garbage newly acquired by the data acquisition unit as input;
Function as.

An information processing method according to a tenth aspect of the present invention comprises:
1. A computer-implemented information processing method, comprising:
A step of acquiring measurement data from a laser level sensor that irradiates an infrared laser onto the surface of the waste stored in the waste pit and measures the reflection intensity of the surface of the waste;
using a trained model that has learned by machine learning the relationship between measurement data of the surface of the garbage previously acquired from the laser level sensor and a value indicating a characteristic of the garbage or a label classifying the quality of the garbage, using newly acquired measurement data of the surface of the garbage from the laser level sensor as input, to estimate and output a value indicating a characteristic of the garbage or a label classifying the quality of the garbage;
Includes.

An information processing device according to an eleventh aspect of the present invention includes:
The system is equipped with a teacher data generation unit that generates teacher data by associating measurement data of the surface of the garbage measured by a laser level sensor that irradiates an infrared laser onto the surface of the garbage stored in the garbage pit and measures the reflection intensity of the surface of the garbage with values indicating the characteristics of the garbage or labels classifying the quality of the garbage as correct answers.

An information processing device according to a twelfth aspect of the present invention is the information processing device according to the eleventh aspect,
The apparatus further includes a model construction unit that constructs a trained model by machine learning the training data generated by the training data generation unit.

An information processing device according to a thirteenth aspect of the present invention comprises:
a data acquisition unit that acquires measurement data from a laser level sensor that irradiates an infrared laser onto the surface of the waste placed in the waste hopper and measures the reflection intensity of the surface of the waste;
an estimation unit that uses a trained model that has learned by machine learning the relationship between the measurement data of the surface of the garbage previously acquired by the data acquisition unit and the value indicating the characteristic of the garbage or the label classifying the quality of the garbage, and estimates and outputs the value indicating the characteristic of the garbage or the label classifying the quality of the garbage, using the measurement data of the surface of the garbage newly acquired by the data acquisition unit as input;
Equipped with.

According to this embodiment, a laser-type level sensor that directly measures the infrared laser emitted by itself as a light source is used, which is economical because there is no need to prepare an infrared camera and an infrared irradiation device separately. Also, when an infrared irradiation device serves as a light source and irradiates infrared rays, if there are irregularities on the surface of the garbage put into the input hopper, a shadow will be cast on the opposite side to the light source, creating a blind spot, making it impossible to measure the moisture content accurately, and in order to prevent the shadows from being cast, it is necessary to install multiple infrared irradiation devices around the camera. However, according to this embodiment, a laser-type level sensor that directly measures the infrared laser emitted by itself as a light source is used, so there is no blind spot caused by a shadow on the opposite side to the light source, and the value indicating the characteristics of the garbage or the label classifying the quality of the garbage can be more accurately estimated.

An information processing device according to a fourteenth aspect of the present invention comprises:
A data acquisition unit is installed near the input door and acquires measurement data from a laser level sensor that irradiates an infrared laser onto the surface of the waste discharged from a waste delivery vehicle, such as a compactor truck, and measures the reflection intensity of the surface of the waste.
an estimation unit that uses a trained model that has learned by machine learning the relationship between the measurement data of the surface of the garbage previously acquired by the data acquisition unit and the value indicating the characteristic of the garbage or the label classifying the quality of the garbage, and estimates and outputs the value indicating the characteristic of the garbage or the label classifying the quality of the garbage, using the measurement data of the surface of the garbage newly acquired by the data acquisition unit as input;
Equipped with.

According to this embodiment, a laser-type level sensor that directly measures the infrared laser emitted by itself as a light source is used, which is economical because there is no need to prepare an infrared camera and an infrared irradiation device separately. Also, when an infrared irradiation device serves as a light source and irradiates infrared rays, if there are irregularities on the surface of the garbage removed from the garbage delivery vehicle (e.g., a compactor truck), a shadow will be cast on the side opposite the light source, creating a blind spot, making it impossible to accurately measure the moisture content, and in order to prevent the shadow from being cast, multiple infrared irradiation devices must be installed around the camera. However, according to this embodiment, a laser-type level sensor that directly measures the infrared laser emitted by itself as a light source is used, so there is no blind spot caused by a shadow cast on the side opposite the light source, and the value indicating the characteristics of the garbage or the label classifying the quality of the garbage can be more accurately estimated.

An information processing device according to a fifteenth aspect of the present invention comprises:
A data acquisition unit that acquires measurement data from a laser level sensor that irradiates an infrared laser onto the surface of the incineration ash stored in the ash pit and measures the reflection intensity of the surface of the incineration ash;
An estimation unit that uses a trained model that has machine-learned the relationship between the measurement data of the surface of the incineration ash previously acquired by the data acquisition unit and the value indicating the characteristics of the incineration ash or the label classifying the quality of the incineration ash, and estimates and outputs the value indicating the characteristics of the incineration ash or the label classifying the quality of the ash, using the measurement data of the surface of the incineration ash newly acquired by the data acquisition unit as input;
Equipped with.

According to this embodiment, a laser-type level sensor that directly measures the infrared laser emitted by itself as a light source is used, which is economical because there is no need to prepare an infrared camera and an infrared irradiation device separately. Also, when an infrared irradiation device serves as a light source and irradiates infrared rays, if there are irregularities on the surface of the incineration ash stored in the ash pit, a shadow will be cast on the opposite side of the light source, creating a blind spot, making it impossible to measure the moisture content accurately. In order to prevent shadows, however, multiple infrared irradiation devices must be installed around the camera. According to this embodiment, a laser-type level sensor that directly measures the infrared laser emitted by itself as a light source is used, so there is no blind spot caused by a shadow on the opposite side of the light source, and the value indicating the characteristics of the incineration ash or the label classifying the quality of the incineration ash can be more accurately estimated.

The present invention can solve at least one of the above problems associated with methods that use spectroscopic images.

FIG. 1 is a schematic diagram showing the configuration of a waste incineration plant according to one embodiment. FIG. 2 is a block diagram showing a configuration of an information processing device according to an embodiment. FIG. 3 is a flowchart showing an example of an information processing method performed by the information processing device according to an embodiment. FIG. 4 is a diagram for explaining an example of the operation of the estimation unit. FIG. 5 is a schematic diagram showing the configuration of a waste incineration plant according to a first modified example of an embodiment. FIG. 6 is a schematic diagram showing the configuration of a waste incineration plant according to a second modified example of an embodiment. FIG. 7 is a schematic diagram showing the configuration of a waste incineration plant according to a third modified example of an embodiment.

Below, an embodiment of the present invention will be described in detail with reference to the attached drawings. Note that in the following description and in the drawings used in the following description, the same reference numerals will be used for parts that can be configured identically, and duplicate descriptions will be omitted.

Figure 1 is a schematic diagram showing the configuration of a waste incineration plant 100 according to one embodiment.

As shown in FIG. 1, the waste incineration plant 100 includes a platform 21 on which a waste delivery vehicle (e.g., a packer truck) 22 carrying waste is parked, a waste pit 3 in which waste input from the platform 21 is stored, a crane 5 for stirring and transporting the waste stored in the waste pit 3, a feed hopper 4 into which the waste transported by the crane 5 is input, an incinerator 1 for incinerating the waste input from the feed hopper 4, a waste heat boiler 2 for recovering exhaust heat from the exhaust gas generated in the incinerator 1, and an ash pit 7 in which incineration ash is stored. The type of incinerator 1 is not limited to a stoker furnace as shown in FIG. 1, but also includes a fluidized furnace (also called a fluidized bed furnace). The structure of the waste pit 3 is not limited to a one-stage pit as shown in FIG. 1, but also includes a two-stage pit. The waste incineration plant 100 is also provided with a crane control device 50 that controls the operation of the crane 5, and a combustion control device 20 that controls the combustion of waste in the incinerator 1.

Waste brought in loaded on a waste delivery vehicle (e.g., a compactor truck) 22 is dumped into the waste pit 3 from a platform 21 near the dump door, and stored in the waste pit 3. The waste stored in the waste pit 3 is stirred by a crane 5 and transported to the dump hopper 4 by the crane 5, dumped into the incinerator 1 via the dump hopper 4, and incinerated and disposed of in the incinerator 1.

As shown in FIG. 1, the waste incineration plant 100 is equipped with a laser level sensor 6 that measures the surface of the waste stored in the waste pit 3 by irradiating it with an infrared laser, and an information processing device 10 that estimates values indicating the characteristics of the waste or labels that classify the quality of the waste.

The laser level sensor 6 is installed above the garbage pit 3, and measures the reflection intensity of the garbage surface by irradiating an infrared laser onto the surface of the garbage stored in the garbage pit 3. The laser level sensor 6 may also measure the reflection intensity and height (level) of the garbage surface simultaneously by irradiating an infrared laser onto the surface of the garbage stored in the garbage pit 3. In this specification, the infrared laser refers to laser light whose central wavelength is in the near-infrared wavelength range (700 to 2000 nm). The laser level sensor 6 is preferably one in which the central wavelength of the infrared laser it irradiates is around 970 nm, around 1190 nm, around 1450 nm, or around 1940 nm, where a peak appears in the absorption spectrum of water in the near-infrared wavelength range. Multiple laser level sensors 6 may be installed.

A LIDAR (Light Detection and Ranging) type sensor may be used as the laser level sensor 6. The laser level sensor 6 may be a 2D type that measures the area below the sensor in two dimensions, or a 3D type that measures the area below the sensor in three dimensions.

In the illustrated example, the laser level sensor 6 is installed on the crane girder (the structure that suspends the crane 5 and moves it), but this is not limited to this and it may be set on the ceiling or upper part of the side wall of the garbage pit 3. When the laser level sensor 6 is installed on the crane girder, it is possible to measure the surface of the garbage stored in the garbage pit 3 each time the crane 5 moves.

Next, we will explain the configuration of the information processing device 10 that estimates values indicating the characteristics of the garbage in the garbage pit 3 or labels that classify the quality of the garbage. Figure 2 is a block diagram showing the configuration of the information processing device 10. The information processing device 10 is composed of one or more computers.

As shown in FIG. 2, the information processing device 10 has a communication unit 11, a control unit 12, and a storage unit 13. Each unit is connected to each other via a bus so that they can communicate with each other.

Of these, the communication unit 11 is a communication interface between the information processing device 10 and each of the devices, the laser level sensor 6, the crane control device 50, and the combustion control device 20. The communication unit 11 transmits and receives information between the information processing device 10 and each of the devices, the laser level sensor 6, the crane control device 50, and the combustion control device 20.

The storage unit 13 is a non-volatile data storage such as a hard disk or a flash memory. The storage unit 13 stores various data handled by the control unit 12. The storage unit 13 also stores teacher data 13b generated by a teacher data generation unit 12d (described later), a trained model 13a constructed by a model construction unit 12e, and measurement data 13c acquired by a data acquisition unit 12a.

The control unit 12 is a control means that performs various processes of the information processing device 10. As shown in FIG. 2, the control unit 12 has a data acquisition unit 12a, an estimation unit 12b, an instruction unit 12c, a teacher data generation unit 12d, and a model construction unit 12e. Each of these units may be realized by a processor in the information processing device 10 executing a predetermined program, or may be implemented in hardware.

The data acquisition unit 12a acquires the reflection intensity (or reflection intensity and height) of the surface of the garbage measured by the laser level sensor 6 as measurement data 13c. The acquired measurement data 13c is stored in the memory unit 13.

The teacher data generating unit 12d generates teacher data 13b by associating the measurement data 13c (measurement data of multiple points included in one area) of the surface of the garbage measured by the laser level sensor 6 with a value (one value) indicating the characteristics of the garbage in that area or a label (one label) classifying the quality of the garbage as the correct answer. The generated teacher data 13b is stored in the memory unit 13.

Here, the characteristics of the waste may be, for example, at least one of the weight, density, moisture content, and heat value of the waste. Among the values indicating the characteristics of the waste, for example, the weight and density can be determined based on the operation history of the waste incineration plant 100 when the waste measured by the laser level sensor 6 was fed into the feed hopper 4. The moisture content and heat value can be determined based on the operation history of the waste incineration plant 100 at a time taking into consideration the time lag between the time the waste measured by the laser level sensor 6 is fed into the feed hopper 4 and the time it affects the combustion state. The time lag between the time the waste is fed into the feed hopper 4 and the time it affects the combustion state may be appropriately determined based on experiments, simulations, and the operator's experience.

The label classifying the quality of the waste may be, for example, a label classifying the waste into three levels, namely, normal waste, high-quality waste (i.e., high-calorie waste), and low-quality waste (i.e., low-calorie waste), or a label classifying the waste into two levels, or four or more levels. The label classifying the quality of the waste may be identified by the operator based on empirical rules by visually checking an image of the surface of the waste captured at the same time as the measurement by the laser level sensor 6.

The teacher data generating unit 12d may divide the surface of the garbage stored in the garbage pit 3 into multiple blocks, and generate teacher data 13b for each block by associating the block-based measurement data 13c (measurement data for multiple points contained in the block) with a value (one value) indicating the characteristics of the garbage in that block or a label (one label) classifying the quality of the garbage as the correct answer.

The model construction unit 12e constructs the learned model 13a by machine learning the teacher data 13b generated by the teacher data generation unit 12d. The model construction unit 12e may construct the learned model 13a by machine learning the teacher data 13b in block units generated by the teacher data generation unit 12d. The constructed learned model 13a (learned parameters for realizing the learned model 13a in the program structure) is stored in the memory unit 13.

The machine learning algorithm may be at least one of maximum likelihood classification, Boltzmann machines, neural networks, support vector machines, Bayesian networks, sparse regression, decision trees, statistical inference using random forests, reinforcement learning, and deep learning.

A specific method of model construction will be described below. That is, the model construction unit 12e uses the measurement data 13c of multiple points (e.g., 5 x 4 = 20 points) included in a block from the block-based teacher data 13b as input values, and a value (one value) indicating the characteristics of the garbage corresponding to the block or a label (one label) classifying the quality of the garbage as output values, and machine-learns the relationship between the input values and the output values using the machine learning algorithm described above to find parameters that relate the input values and the output values.

For example, when a convolutional neural network (CNN), which is one of the deep learning methods, is used as the machine learning algorithm, the model construction unit 12e uses the multiple measurement data 13c included in the teacher data 13b of the block as input values of the convolutional neural network, and uses values indicating the characteristics of the garbage corresponding to the block or labels classifying the quality of the garbage as output values, and machine-learns the relationship between the input values and the output values using methods such as gradient descent and backpropagation, and optimizes the parameters of each node of the network to obtain parameters that relate the input values and the output values, i.e., a model 13a as learned parameters.

Figure 4 is a diagram for explaining an example of the operation of the estimation unit 12b. As shown in Figure 4, the estimation unit 12b receives as input the measurement data (measurement data of multiple points included in one area) of the surface of the garbage newly acquired by the data acquisition unit 12a, and uses the trained model 13a to estimate and output a value (one value) indicating the characteristics of the garbage in that area or a label (one label) classifying the quality of the garbage.

As shown in FIG. 4, the estimation unit 12b may divide the measurement data of the surface of the garbage acquired by the data acquisition unit 12a into a plurality of blocks (3×3=9 in the illustrated example), input the measurement data of each divided block (5×4=20 points of measurement data included in one block in the illustrated example) to the trained model 13a constructed by machine learning the teacher data 13b of each block, and estimate and output a value (one value) indicating the characteristics of the garbage of the block or a label (one label) classifying the quality of the garbage for each block. For example, when dividing into 3×3=9 blocks, the estimation unit 12b may obtain one estimated value (a value indicating the characteristics of the garbage or a label classifying the quality of the garbage) for each block, thereby finally obtaining nine estimated values corresponding to each of the nine blocks. The estimation unit 12b may link each block into which the garbage pit 3 is divided and the address allocation of the garbage pit 3, and estimate and output a value indicating the characteristics of the garbage or a label classifying the quality of the garbage for each address.

The instruction unit 12c transmits operation instructions to the crane control device 50 that controls the crane 5 that mixes or transports the garbage, or transmits operation instructions to the combustion control device 20 that controls the combustion of the garbage, based on the value indicating the garbage characteristics estimated by the estimation unit 12b or the label that classifies the garbage quality.

Specifically, for example, the instruction unit 12c may send operation instructions to the crane control device 50 so that the characteristics or quality of the garbage stored in the garbage pit 3 are uniform in each block, or so that the characteristics or quality of the garbage to be currently put into the incinerator 1 are close to the characteristics or quality of the garbage put into the incinerator 1 last time, based on the values indicating the characteristics of the garbage estimated by the estimation unit 12b on a block-by-block basis or the labels classifying the quality of the garbage.

Alternatively, the instruction unit 12c may grasp the characteristics or quality of the garbage that is grabbed by the crane 5 and transported together with the garbage from the garbage pit 3 to the input hopper 4 based on the values indicating the characteristics of the garbage estimated by the estimation unit 12b on a block-by-block basis or the labels classifying the quality of the garbage, and send an operation instruction to the combustion control device 20 to control the combustion of the garbage (for example, increasing or decreasing the combustion temperature, increasing or decreasing the combustion time, increasing or decreasing the amount of air, increasing or decreasing the amount of garbage fed, or increasing or decreasing the amount of chemical sprayed) according to the characteristics or quality of the garbage that is input together with the garbage into the incinerator 1 via the input hopper 4. In this case, the time lag from input into the incinerator 1 to the time when the combustion state is affected may be appropriately determined based on experiments, simulations, or the experience of the operator.

Next, an example of an information processing method using the information processing device 10 having such a configuration will be described. FIG. 3 is a flowchart showing an example of the information processing method.

As shown in FIG. 3, first, the teacher data generating unit 12d generates teacher data 13b by associating the measurement data 13c of the surface of the garbage previously measured by the laser level sensor 6 with a value indicating the characteristics of the garbage or a label classifying the quality of the garbage as a correct answer (step S11). The generated teacher data 13b is stored in the memory unit 13.

Next, the model construction unit 12e constructs a trained model 13a by machine learning the teacher data 13b generated by the teacher data generation unit 12d (step S12). The constructed trained model 13a (trained parameters for realizing the trained model 13a in the program structure) is stored in the storage unit 13.

Next, the data acquisition unit 12a acquires the reflection intensity (or reflection intensity and height) of the surface of the garbage newly measured by the laser level sensor 6 as measurement data 13c (step S13). The acquired measurement data 13c is stored in the memory unit 13.

Next, the estimation unit 12b divides the measurement data of the surface of the garbage acquired by the data acquisition unit 12a into multiple blocks, inputs the measurement data for each divided block to the trained model 13a, and estimates and outputs a value indicating the characteristics of the garbage or a label classifying the quality of the garbage for each block (step S14).

Next, the instruction unit 12c transmits an operation instruction to the crane control device 50 or the combustion control device 20 based on the value indicating the characteristics of the garbage or the label classifying the quality of the garbage estimated on a block-by-block basis by the estimation unit 12b (step S15).

As mentioned in the section on problems to be solved by the invention, the method using spectral images requires an infrared camera and an infrared irradiation device, taking into account the amount of light. Also, in the method using spectral images, images are taken with a fixed camera, so if there are cliffs or unevenness (i.e. areas where light does not reach) in the garbage pit, or if something other than garbage, such as a crane, appears in the image, that area becomes a blind spot, and it may not be possible to measure the moisture content accurately.

In contrast, according to the present embodiment, the laser level sensor 6, which acts as a light source and directly measures the infrared laser emitted by itself, is used, which is economical because there is no need to prepare an infrared camera and an infrared emitting device separately. Also, when the infrared emitting device acts as a light source and emits infrared rays, if there are irregularities on the surface of the garbage stored in the garbage pit 3, a shadow will be cast on the opposite side of the light source, creating a blind spot, and accurate moisture content cannot be measured. In order to prevent shadows, however, multiple infrared emitting devices must be installed around the camera. According to this embodiment, the laser level sensor 6, which acts as a light source and directly measures the infrared laser emitted by itself, is used, so there is no blind spot caused by a shadow on the opposite side of the light source, and the value indicating the characteristics of the garbage or the label classifying the quality of the garbage can be more accurately estimated.

In addition, according to this embodiment, both the reflection intensity and height of the garbage surface can be measured simultaneously by the laser level sensor 6, and there is no need to install a separate height measuring device to measure the height of the garbage surface, which is economical. Furthermore, more accurate estimation is possible by using the trained model 13a to estimate values indicating the characteristics of the garbage or labels classifying the quality of the garbage using both the reflection intensity and height of the garbage surface as input.

In addition, according to this embodiment, since the laser level sensor 6 is installed on the crane girder, even if it is a 2D type laser level sensor that measures the area below the sensor two-dimensionally, it is possible to measure the entire area within the garbage pit 3 by taking measurements each time the crane 5 moves.

Note that various modifications can be made to the above-described embodiment. Below, we will explain some variations of the above-described embodiment.

Figure 5 is a schematic diagram showing the configuration of a waste incineration plant 100 according to a first modified example. In the example shown in Figure 5, a laser level sensor 61 is installed above the input hopper 4, and an infrared laser is irradiated onto the surface of the waste input into the input hopper 4 to measure the reflection intensity (or reflection intensity and height) of the surface of the waste.

The data acquisition unit 12a of the information processing device 10 acquires measurement data of the surface of the garbage put into the input hopper 4 from the laser level sensor 61. The estimation unit 12b uses a trained model 13a that performs machine learning on the relationship between the measurement data of the surface of the garbage previously acquired by the data acquisition unit 12a and the value indicating the characteristics of the garbage or the label classifying the quality of the garbage to input the measurement data of the surface of the garbage newly acquired by the data acquisition unit 12a, and estimates and outputs the value indicating the characteristics of the garbage or the label classifying the quality of the garbage. The instruction unit 12c transmits an operation instruction to the combustion control device 20 based on the value indicating the characteristics of the garbage or the label classifying the quality of the garbage estimated by the estimation unit 12b. This makes it possible to enhance the combustion control.

Figure 6 is a schematic diagram showing the configuration of a waste incineration plant 100 according to a second modified example. In the example shown in Figure 6, a laser level sensor 62 is installed near the input door, and an infrared laser is irradiated onto the surface of the waste removed from the waste delivery vehicle (e.g., a compactor truck) 22 to measure the reflection intensity (or reflection intensity and height) of the surface of the waste.

The data acquisition unit 12a of the information processing device 10 acquires measurement data of the surface of the garbage carried out from the garbage carrying vehicle (e.g., a packer truck) 22 from the laser level sensor 62. The estimation unit 12b uses the trained model 13a, which performs machine learning on the relationship between the measurement data of the surface of the garbage previously acquired by the data acquisition unit 12a and the value indicating the characteristics of the garbage or the label classifying the quality of the garbage, to input the measurement data of the surface of the garbage newly acquired by the data acquisition unit 12a, and estimates and outputs the value indicating the characteristics of the garbage or the label classifying the quality of the garbage. The instruction unit 12c transmits operation instructions to the crane control device 50 based on the value indicating the characteristics of the garbage or the label classifying the quality of the garbage estimated by the estimation unit 12b. This makes it possible to enhance the pit work in the garbage pit 3. It is also possible to utilize it for settling the processing outsourcing fee according to the quality of the garbage (i.e., moisture content, etc.).

Figure 7 is a schematic diagram showing the configuration of a waste incineration plant 100 according to a third modified example. In the example shown in Figure 7, a laser level sensor 63 is installed above the ash pit 7, and an infrared laser is irradiated onto the surface of the incineration ash stored in the ash pit 7 to measure the reflection intensity (or reflection intensity and height) of the surface of the incineration ash.

The data acquisition unit 12a of the information processing device 10 acquires measurement data of the surface of the incineration ash stored in the ash pit 7 from the laser level sensor 63. The estimation unit 12b uses the trained model 13a, which performs machine learning on the relationship between the measurement data of the surface of the incineration ash previously acquired by the data acquisition unit 12a and the value indicating the characteristics of the incineration ash or the label classifying the quality of the incineration ash, to input the measurement data of the surface of the incineration ash newly acquired by the data acquisition unit 12a, and estimates and outputs the value indicating the characteristics of the incineration ash or the label classifying the quality of the incineration ash. The instruction unit 12c transmits an operation instruction to an ash transport crane control device (not shown) installed in the ash pit 7 based on the value indicating the characteristics of the incineration ash or the label classifying the quality of the incineration ash estimated by the estimation unit 12b. This makes it possible to enhance the pit work in the ash pit 7. The instruction unit 12c may control the amount of water sprayed in an ash extrusion device or ash discharge device (not shown) based on the value indicating the characteristics of the incineration ash estimated by the estimation unit 12b or the label classifying the quality of the incineration ash.

Here, the value indicating the characteristics of the incineration ash may be, for example, the moisture content (amount of moisture). Also, the label classifying the quality of the incineration ash may be, for example, a label classifying the ash into two levels, high moisture content and low moisture content, or a label classifying the ash into three or more levels.

In the above-mentioned embodiment and modified example, a laser level sensor is installed in a waste incineration plant. However, the information processing device may be configured to install a laser level sensor not only in waste incineration plants, but also on the ceiling of a fuel storage facility in a wood biomass power plant, for example, and to determine which parts of the fuel are wet or dry based on the measurement data of the laser level sensor and select the fuel to be input.

That is, a laser level sensor is installed on the ceiling of the fuel storage facility of a wood biomass power plant, and an infrared laser is irradiated onto the surface of the fuel stored in the fuel storage facility to measure the reflection intensity (or reflection intensity and height) of the fuel surface.

The data acquisition unit 12a of the information processing device 10 acquires measurement data of the surface of the fuel stored in the fuel storage facility from a laser level sensor. The estimation unit 12b uses a trained model 13a that performs machine learning on the relationship between the measurement data of the fuel surface previously acquired by the data acquisition unit 12a and the value indicating the characteristics of the fuel or the label classifying the quality of the fuel, and estimates and outputs the value indicating the characteristics of the fuel or the label classifying the quality of the fuel, using the measurement data of the fuel surface newly acquired by the data acquisition unit 12a as input. The instruction unit 12c transmits an operation instruction regarding which part of the fuel should be poured to a fuel transport means (not shown) installed in the fuel storage facility, based on the value indicating the characteristics of the fuel or the label classifying the quality of the incineration ash estimated by the estimation unit 12b.

In the above-described embodiment and modified examples, part of the processing of the control unit 12 may be performed on a cloud server other than the information processing device 10, rather than on the information processing device 10. Part of the storage unit 13 may be located on a cloud server other than the information processing device 10, rather than on the information processing device 10.

For example, the processing of the teacher data generation unit 12d may be executed on a cloud server to generate the teacher data 13b, or the processing of the model construction unit 12e may be executed on a cloud server to construct the trained model 13a. In addition, the processing of the estimation unit 12b may be executed on the cloud server using the trained model 13a constructed on the cloud server, or the information processing device 10 may download the trained model 13a (trained parameters) from the cloud server and use it within the information processing device 10 to execute the processing of the estimation unit 12b.

The control unit 12 may periodically monitor the estimation results of the estimation unit 12b and determine whether or not the trained model 13a needs to be revised and updated.

For example, the control unit 12 uses an edge server to determine whether the estimation result of the estimation unit 12b is normal or abnormal, and if an abnormality is detected, it determines whether the measurement data and estimation result pose a problem in the operation of the incinerator 1. If it is determined that there is an operational problem, the teacher data generation unit 12d prepares new teacher data 13b based on the labels that classify the quality of the waste that were reassigned by an experienced operator to the measurement data where an abnormality was detected, and the model construction unit 12e learns the newly prepared teacher data 13b to generate a trained model 13a.

Although the above describes the embodiments and modifications by way of example, the scope of the present invention is not limited to these, and can be modified or altered according to the purpose within the scope of the claims. Furthermore, each embodiment and modification can be combined as appropriate within the scope that does not cause any contradiction in the processing content.

In addition, the information processing device 10 according to this embodiment may be configured by one or more computers, but the program for realizing the information processing device 10 in one or more computers and the recording medium on which the program is non-temporarily recorded are also covered by this case.

1 Incinerator 2 Combustion device 3 Garbage pit 4 Input hopper 5 Crane 6, 61 to 63 Laser level sensor 7 Ash pit 10 Information processing device 11 Communication unit 12 Control unit 12a Data acquisition unit 12b Estimation unit 12c Instruction unit 12d Teacher data generation unit 12e Model construction unit 13 Memory unit 13a Learned model 13b Teacher data 13c Measurement data 20 Combustion control device 21 Platform 22 Garbage delivery vehicle 50 Crane control device 100 Garbage incineration plant

Claims (15)

a data acquisition unit that acquires measurement data from a laser level sensor that irradiates an infrared laser onto the surface of the waste stored in the waste pit and measures the reflection intensity of the surface of the waste;
an estimation unit that uses a trained model that has undergone machine learning to learn the relationship between measurement data of multiple points included in one area on the surface of the garbage previously acquired by the data acquisition unit and a value indicating the characteristics of the garbage in that area or a label classifying the quality of the garbage, and estimates and outputs a value indicating the characteristics of the garbage in that area or a label classifying the quality of the garbage, using measurement data of multiple points included in one area on the surface of the garbage newly acquired by the data acquisition unit as input;
An information processing device comprising: The laser level sensor measures the reflection intensity and height of the surface of the garbage stored in the garbage pit by irradiating an infrared laser onto the surface of the garbage,
the estimation unit uses a trained model that has undergone machine learning to learn the relationship between the measurement data of the reflection intensity or the reflection intensity and height of multiple points included in one area of the surface of the garbage previously acquired by the data acquisition unit and a value indicating the characteristics of the garbage in that area or a label classifying the quality of the garbage, and estimates and outputs the value indicating the characteristics of the garbage in that area or the label classifying the quality of the garbage, using the measurement data of the reflection intensity or the reflection intensity and height of multiple points included in one area of the surface of the garbage newly acquired by the data acquisition unit as input;
2. The information processing apparatus according to claim 1, the estimation unit divides the measurement data of the surface of the garbage acquired by the data acquisition unit into a plurality of blocks, inputs the measurement data for each divided block, which is the measurement data for a plurality of points included in one block, into the trained model, and estimates and outputs a value indicating the characteristics of the garbage in that block or a label classifying the quality of the garbage on a block-by-block basis.
3. The information processing apparatus according to claim 1, wherein the information processing apparatus is a computer. an instruction unit that transmits an operation instruction to a crane control device that controls a crane that mixes or transports the waste, or transmits an operation instruction to a combustion control device that controls the combustion of the waste, based on a value indicating the characteristics of the waste or a label that classifies the quality of the waste estimated by the estimation unit;
4. The information processing apparatus according to claim 1, further comprising: The characteristic of the waste is at least one of weight, density, moisture content, and heat value.
5. The information processing device according to claim 1, wherein the information processing device is a computer. The machine learning algorithm is at least one of a maximum likelihood classification method, a Boltzmann machine, a neural network, a support vector machine, a Bayesian network, a sparse regression, a decision tree, a statistical estimation using a random forest, a reinforcement learning, and a deep learning;
6. The information processing device according to claim 1 , A laser level sensor that irradiates an infrared laser onto the surface of the garbage stored in the garbage pit and measures the reflection intensity of the surface of the garbage;
An information processing device according to any one of claims 1 to 6 ,
A waste treatment plant comprising: The laser level sensor is installed on a crane girder.
8. A waste treatment plant according to claim 7. Computer,
a data acquisition unit that acquires measurement data from a laser level sensor that irradiates an infrared laser onto the surface of the waste stored in the waste pit and measures the reflection intensity of the surface of the waste;
an estimation unit that uses a trained model that has undergone machine learning to learn the relationship between measurement data of multiple points included in one area on the surface of the garbage previously acquired by the data acquisition unit and a value indicating the characteristics of the garbage in that area or a label classifying the quality of the garbage, and estimates and outputs a value indicating the characteristics of the garbage in that area or a label classifying the quality of the garbage, using measurement data of multiple points included in one area on the surface of the garbage newly acquired by the data acquisition unit as input;
An information processing program characterized by causing the program to function as follows. 1. A computer-implemented information processing method, comprising:
A step of acquiring measurement data from a laser level sensor that irradiates an infrared laser onto the surface of the waste stored in the waste pit and measures the reflection intensity of the surface of the waste;
A step of estimating and outputting a value indicating the characteristics of the garbage in one area or a label classifying the quality of the garbage using a trained model that has machine-learned the relationship between measurement data of multiple points included in one area of the garbage surface previously acquired from the laser level sensor and a value indicating the characteristics of the garbage in that area or a label classifying the quality of the garbage, using measurement data of multiple points included in one area of the garbage surface newly acquired from the laser level sensor as input;
13. An information processing method comprising: An information processing device characterized by comprising a teacher data generation unit that generates teacher data by associating, as a correct answer, one value indicating the characteristics of the garbage in the area or one label classifying the quality of the garbage, with measurement data of a plurality of points included in one area of the garbage surface measured by a laser level sensor that irradiates an infrared laser onto the surface of the garbage stored in a garbage pit and measures the reflection intensity of the garbage surface. The information processing apparatus according to claim 11 , further comprising a model construction unit that constructs a trained model by machine learning the training data generated by the training data generation unit. a data acquisition unit that acquires measurement data from a laser level sensor that irradiates an infrared laser onto the surface of the waste placed in the waste hopper and measures the reflection intensity of multiple points included in one area on the surface of the waste;
an estimation unit that uses a trained model that has undergone machine learning to learn the relationship between measurement data of multiple points included in one area on the surface of the garbage previously acquired by the data acquisition unit and a value indicating the characteristics of the garbage in that area or a label classifying the quality of the garbage, and estimates and outputs a value indicating the characteristics of the garbage in that area or a label classifying the quality of the garbage, using measurement data of multiple points included in one area on the surface of the garbage newly acquired by the data acquisition unit as input;
An information processing device comprising: a data acquisition unit that is installed near the input door and acquires measurement data from a laser level sensor that irradiates an infrared laser onto the surface of the waste discharged from the waste delivery vehicle and measures the reflection intensity of the surface of the waste;
an estimation unit that uses a trained model that has undergone machine learning to learn the relationship between measurement data of multiple points included in one area on the surface of the garbage previously acquired by the data acquisition unit and a value indicating the characteristics of the garbage in that area or a label classifying the quality of the garbage, and estimates and outputs a value indicating the characteristics of the garbage in that area or a label classifying the quality of the garbage, using measurement data of multiple points included in one area on the surface of the garbage newly acquired by the data acquisition unit as input;
An information processing device comprising: A data acquisition unit that acquires measurement data from a laser level sensor that irradiates an infrared laser onto the surface of the incineration ash stored in the ash pit and measures the reflection intensity of the surface of the incineration ash;
An estimation unit that uses a trained model that performs machine learning on the relationship between the measurement data of multiple points included in one area of the surface of the incineration ash previously acquired by the data acquisition unit and one value indicating the characteristics of the incineration ash in that area or one label classifying the quality of the incineration ash, and estimates and outputs one value indicating the characteristics of the incineration ash in that area or one label classifying the quality of the incineration ash, using the measurement data of multiple points included in one area of the surface of the incineration ash newly acquired by the data acquisition unit as input;
An information processing device comprising: