JP3562871B2 - Image processing device - Google Patents

Description

[0001]
[Industrial applications]
The present invention relates to an image processing apparatus that calculates a feature amount such as a center of gravity of an object included in a two-dimensional image.
[0002]
[Prior art]
2. Description of the Related Art Conventionally, the following method is known as a method of detecting the center of gravity of an object included in a two-dimensional image. That is, when capturing the target object image for each line, the moment from the reference coordinates of the target object is obtained from the distance of the portion occupied by the target object and the distance from the reference coordinates to the center of the target object. This operation is performed for each line where the object exists, and the sum of all moments is obtained. Further, a method of calculating the sum of all the lines for the distance occupied by the target object and then dividing the sum of all moments by the sum of all run lengths to detect the position of the center of gravity of the target object with respect to the reference coordinates.
[0003]
The method of detecting the position of the center of gravity is disclosed in, for example, Japanese Patent Publication No. 56-30802, and the method of detecting the position of the center of gravity of the related art will be specifically described with reference to FIG. In FIG. 11, the center of gravity (x, y) of the object image indicated by oblique lines is calculated by the following equation (1).
_{x = Σ {l i × (} a i + l i / 2)} / Σl i
_{y = Σ (l i × p} i) / Σl i ········ (1)
Here, l _i is the run length of the portion occupied by the object in each line, a _i is the run length from the y reference coordinate axis until the object is detected, and p _i is the run length from the x reference coordinate axis until the object is detected. , And i = 1 to n.
[0004]
Therefore, in the target image shown in FIG. 11, the center of gravity (x, y) can be calculated from the above equation (1) as shown in the following equation (2).
x = [{3 × (2 + 3/2)} + {5 × (1 + 5/2)} + {3 × (2 + 3/2)}] / (3 + 5 + 3) = 3.5
y = {(3 × 1) + (5 × 2) + (3 × 3)} / (3 + 5 + 3) = 2 (2)
[0005]
[Problems to be solved by the invention]
In an information processing apparatus, an image inspection apparatus, and the like using an imaging device (CCD camera or the like), how to calculate the position of the center of gravity of the imaged object at high speed and accurately is an issue.
[0006]
By the way, in the above-described conventional center-of-gravity point detection method, as shown in FIG. 12, when a defect such as a scratch is mixed in an image corresponding to an object, an arithmetic processing is required to calculate an accurate center-of-gravity position. There was a problem that control became complicated and hardware for realizing this became large-scale. Also, when there are a plurality of objects in the field of view of the imaging apparatus and it is necessary to calculate the position of the center of gravity or the like for each of these objects, the moment for the plurality of objects is obtained for each line, and this is calculated as the object. It has been very difficult to sort a plurality of objects by sorting each time.
[0007]
SUMMARY OF THE INVENTION The present invention has been made to solve the above problems in the conventional image processing apparatus, and the invention according to claim 1 uses input information based on foresight information set in advance from a specific format of an image handled by an imaging apparatus. extracts only the object image from the image, thereby enabling the determination of the characteristic quantity of the object image accurately and at high speed, the image data other than the object image data from the image data cut out from the input image relatively easily It is an object of the present invention to provide an image processing device capable of making a determination . The invention according to claim 2, intended to provide an image processing apparatus that the image data other than the object image data can be easily processed as unnecessary data and invention according to claim 3, the object It is an object of the present invention to provide an image processing apparatus capable of determining the position of the center of gravity of an image , and the invention according to claim 4 can more easily process image data other than target object image data as unnecessary data. It is an object of the present invention to provide an image processing apparatus as described above .
[0008]
Means and Action for Solving the Problems
In order to solve the above problem, the invention according to claim 1 two-dimensionally stores an m × n image area including an object image for which a feature amount in an image input by an imaging unit is to be calculated and processed. Means for detecting and determining the presence of image data other than the object image data included in the storage means based on preset foresight information; and means for detecting and detecting the image data other than the object image data in accordance with the result of the detection / determination means. and means for processing the image data as unnecessary data, in the image processing apparatus after performing the extraction of the object image data to perform the calculation processing of the feature amount of the object image, the object image data Means for detecting and determining the presence of image data other than the image data include means for performing X projection and Y projection of image data included in the m × n image area including the target object image, and reference pixels for the m × n image area. Out of Means for examining the X projection and the Y projection toward, determining a coordinate axis having no projection data as a discontinuous point, and determining image data outside the coordinate axis as image data other than the object image data. It is characterized by having.
[0009]
In the image processing apparatus configured as described above, it is possible to remove adjacent image data other than the target object image data or erroneous image data due to scratches or dirt by using preset information, and to obtain an accurate feature amount of the target object image data. Can be calculated at high speed. Further , a means for detecting and determining the presence of image data other than the target object image data in the mxn image area can be realized with a simple configuration.
[0010]
According to a second aspect of the present invention, in the image processing apparatus according to the first aspect, the means for processing the image data other than the target object image data as unnecessary data includes a unit other than the target object image data determined by the determination unit. It is constituted by means for masking image data. Thus, a means for processing image data other than target object image data as unnecessary data can be realized with a simple configuration. According to a third aspect of the present invention, in the image processing apparatus according to the first or second aspect, the calculation of the feature amount of the target image is a center of gravity calculation of the target image. As described above, by setting the feature amount of the object image as the center of gravity, it is possible to calculate the center of gravity of a plurality of object images at high speed and accurately. According to a fourth aspect of the present invention, in the image processing apparatus according to the second aspect, the mask unit is configured to control the other projection based on a determination result of the determination unit regarding one of the X projection and the Y projection. And masking image data other than the target object image data. As described above, unnecessary data can be more easily processed by masking image data other than the target object image data related to the other projection based on the determination result of the determination unit related to one projection.
[0011]
【Example】
Next, examples will be described. The image processing apparatus for extracting features from an object image according to the present invention can be realized by relatively simple hardware or software. Here, an embodiment realized by hardware will be specifically described.
[0012]
First, an example of an image format targeted by the image processing apparatus according to the present invention will be described with reference to FIG. In this example of the image format, as shown in FIG. 1, a plurality of code data groups 2 exist between the reference images 1A and 1B, a distance L1 from the reference images 1A and 1B to the code data, and a distance L2 between the code data. Has a predetermined value (foresight information). In this embodiment, the center coordinates of all the code data and the center position of the reference image estimated from the foresight information (L1, L2) are set as true center coordinates, and the subsequent processing is performed.
[0013]
FIG. 2 shows an enlarged image of a portion surrounded by an ellipse in FIG. In FIG. 2, at least one code data is cut out from a two-dimensional image input from the image input device, and this image signal is set as an m × n area for storing in a matrix, where m = 5 and n = 7. Is shown. In the example shown in FIG. 2, adjacent code data (A) and code data (C) are mixed in the m × n image area in addition to the target code data (B) for which the center of gravity is to be obtained.
[0014]
Here, as foresight information determined by an image format, at least one of each code data is included in the 5 × 7 matrix image area, and data of at least one pixel or more is provided between adjacent code data. Is given.
[0015]
FIG. 3 is a block diagram showing a first embodiment of the present invention. In FIG. 3, reference numeral 11 denotes a unit for cutting out and storing an m × n matrix image area including an object image from an input image signal in accordance with the foresight information provided by the foresight information holding unit 18. It is shown in FIG. In this embodiment, an m × n matrix image area has a 5 × 7 matrix configuration of 5 pixels in the m direction and 7 pixels in the n direction, and the center of the 5 × 7 matrix image area is a reference pixel A _ij. (In this example, A _ij = a ₃₄ ). Therefore, it is assumed that the look-ahead information holding means 18 holds the look-ahead information L1, L2, A _ij = a ₃₄ , m = 5, and n = 7.
[0016]
In FIG. 3, reference numeral 12 denotes first calculating means for performing Y projection on the object image in the m × n matrix image area, and reference numeral 13 denotes a second calculating means for obtaining a moment from the reference coordinates in the Y direction. Calculation means. Similarly, reference numeral 14 denotes a third calculating means for performing X projection on the object image in the mxn matrix image area, and 15 denotes a fourth calculating means for obtaining a moment from the reference coordinates in the X direction. It is an operation means. Reference numeral 16 denotes the correction amount in the Y direction of the center position of the object image from the reference coordinates included in the storage unit 11 of the mxn matrix image area, and the output of the first and second calculation units 12 and 13. Fifth calculating means for calculating from the result, reference numeral 17 denotes a correction amount in the X direction of the center position of the object image from the reference coordinates contained in the storage means 11 of the mxn matrix image area, and This is a sixth calculating means that calculates from the output results of the fourth calculating means 14 and 15.
[0017]
Next, a configuration example of the first calculating means 12 will be described with reference to FIG. The calculating means 12 includes a seventh calculating means 12-1 for calculating a sum (Y projection) of column vectors in the Y direction represented by the following equation (3), and a discontinuous point in the Y direction of the object image from the result. Calculating means 12-2 comprising a masking circuit for detecting a column pattern and masking a column vector element from a discontinuous point in the X direction, which will be described later, and the sum of m calculation results of the eighth calculating means 12-2 And ninth operation means 12-3 comprising an adder for obtaining (ΣYs = ΣYs [i]).
Ys [1-5] _{'= (a 11 + a 12} ··· + a 17, a 21 + a 22 ··· + a 27, ····· a 51 + a 52 ··· + a 57) ····· ( 3)
[0018]
FIG. 5 is a diagram showing a configuration example of the second calculating means 13. The second calculating means 13 calculates a weighting coefficient in the Y direction, and the weighting coefficient Cy [1-5] It is set around the reference pixel. In this embodiment, Cy [1] = − 2, Cy [2] = − 1, Cy [3] = 0, Cy [4] = 1, and Cy [5] = 2. Then, a tenth calculating means 13-1 for obtaining the calculation of the weighting coefficient is added to the sum of the row vectors of each row, and the sum of m calculation results of the tenth calculating means 13-1 (ΣYp = ΣYp [i]) ) Constitutes a second calculating means.
[0019]
Next, a configuration example of the third calculating means 14 will be described with reference to FIG. The calculating means 14 includes a twelfth calculating means 14-1 for calculating a sum (X projection) of row vectors in the X direction represented by the following equation (4), and calculating a discontinuous point in the X direction of the target image from the result. A thirteenth operation means 14-2 comprising a mask circuit for detecting and masking the elements of the row vector from the above-mentioned discontinuous point in the Y direction, and a sum of n operation results of the thirteenth operation means 14-2 ( (Xs = 演算 Xs [i])). .
Xs [1-7] _{'= (a 11 + a 21} ··· + a 51, a 12 + a 22 ··· + a 52, ····· a 17 + a 27 ··· + a 57) ···· (4 )
[0020]
FIG. 7 is a diagram showing a configuration example of the fourth calculating means 15. The fourth calculating means 15 calculates a weighting coefficient in the X direction. The weighting coefficient Cx [1-7] It is set around the pixel. In the present embodiment, Cx [1] =-3, Cx [2] =-2, Cx [3] =-1, Cx [4] = 0, Cx [5] = 1, Cx [6] = 2 Set Cx [7] = 3. Then, a fifteenth calculating means 15-1 for calculating the weighting coefficient is added to the sum of the vectors in each column, and a total sum (ΣXp = ΣXp [i]) of n calculation results of the fifteenth calculating means 15-1 ) To form a fourth calculating means 15.
[0021]
FIG. 8A is a diagram showing a configuration example of the fifth calculating means 16 for obtaining the correction amount ΔY of the position of the center of gravity from the reference pixel in the Y direction. The output result ΔYs of the first calculating means 12 is represented by the second And a divider 16-1 for dividing by the output result ΣYp of the calculating means 13. FIG. 8B is a diagram showing a configuration example of the sixth calculating means 17 for obtaining the correction amount ΔX of the position of the center of gravity from the reference pixel in the X direction, and the output result ΣXs of the third calculating means 14 is shown in FIG. 4 is composed of a divider 17-1 for dividing by the output result ΣXp of the operation means 15 of FIG.
[0022]
Next, a description will be given of a means for detecting a discontinuous point of the target object image data. In FIG. 2, a 5 × 7 matrix image area using A _ij as a reference pixel in the figure is cut out, and an image stored in the storage unit 11 is shown in FIG. Among these, the portion shown by the diagonal left is the target image (black image), and the portion shown by the diagonal right is the adjacent code data or the non-target image due to scratches or dirt. In the present embodiment, the detection of the discontinuous point of the image stored in the storage unit 11 is performed by detecting “0” based on the projection result. In this embodiment, no discontinuous point is detected in the Y projection as shown by the following equation (5).
Ys [1-5] '= (1,2,3,3,1) (5)
In the X projection, discontinuous points Xs [2] 'and Xs [6]' are detected as shown by the following equation (6).
Xs [1-7] = (1,0,3,3,2,0,1) (6)
[0023]
Therefore, as a result, the control signals for masking the image signals of the pixels (a ₁₁ , a ₂₁ , a ₃₁ , a ₄₁ , a ₅₁ ) and (a ₁₇ , a ₂₇ , a ₃₇ , a ₄₇ , a ₅₇ ) are changed. appear. That is, by detecting a portion where the result of the X projection and the Y projection is “0”, a discontinuous point of the object image is detected, and the mask circuit included in the first and third arithmetic means 12 and 14 is detected. The control signal is applied to the eighth and thirteenth operation means 12-2 and 14-2, which are composed of the following, and the operation results of the eighth and thirteenth operation means expressed by the following equations (7) and (8) are obtained. Can be
Ys [1-5] = (0,2,3,3,0) (7)
Xs [1-7] = (0,0,3,3,2,0,0,) (8)
In FIG. 9, Yp (= Cy × Ys) and Xp (= Cx × Xs) indicate calculation results of the second and fourth calculation means 13 and 15, respectively.
[0024]
With respect to this result, the correction amounts ΔY and ΔX from the reference pixel A _ij are obtained assuming that the sum of the calculation results of the weighting coefficients is ΣXp and ΣYp, and are expressed by the following equations (9) and (10).
ΔY = ΣYp / ΣYs = 1/8 = 0.125 (9)
ΔX = ΣXp / ΣXs = − ／ = − 0.125 (10)
By adding the correction amounts ΔY and ΔX to the coordinate value of the reference pixel A _ij , it is possible to calculate the accurate center coordinates (centroid) of the target image.
[0025]
Next, a second embodiment will be described with reference to FIG. The two-dimensional image to be input and the basic configuration in this embodiment are almost the same as those in the first embodiment, and the eighth and thirteenth operation means 12-2 and 14-2 each comprising a mask circuit shown in FIGS. Can be dealt with by slightly changing the processing of FIG. FIG. 10 shows a portion obtained by cutting out a 5 × 7 matrix image area using A _ij as a reference pixel in this embodiment. Among these, the portion shown by the diagonal left is the target image (black image), and the portion shown by the diagonal right is the adjacent code data or the non-target image due to scratches or dirt.
[0026]
In this embodiment, a discontinuous point Ys [2] ′ is detected in the Y projection as shown by the following equation (11).
Ys [1-5] '= (3,0,3,4,2) (11)
In the X projection, a discontinuous point Xs [6] 'is detected as shown by the following equation (12).
Xs [1-7] '= (1,1,3,3,2,0,2) (12)
Therefore, this result, the mask image signals of the pixels _{(a 11, a 12, a} 13, a 14, a 15, a 16, a 17) and _{(a 17, a 27, a} 37, a 47, a 57) A control signal is generated. That is, by detecting a portion where the result of the X projection and the Y projection is “0”, a discontinuous point of the object image is detected, and the mask circuit included in the first and third arithmetic means 12 and 14 is detected. The control signal is applied to the eighth and thirteenth operation means 12-2 and 14-2, which are composed of the following, and the operation results of the eighth and thirteenth operation means expressed by the following equations (13) and (14) are obtained. Can be
Ys [1-5] = (0,0,3,3,1) (13)
Xs [1-7] = (0,0,2,3,2,0,0,) (14)
[0027]
With respect to this result, the correction amounts ΔY and ΔX from the reference pixel A _ij are obtained assuming that the sum of the calculation results of the weighting coefficients is ΣXp and ΣYp, and are expressed by the following equations (15) and (16).
ΔY = ΣYp / ΣYs = 4/7 = 0.57 (15)
ΔX = ΣXp / ΣXs = 0/7 = 0 (16)
By adding the correction amounts ΔY and ΔX to the coordinate value of the reference pixel A _ij , it is possible to calculate the accurate center coordinates (centroid) of the target image.
[0028]
The above processing is to temporarily store the calculation result of the discontinuous point detecting means in the first and third calculating means, and to provide the eighth and thirteenth calculating means comprising mask circuits in the X and Y directions as a mask signal, As shown in FIG. 3, it can be realized by giving by crossing.
[0029]
By the way, in each of the above embodiments, the storage means of the m × n matrix image area has the reference pixel A _ij arranged at the center of the matrix. However, the reference pixel may be set anywhere on the matrix. I do not care. Further, in each of the above embodiments, the storage means stores m × n image signals in a matrix, and the case where m = 5, n = 7 has been described. It is necessary to include an object image to be calculated. For example, when m = 7 and n = 9, the configuration is almost the same as that of the present embodiment, and by changing the bit length of each arithmetic unit in FIG. It can be easily realized.
[0030]
As described above, the control circuit for reading data from the m × n matrix storage means and the two-dimensional image storage means input from the imaging means to the m × n matrix storage means is slightly changed in accordance with the size of the object image. By doing so, it is possible to easily calculate the center-of-gravity correction amount of the target object image from the reference pixel.
[0031]
【The invention's effect】
As described based on the above embodiments, according to the first aspect of the invention, along with the feature amount of an object image can be and accurately calculated at high speed, the image data other than the object image data Can be realized with a simple configuration. According to the second aspect of the present invention, a means for processing image data other than the target object image data as unnecessary data can be realized with a simple configuration. According to the third aspect of the present invention, it is possible to quickly and accurately calculate the center of gravity of a plurality of target object images. According to the fourth aspect of the present invention, a masking means for processing image data other than target object image data as unnecessary data can be realized with a simpler configuration.
[Brief description of the drawings]
FIG. 1 is a diagram showing an example of an image format targeted by an image processing apparatus according to the present invention.
FIG. 2 is an enlarged view showing a part of the image format example shown in FIG. 1;
FIG. 3 is a block diagram showing a first embodiment of the present invention.
FIG. 4 is a diagram showing a configuration example of a first calculating means in the first embodiment shown in FIG. 3;
FIG. 5 is a diagram showing a configuration example of a second calculating means in the first embodiment shown in FIG. 3;
FIG. 6 is a diagram showing a configuration example of a third calculating means in the first embodiment shown in FIG. 3;
FIG. 7 is a diagram showing a configuration example of a fourth calculating means in the first embodiment shown in FIG. 3;
FIG. 8 is a diagram showing a configuration example of fifth and sixth calculation means in the first embodiment shown in FIG. 3;
FIG. 9 is an explanatory diagram for explaining the operation of the first embodiment shown in FIG. 3;
FIG. 10 is a diagram illustrating a second embodiment of the present invention and its operation.
FIG. 11 is an explanatory diagram for explaining a conventional center-of-gravity point detection method.
FIG. 12 is an explanatory diagram for explaining a conventional center-of-gravity point detection method.
[Explanation of symbols]
1A, 1B Reference image 2 Code data group A, C Adjacent code data B Target code data 11 Storage means 12 First calculation means 13 Second calculation means 14 Third calculation means 15 Fourth calculation means 16 Fifth Arithmetic means 17 Sixth arithmetic means 18 Foresight information holding means 12-1 Seventh arithmetic means 12-2 Eighth arithmetic means 12-3 Ninth arithmetic means 13-1 Tenth arithmetic means 13-2 Eleventh Calculation means 14-1 twelfth calculation means 14-2 thirteenth calculation means 14-3 fourteenth calculation means 15-1 fifteenth calculation means 15-2 sixteenth calculation means 16-1, 16-2 Divider

Claims (4)

Means for two-dimensionally storing an m × n image area including an object image whose feature amount in an image input by the imaging means is to be calculated, and which is included in the storage means based on preset look-ahead information Means for detecting and determining the presence of image data other than the object image data to be detected, and means for processing image data other than the object image data as unnecessary data according to the result of the detection / determination means. In an image processing apparatus configured to perform a process of calculating a feature amount of the target image after extracting the target image data, the unit that detects and determines the presence of image data other than the target image data includes: Means for performing X-projection and Y-projection of image data included in an m × n image area including an image, and inspecting the X-projection and Y-projection outward from a reference pixel of the m × n image area; De It regarded coordinate axes data does not exist a discontinuity, an image processing apparatus characterized by being composed of the image data and determining means other than the object image data outside the image data from the coordinate axes. Means for processing the image data other than the object image data as unnecessary data, according to claim 1, characterized in that it is constituted by a means for masking the image data other than the object image data determined by said determining means An image processing apparatus as described in the above. The process of calculating the feature quantity of the object image, the image processing apparatus according to claim 1 or 2, wherein the a centroid calculation process of the object image. 2. The image processing apparatus according to claim 1, wherein the masking unit masks image data other than the target object image data according to the other projection based on a determination result of the determining unit regarding one of the X projection and the Y projection. 2. The image processing device according to 2.

Images