TWI485634B - System and method for identifying difference of two images - Google Patents

Description

Differential image automatic recognition system and method

The present invention relates to an image processing system and method, and more particularly to a system and method for automatically identifying a difference image.

When shooting an image of an object under test (such as a main board) under normal external light source illumination, the difference between the captured image and the actual image may be caused by the change in the brightness of the light caused by the change of the light source from one fluorescent lamp to two lamps. Therefore, the industry needs to use optical automatic detection (AOI) equipment to automatically detect the captured image of the object under test to ensure the change of brightness and the surrounding light and shadow in the natural light shooting, to achieve the fixed light source shooting object. The effect of the image.

In optical automatic detection, it is common to obtain an image of the object to be measured and compare it with the sample image. In practical applications of optical automatic detection, it is usually required that the production line AOI test engineer inputs a number of threshold operators (Threshold) to automatically identify the bad position of the image to be tested. However, when detecting whether a motherboard image is the same as the sample image, you need to manually input all the thresholds of the motherboard image, which is about 3,000 to 4,000. Therefore, manually inputting the difference threshold operator of the target image and the sample image to complete the image comparison, the accuracy is not high, and the labor cost is high.

In view of the above, it is necessary to provide a differential image automatic recognition system and method, which can automatically generate a threshold operator required for comparing two images, and identify the two images by comparing the image of the object to be measured with the actual image. The difference image.

The differential image automatic identification system is installed and operated in a computer, and the computer is connected to the image capturing device. The system includes: an image acquisition module, configured to capture a captured image of the object to be tested by the image capturing device, and obtain a standard image of the object to be tested from a memory of the computer; and an image computing module for capturing the image according to the image Resolution and primitive RGB Calculate the absolute value of the IAED average difference between the captured image and the standard image on the R channel, the absolute value of the IAED average difference on the G channel and the absolute value of the IAED average difference on the B channel, and segment the captured image. N N primitive blocks and dividing the standard image into N primitive blocks, and calculating the difference value of each primitive block in the captured image corresponding to the IAED of the primitive block in the standard image on the RGB three channels, wherein IAED represents Image average energy density; threshold generation module for generating a threshold operator required for comparing captured images with standard images; image comparison module for each primitive block in the captured image The absolute value of the IAED difference value of the corresponding primitive block in the standard image on the RGB three-channel, the absolute value of the average difference between the captured image and the standard image on the RGB three-channel IAED, and the threshold value generated to determine the captured image Whether each primitive block is the same as the corresponding primitive block in the standard image; the image generation module is configured to clear the image from the captured image when the primitive block of the captured image is the same as the corresponding primitive block in the standard image The metablock generates a difference image according to the remaining primitive blocks in the captured image, and outputs the difference image to a display on the computer.

The differential image automatic identification method identifies the difference primitive block in the captured image of the object to be detected by the computer, and the computer is connected to the image capturing device. The method includes the steps of: taking a captured image of the object to be tested by the image capturing device, and obtaining a standard image of the object to be tested from the memory of the computer; calculating the captured image and the standard image according to the resolution of the captured image and the primitive RGB The absolute value of the IAED average difference on the R channel, the absolute value of the IAED average difference on the G channel and the absolute value of the IAED average difference on the B channel, where IAED represents the average energy density of the image; the captured image is segmented into N primitives Blocking and dividing the standard image into N primitive blocks; calculating the difference value of each primitive block in the captured image corresponding to the IAED of the primitive block in the standard image on the RGB three channels; generating a comparison image and standard image Threshold operator required for the time; according to the image element in each image block and the corresponding image block in the standard image, the absolute value of the IAED difference value on the RGB three channels, the captured image and the standard image on the RGB three-channel IAED The absolute value of the average difference value and the generated threshold operator determine whether each primitive block of the captured image is identical to the corresponding primitive block in the standard image; if the image element block and the standard image are captured If the corresponding primitive block is the same, the primitive block is cleared from the captured image; if the primitive block of the captured image is different from the corresponding primitive block in the standard image, the primitive block is retained in the captured image; The remaining primitive block in the image produces a difference image and outputs the difference image to a display on a computer.

Compared with the prior art, the differential image automatic identification system and method of the present invention can automatically generate a threshold operator required for comparing two images, and identify the image by comparing the image of the object to be measured with the actual image. The difference image between the two images can omit the process of manually creating the threshold operator, thereby improving accuracy and saving labor costs.

As shown in FIG. 1, it is an architectural diagram of a preferred embodiment of the differential image automatic recognition system 11 of the present invention. In the embodiment, the differential image automatic identification system 11 is installed and operated in the computer 1, and can automatically generate a threshold operator (Threshold) required to compare two images, and identify the image and actual object to be tested. The image difference primitive block, and the difference image of the two images is generated according to the difference primitive block. The computer 1 is connected to an image capturing device 2 for taking a captured image of the object to be tested 3, as shown in FIG. When the image of the object to be tested 3 (for example, the main board) is photographed under the condition that the external light source is irradiated, the light source may change from one fluorescent lamp to two lamps, and the brightness of the light is changed to cause a difference between the captured image and the actual image.

The computer 1 includes a memory 12, a central processing unit 13, and a display 14. The memory 12 stores a standard image that is compared with the captured image, such as the standard image b shown in FIG. The central processing unit 13 is configured to execute the difference image automatic recognition system 11 to automatically recognize the difference image of the captured image by comparing the difference primitive block between the captured image and the standard image, and display the difference image on the display 14.

The difference image automatic recognition system 11 includes an image acquisition module 111, an image calculation module 112, an image comparison module 113, a threshold generation module 114, and an image generation module 115. The module referred to in the present invention is a computer program segment composed of a series of calculation instructions. In the embodiment, the module is a computer program segment that can be executed by the central processing unit 13 and can perform a fixed function, and is stored in the memory 12.

The image capturing module 111 is configured to capture the captured image a of the object to be tested 3 by the image capturing device 2 and obtain the standard image b of the object to be tested 3 from the memory 12 . The captured image a is an image to be tested of the object to be tested 3 when the light source is disturbed by the change of the external light source. The standard image b is a sample image of the object 3 to be tested when there is no external light interference.

The image calculation module 112 is configured to calculate an Image Average Energy Density (IAED) value on the RGB three channels of the captured image a and the standard image b, respectively. The RGB three channels include an R grayscale channel, a G grayscale channel, and a B grayscale channel of an image. The image average energy density IAED refers to the average energy density of each primitive in the N×N image, and is used to measure the reference energy of the image element according to the formula IAED=(R+G+B). /N/N to calculate the average energy density of the primitive of the image. In this formula, R represents the sum of the R values of all the primitives in the image, and G represents the sum of the G values of all the primitives in the image. And B represents the sum of the B values of all the primitives in the image.

In this embodiment, it is assumed that the picture element of the captured image a is 32×32, and the sum of the luminance values of the RGB three channels of the primitive block (for example, A1) to be processed is R=147248, G=147760, B=144176, Then, the average IAED value of the primitive block A1 = (R + G + B) / 32 / 32 = 428.89. The image calculation module 112 calculates the R channel IAED_a_R=R/32/32=147248/32/32=143.80 of the captured image a, G channel IAED_a_G=G/32/32=144760/32/32=141.37, B channel IAED_a_B= B/32/32=144176/32/32=140.80. Assume that the primitive of the standard image b is 32×32, wherein the sum of the luminance values of the three channels of RGB is R=152179, G=135539, B=31091, and the image calculation module 112 calculates the R channel IAED_b_R=R/32 of the standard image b. /32=152179/32/32=148.61, G channel IAED_b_G=G/32/32=135539/32/32 = 132.36, B channel IAED_b_B value = 31091/32/32 = 30.36.

The image calculation module 112 is further configured to calculate an absolute value of the average difference D_R of the IAED on the R channel of the captured image a and the standard image b, an absolute value of the average difference D_G of the IAED on the G channel, and the B channel. The absolute value of the average difference D_B of IAED. In this embodiment, the average difference D_R of the IAED on the R channel of the captured image a and the standard image b is equal to the IAED_a_R value minus the absolute value of the IAED_b_R value, and the average difference D_G of the IAED on the G channel is equal to the IAED_a_G value minus the IAED_b_G value. The absolute value, and the average difference D_B of the IAED on the B channel is equal to the IAED_a_B value minus the absolute value of the IAED_b_B value. For example, the average difference D_R=IAED_a_R-IAED_b_R=143.80-148.61=-4.81, the absolute value of rounding is 5; the average difference D_G=IAED_a_G-IAED_b_G=141.37-132.36=9.01, and the absolute value of rounding is 9; The difference D_B=IAED_a_B-IAED_b_B=140.80–30.36=110.44, and the absolute value of rounding is 110.

The image calculation module 112 is configured to divide the captured image a into N primitive blocks, such as the primitive blocks A1, A2, ... An shown in FIG. 4, and divide the standard image b into N maps. The metablock, such as the primitive blocks B1, B2, ... Bn shown in FIG. The image calculation module 112 is further configured to calculate, respectively, a difference value of each primitive block An in the captured image a corresponding to the IAED of each primitive block Bn in the standard image b on the RGB three channels. In this embodiment, the image computing module 112 establishes a primitive block index n_Index for each primitive block, and makes n_Index=1, and starts to calculate a difference value D_Rn between the primitive block An and the primitive block Bn on the R channel IAED. The absolute value, the absolute value of the difference value D_Gn of the IAED on the G channel, and the absolute value of the difference value D_Bn of the IAED on the B channel. In this embodiment, the calculation method of the difference values D_Rn, D_Gn, and D_Bn of each IAED of each primitive block Bn corresponding to each primitive block Bn on the RGB three-channel is compared with the above-mentioned calculation of the captured image a and the standard image b in RGB three. The average difference between the IAED of the IAED on the channel is D_R, D_G, and D_B, and the values of R, G, and B taken here are limited to the sum of all the points RGB on each primitive block.

The image comparison module 113 is configured to use the difference value of the IAED on each of the RGB three channels according to each primitive block An and the corresponding primitive block Bn, the threshold value of the image comparison, and the captured image a and the standard image. b Determine whether the image data in the primitive block An and the primitive block Bn are the same by the average difference of the IAED on the RGB three channels. In this embodiment, the image comparison module 113 determines whether the absolute value of the difference value D_Rn is greater than the product of the absolute value of the threshold operator and the average difference value D_R, and whether the absolute value of the difference value D_Gn is greater than the threshold operator and the average value. The product of the absolute value of the difference D_G, and the absolute value of the difference value D_Bn is greater than the product of the absolute value of the threshold operator and the average difference value D_B. That is, the image matching module 113 determines whether two or more inequalities in the inequality D_Rn>T×D_R, the inequality D_Gn>T×D_G, and the inequality D_Gn>T×D_G are satisfied, where T is a threshold operator.

The threshold generating module 114 is configured to automatically generate a threshold operator T required for comparing the captured image a with the standard image b. The method for generating the threshold operator T will be described in FIG.

The image generation module 115 is configured to clear the primitive block An from the captured image a when the image data in the primitive block An and the primitive block Bn are the same. When the image block An and the image data in the primitive block Bn are different, the primitive block An is retained in the captured image a. After all the primitive blocks An in the captured image a are compared with the primitive block Bn, the image generation module 115 generates a difference image according to all the retained primitive blocks An, and outputs the difference image to the display 14. Displaying, or storing the difference image in the memory 12.

As shown in FIG. 2, it is a flowchart of a preferred embodiment of the differential image automatic identification method of the present invention. In this embodiment, the method can automatically generate a threshold operator required for two image comparisons, identify the difference primitive block by comparing the measured image with the actual image, and generate two pieces according to the difference primitive block. The difference image of the image.

In step S20, the image capturing module 111 captures the captured image a of the object to be tested 3 by the image capturing device 2, and acquires the standard image b of the object to be tested 3 from the memory 12. The captured image a is an image to be tested of the object to be tested 3 when the light source is disturbed by the change of the external light source. The standard image b is a sample image of the object 3 to be tested when there is no external light interference.

In step S21, the image calculation module 112 calculates the IAED values on the RGB three channels of the captured image a and the standard image b, and calculates the absolute value of the average difference D_R of the IAED values of the captured image a and the standard image b on the R channel, respectively. The absolute value of the average difference D_G of the IAED value on the G channel, and the absolute value of the average difference D_B of the IAED value on the B channel. The IAED refers to the average energy density of each primitive in the resolution of the N×N image, and is used to measure the reference energy of the image element of the image according to the formula IAED=(R+G+B)/N/N. To calculate the average energy density of the image of the image.

In step S22, the image computing module 112 divides the captured image a into N primitive blocks, such as the primitive blocks A1, A2, ... An shown in FIG. 4, and divides the standard image b into N primitive blocks. , as shown in Figure 4, the primitive blocks B1, B2, ... Bn.

In step S23, the image computing module 112 creates a primitive block index n_Index for each primitive block An and each primitive block Bn, and makes n_Index=1.

In step S24, the image computing module 112 calculates the absolute value of the difference value D_Rn between the primitive block An and the primitive block Bn on the R channel, and the absolute value of the difference value D_Gn of the IAED on the G channel, and on the B channel. The absolute value of the difference value D_Bn of IAED.

In step S25, the image comparison module 113 is based on the difference value between the primitive block An and the primitive block Bn on the RGB three channels, the threshold value of the image comparison, and the captured image a and the standard image b on the RGB three channels. The average difference of the IAED is used to determine whether the image data in the primitive block An and the primitive block Bn are the same. In this embodiment, the image matching module 113 determines whether two or more of the inequality D_Rn>T×D_R, the inequality D_Gn>T×D_G, and the inequality D_Gn>T×D_G are satisfied, where T is a threshold operator, The method of generating the threshold operator T will be described in FIG.

If the image data in the primitive block An and the primitive block Bn are different, the image generation module 115 retains the primitive block An in the captured image a in step S26. If the image data in the primitive block An and the primitive block Bn are the same, in step S27, the image generation module 115 clears the primitive block An from the captured image a.

In step S28, the image comparison module 113 determines whether the value of the primitive index n_Index is smaller than the number N of the primitive blocks of the captured image a, that is, whether each primitive block An in the captured image a has been compared with the primitive block Bn. . If there is still no comparison between the primitive block An and the primitive block Bn, step S29 is performed; if all the primitive blocks An and the primitive block Bn have been compared, step S30 is performed.

In step S29, the image matching module 113 performs the n_Index=n_Index+1 operation on the primitive index n_Index value, and then executes step S24. In step S30, the image generation module 115 calculates the total number D_n of the primitive blocks An retained in the captured image a, generates a difference image according to all the reserved primitive blocks An, and outputs the difference image to the display 14 for display. Or storing the difference image in the memory 12.

As shown in FIG. 3, it is a flowchart of the threshold value generating method required in step S25 in FIG. In step S31, the threshold value generating module 114 sets the threshold value Tn, sets the minimum threshold x and the maximum threshold y of Tn, and divides Tn by a predetermined step in the range of the minimum threshold x and the maximum threshold y of Tn. It is a sequence of thresholds with T thresholds (Tn1, Tn2, ..., Tni, ..., Tns). In the present embodiment, the minimum threshold x is set to 2.0, the maximum threshold y is set to 30, and the predetermined step is set to 0.1.

In step S32, the threshold value generating module 114 calculates, by using each of the threshold values Tni of the threshold sequence, the number of difference primitive blocks D_mi of the captured image a divided into M primitive blocks, and constructs a sequence of first difference primitive values. (Pm1, Pm2, ... Pmi, ..., Pms), where Pmi = D_mi × M × M, i = 1, 2, ..., S.

In step S33, the threshold value generating module 114 calculates, by using each of the threshold values Tni of the threshold sequence, the number of difference primitive blocks D_ni of the captured image a divided into N primitive blocks, and constructs a sequence of second difference primitive values. (Pn1, Pn2, ... Pni, ..., Pns), where Pni = D_ni × N × N, i = 1, 2, ..., S.

Step S34, the threshold generation module 114 gradually compares each of the first difference primitive value sequence (Pm1, Pm2, ... Pmi, ..., Pms) with the second difference primitive value sequence (Pn1, Pn2, ... Each element of Pni,..., Pns) finds the value of i corresponding to the minimum value of the difference primitive D_p=Pmi-Pni.

In step S35, the threshold value generation module 114 takes the threshold value Tni corresponding to the i value in the threshold sequence (Tn1, Tn2, ..., Tni, ..., Tns) as the optimal threshold T. If the threshold generation module 114 finds that the difference value of the difference primitive D_p corresponds to 10, the Tn10=2.0+0.1×10=3.0 in the threshold sequence (Tn1, Tn2, ...Tns), that is, the valve The value generation module 114 produces an optimal threshold T of 3.0.

The above is only the preferred embodiment of the present invention, and has been used in a wide range of applications. Any other equivalent changes or modifications that are not departing from the spirit of the present invention should be included in the following patent application. Within the scope.

1. . . computer

11. . . Differential image automatic recognition system

111. . . Image acquisition module

112. . . Image computing module

113. . . Image comparison module

114. . . Threshold generation module

115. . . Image generation module

12. . . Memory

13. . . CPU

14. . . monitor

2. . . Image capture device

3. . . Object to be tested

1 is a block diagram of a preferred embodiment of the differential image automatic recognition system of the present invention.

2 is a flow chart of a preferred embodiment of the differential image automatic recognition method of the present invention.

FIG. 3 is a flow chart of a method for generating a threshold operator required in step S25 of FIG. 2.

FIG. 4 is a schematic diagram of a captured image and a standard image divided into N primitive blocks.

1. . . computer

11. . . Differential image automatic recognition system

111. . . Image acquisition module

112. . . Image computing module

113. . . Image comparison module

114. . . Threshold generation module

115. . . Image generation module

12. . . Memory

13. . . CPU

14. . . monitor

2. . . Image capture device

3. . . Object to be tested

Claims (10)

A differential image automatic identification system installed and operated in a computer connected to an image capturing device, the system comprising:
The image acquisition module is configured to capture a captured image of the object to be tested by the image capturing device, and obtain a standard image of the object to be tested from the memory of the computer;
The image calculation module is configured to calculate the absolute value of the average difference of the IAED between the captured image and the standard image on the R channel according to the resolution of the captured image and the primitive RGB, the absolute value of the average difference of the IAED on the G channel, and the IAED on the B channel. The absolute value of the average difference is divided into N primitive blocks and the standard image is divided into N primitive blocks, and each primitive block in the captured image is calculated corresponding to the primitive block in the standard image in RGB. The difference value of IAED on the three channels, where IAED represents the average energy density of the image;
Threshold generation module for generating a threshold operator required for comparing images and standard images;
The image comparison module is configured to determine the absolute value of the IAED difference value on the RGB three channels according to the primitive block in the captured image and the corresponding image block in the standard image, and the IAED average of the captured image and the standard image on the RGB three channels. The absolute value of the difference and the generated threshold operator determine whether each primitive block of the captured image is identical to the corresponding primitive block in the standard image;
The image generation module is configured to: when the primitive block of the captured image is the same as the corresponding primitive block in the standard image, clear the primitive block from the captured image, and generate a difference image according to the remaining primitive block in the captured image. And output the difference image to the display on the computer. The differential image automatic identification system according to claim 1, wherein the IAED refers to an average energy density of each primitive in the N×N image, which is according to the formula IAED=(R+G). +B)/N/N to calculate the average energy density of the image of the image. The difference image automatic identification system of claim 1, wherein the image calculation module is further configured to create a primitive block for each primitive block in the captured image and each primitive block in the standard image. index. The difference image automatic identification system of claim 1, wherein the image comparison module determines whether each primitive block of the captured image is identical to the corresponding primitive block in the standard image, and includes the steps of:
Determining whether the absolute value of the IAED difference value between the primitive block in the captured image and the corresponding primitive block in the standard image on the R channel is greater than the absolute difference between the threshold operator and the average difference of the IAED between the captured image and the standard image on the R channel Product of values;
Determining whether the absolute value of the IAED difference value of each primitive block in the captured image and the corresponding primitive block in the standard image on the G channel is greater than the absolute difference between the threshold operator and the average difference of the IAED between the captured image and the standard image on the G channel. Product of values;
Determining whether the absolute value of the IAED difference value of each primitive block in the captured image and the corresponding primitive block in the standard image on the G channel is greater than the absolute difference between the threshold operator and the average difference of the IAED between the captured image and the standard image on the G channel. Product of values;
If two or more of the above three determination conditions are satisfied, each primitive block of the captured image is the same as the corresponding primitive block in the standard image. The differential image automatic identification system of claim 1, wherein the threshold generating module generates a threshold operator comprises the following steps:
Constructing a sequence of thresholds having S thresholds;
Calculating, by using each threshold value in the threshold sequence, a number of difference first meta-blocks of the captured image that is divided into M primitive blocks, and constructing a first difference primitive value sequence according to the first difference primitive block number;
Calculating, by using each threshold value in the threshold sequence, the number of second difference primitive blocks in which the captured image is divided into N primitive blocks, and constructing a second differential primitive value sequence according to the second difference primitive block number;
Gradually comparing each element in the first difference primitive value sequence with each element in the second difference primitive value sequence to find a difference primitive block number i value corresponding to the minimum value of the difference primitive, where i=1, 2, ..., S;
The threshold value corresponding to the value of i in the threshold sequence is taken as the optimal threshold operator. A method for automatically identifying a difference image, wherein a computer detects an image of a difference element in a captured image of the object to be detected, and the computer is connected with an image capturing device, the method comprising the steps of:
Taking a captured image of the object to be tested by the image capturing device, and obtaining a standard image of the object to be tested from the memory of the computer;
Calculate the absolute value of the average IAED difference between the captured image and the standard image on the R channel according to the resolution of the captured image and the primitive RGB, the absolute value of the average difference of the IAED on the G channel, and the absolute value of the average difference of the IAED on the B channel. Where IAED represents the average energy density of the image;
Dividing the captured image into N primitive blocks and dividing the standard image into N primitive blocks;
Calculating the difference value of each primitive block in the captured image corresponding to the IAED of the primitive block in the standard image on the RGB three channels;
Produce a threshold operator required to compare images and standard images;
According to the absolute value of the IAED difference value on the RGB three channels of each primitive block in the captured image and the corresponding primitive block in the standard image, the absolute value of the IAED average difference between the captured image and the standard image on the RGB three channels and the generated value The threshold operator determines whether each primitive block of the captured image is identical to the corresponding primitive block in the standard image;
If the primitive block of the captured image is the same as the corresponding primitive block in the standard image, the primitive block is cleared from the captured image;
If the primitive block of the captured image is different from the corresponding primitive block in the standard image, the primitive block is retained in the captured image;
A difference image is generated according to the remaining primitive blocks in the captured image, and the difference image is output to a display on a computer. The method for automatically identifying a difference image according to claim 6, wherein the IAED refers to an average energy density of each primitive in the N×N image, which is according to the formula IAED=(R+G). +B)/N/N to calculate the average energy density of the image of the image. The method for automatically identifying a difference image according to claim 6 of the patent application scope, further comprising the steps of establishing a primitive block index for each primitive block in the captured image and each primitive block in the standard image. The method for automatically distinguishing the difference image according to the sixth aspect of the invention, wherein the step of determining whether each primitive block of the captured image is identical to the corresponding primitive block in the standard image comprises:
Determining whether the absolute value of the IAED difference value between the primitive block in the captured image and the corresponding primitive block in the standard image on the R channel is greater than the absolute difference between the threshold operator and the average difference of the IAED between the captured image and the standard image on the R channel Product of values;
Determining whether the absolute value of the IAED difference value of each primitive block in the captured image and the corresponding primitive block in the standard image on the G channel is greater than the absolute difference between the threshold operator and the average difference of the IAED between the captured image and the standard image on the G channel. Product of values;
Determining whether the absolute value of the IAED difference value of each primitive block in the captured image and the corresponding primitive block in the standard image on the G channel is greater than the absolute difference between the threshold operator and the average difference of the IAED between the captured image and the standard image on the G channel. Product of values;
If two or more of the above three determination conditions are satisfied, each primitive block of the captured image is the same as the corresponding primitive block in the standard image. The method for automatically identifying a difference image according to claim 6, wherein the step of generating a threshold operator comprises:
Constructing a sequence of thresholds having S thresholds;
Calculating, by using each threshold value in the threshold sequence, a number of difference first meta-blocks of the captured image that is divided into M primitive blocks, and constructing a first difference primitive value sequence according to the first difference primitive block number;
Calculating, by using each threshold value in the threshold sequence, the number of second difference primitive blocks in which the captured image is divided into N primitive blocks, and constructing a second differential primitive value sequence according to the second difference primitive block number;
Gradually comparing each element in the first difference primitive value sequence with each element in the second difference primitive value sequence to find a difference primitive block number i value corresponding to the minimum value of the difference primitive, where i=1, 2, ..., S;
The threshold value corresponding to the value of i in the threshold sequence is taken as the optimal threshold operator.