CN1110010C - Automatic Detection Method of Disconnection Phenomenon of Optical Scanner - Google Patents

Abstract

An automatic detection method for the disconnection phenomenon of optical scanner can automatically detect the disconnection phenomenon and calculate the number of disconnection. After reading the oblique line of the test chart, calculating an intermediate gray-scale value, setting a standard deviation, searching a pixel with the gray-scale value closest to the intermediate gray-scale value in the first row of pixels of the oblique line, and starting to compare the gray-scale values by taking the pixel (Xi, Yj) as a reference. The comparison sequence is to compare the gray levels of the pixels located at (Xi, Yj) and (Xi-1, Yj + 1). If the comparison value exceeds a set value, the phenomenon of line drop is indicated. The number of dropped connections can be accumulated during the comparison process, and the place where the line was dropped can be framed.

Description

Automatic Detection Method of Disconnection Phenomenon of Optical Scanner

The invention relates to a method for detecting image quality, in particular to a method for detecting a disconnection phenomenon applied to an optical scanner.

There are many factors that affect the scanned image quality, and different test items can be designed for various possible factors. In the common image quality test, one of the test items is to detect the phenomenon of dropped calls. There are many factors that cause the disconnection phenomenon. For example, when the scanner is moving the optical module, the reading of the image is intermittent due to the out-of-step of the motor or vibration and friction. Or the control of the scanner's transmission system is unstable, or the speed of the transmission system and image reading is inconsistent, etc., all of which may cause disconnection. Therefore, the problem that may occur inside the scanner can be deduced from the phenomenon of disconnection. For example, if the number of strips dropped is the same every time, it may indicate that there may be a problem with the control software. If the dropped wires are at the same location, there may be a problem with the motor or the transmission system.

The test method can use a calibration paper with standard oblique lines drawn on it. After scanning, compare the scanned image with the standard oblique lines to determine the location and number of dropped lines. After the oblique line is zoomed in, the arrangement of pixels on the edge of the oblique line will be in a ladder shape. From the arrangement rules of the ladder, it can be seen whether there is any line drop. If one line is dropped, the step difference of the pixels is two pixels, if two lines are dropped, the step difference of the graph is three pixels, etc., and so on. If the dropped calls are within the tolerable range, then the scanner is of acceptable quality. Otherwise, it is a non-standard product and needs further inspection.

The method for judging dropped calls that is used now must go through very complicated procedures. When scanning a high-resolution image, the scanned image must be magnified dozens of times, and then manually compare the steps of the pixels on the diagonal line one by one, and then compare the step rules of the pixels one by one to find out the missing lines. place. Also, each image must be scanned back and forth several times to determine whether the scanner's controls are aligned. In addition, manual judgment has many problems, not only slow, but also involves subjective judgment, not only has a high error rate, but also its quality standards may vary from time to time, so it cannot provide consistent and effective information on the quality of scanned images. ensure.

The main purpose of the present invention is to provide a method for automatically detecting disconnection, to correctly and quickly judge whether the scanner has a disconnection phenomenon from the scanned image, and to display the place of disconnection and calculate the number of disconnection .

Another object of the present invention is to propose an automatic detection method for offline detection of scanned images applicable to various specifications, so as to replace manual detection and improve detection efficiency.

Another object of the present invention is to provide a method for automatically detecting disconnection phenomena, so as to judge internal problems generated by the disconnection phenomenon, thereby ensuring product quality.

The present invention proposes an automatic detection method for disconnection, which includes the following steps: after reading the oblique line of the test chart, calculate an intermediate gray scale value, and set a standard deviation. , look for a pixel whose grayscale value is closest to the intermediate grayscale value, and use the pixel (Xi, Yj) as a reference to start comparing grayscale values. The comparison sequence is first to compare the gray levels of the pixels located at (Xi, Yj) and (Xi-1, Yj+1). If the comparison value exceeds a set value, it means that there is a phenomenon of disconnection. But in order to calculate how many lines are lost, it must be compared with the pixel (Xi-2, Yj+1) adjacent to (Xi-1, Yj+1) until it meets a set comparison condition, that is , the gray scale of (Xi, Yj) is smaller than (Xi-1, Yj+1), and the gray scale of (Xi, Yj) is greater than or equal to (Xi-2, Yj+1). If the grayscale difference between the pixels of (Xi, Yj) and (Xi-1, Yj+1) is within the error tolerance, it means that there is no disconnection, and the comparison of the next column can be continued. During the comparison process, the number of disconnected lines is accumulated, and the places where the lines are dropped are framed. When the comparison is completed, the total number of dropped lines is output and the places where the lines are dropped are displayed.

The automatic detection method of the present invention is applicable to the image quality judgment of any scanner, not only can effectively replace manual judgment, but also can reduce errors and improve detection efficiency. Moreover, the problematic places can be displayed on the screen for further detection and judgment, making the detection more convenient. Moreover, the setting of the standard deviation can be adjusted according to scanners of different qualities, making it more widely applicable.

Fig. 1 is the schematic diagram of detection method of the present invention;

Fig. 2 is the judgment flowchart of the present invention for the detection of dropped line;

FIG. 3 is an enlarged schematic diagram of oblique lines in a case where two lines are dropped.

If you zoom in on the oblique line until you can see the shape of the pixels, you can see that the pixels on the edge of the normal oblique line are arranged very regularly, and their gray scales are also very similar. The grayscale change should be that the middle part is the darkest, and then gradually lighten to both sides. If the center of the line is used as the boundary, the arrangement and grayscale values on both sides are symmetrical. If there is a disconnection situation, in addition to the change of the gray scale, the step graphic of the pixel will also show the situation of the drop.

Aiming at the characteristics of the gray scale change of the oblique line, the present invention sets an oblique line with a slope of 45 degrees on the test chart. Certainly, slopes with other slopes are also available, but since the slope of 45 degrees is half of 90 degrees, the step difference of displayed pixels is the most obvious, so the resulting effect is the best. As shown in FIG. 1 , after the oblique line 11 with a slope of 45 degrees is enlarged, a ladder-like arrangement will be produced on the edge. In the case of not dropping the line, the step difference of each layer should be one pixel.

In the automatic judgment of the present invention, the number of dropped calls is not judged by the drop of the steps, but judged by the change of the gray scale value. When judging, a middle gray scale value and a standard deviation are used as standard values for judging. Taking the gray scale generated by 8 bits as an example, from completely black 0 to completely white 255, 256 gray scale changes can be produced. The method of the present invention takes its median value (128 ± standard deviation) as the standard value for comparison. This is because the grayscale changes from 128-255 and 128-0 are relatively obvious, because it is easier to judge the difference in grayscale values. The standard deviation can be set according to the specifications of the scanner to be tested.

As shown in FIG. 1 , a small grid in FIG. 1 represents a pixel, and the number in the grid is the grayscale value of the pixel. The gray scale in the middle part is completely black, and the gray scales on both sides are symmetrical (indicated by the gray scale value). The middle gray scale value of 256 is 128, assuming that the standard deviation value is 20, the gray scale value of (128±20) is taken as the standard value for comparison. In FIG. 1 , the abscissa is set to X and the ordinate is set to Y. Assume that index j represents the pixel at position j on the ordinate, and index i represents the pixel at position i on the abscissa. Let j=1; find the gray-scale value closest to (128±20) in the Yj column. And take the position of the found pixel on the abscissa X as the starting position. Taking the grayscale value in Figure 1 as an example, find the pixel closest to the middle grayscale value in row Yj to be 112. Then, the position of the pixel with the grayscale value of 112 on the abscissa X is used as the starting position of Xi, and i is set as 1.

Next, compare the gray scale difference between the vertical seat bar Yj and Yj+1 row, that is, compare the pixel located on the coordinate (Xi, Yj) and the pixel located on (Xi-1, Yj+1), if (Xi , Yj) and the gray scale value of the pixel on (Xi-1, Yj+1) are within the error tolerance value, it means that the connection is not dropped, and you can continue to check the next column, that is, compare Yj The pixels of the +1 row and the Yj+2 row. As shown in Figure 1, the pixel located on (Xi-1, Yj+1) has a grayscale value of 112, so it means that the connection is not disconnected, and the index j can be increased by 1, and the index i can be decreased by 1 to continue checking Comparison for the next column. In this way, pixels with grayscale values of 121, 123, 129, and 120 are compared one by one. Since the grayscale values are all within the error tolerance value, it can be determined that the oblique line is not dropped.

If the grayscale difference between the pixel on (Xi, Yj) and the pixel on (Xi-1, Yj+1) is not within the error tolerance, it means that there is a disconnection phenomenon. It is necessary to continue to compare the pixels on (Xi, Yj), (Xi-1, Yj+1) and (Xi-2, Yj+1). When its comparison value meets a set comparison condition, that is, (Xi, Yj) is less than (Xi-1, Yj+1), and greater than or equal to the pixel located on (Xi-2, Yj+1), then Indicates that a line is dropped. If the set comparison condition is not met, continue to compare the pixels on (Xi, Yj), (Xi-2, Yj+1) and (Xi-3, Yj+1). This is repeated until the comparison result meets the set comparison condition. When comparing, accumulate the number of disconnected lines and frame the places where the lines are dropped. After completion, start from column Yj+1 and repeat the above comparison until the comparison of the entire oblique line is completed. Finally, output the total number of dropped lines and display where there are dropped lines. It must be mentioned that it is not necessary to compare colored lines in the detection of dropped calls, because the problem of dropped calls has nothing to do with color. Therefore, testing in grayscale is sufficient.

In order to clearly show the flow chart of the present invention for judging dropped calls, please refer to FIG. 2 . When the user selects the test item as the detection of oblique lines, the user can set the standard deviation value according to the specifications of the scanner and the required quality. The judgment process is shown in Figure 2:

When testing, first read the standard oblique line on the test chart, preferably 45 degrees, namely step 201 . In step 202, according to the standard oblique line, the scanner generates an image of the standard oblique line. In step 203, the index is initialized, so that i points to the pixel position in the abscissa X whose grayscale value is closest to the middle grayscale value. And make i=1; j=1; L=diagonal line length (in picture element); m=0 (m is the counter used for accumulative disconnection); k=0 (k represents the exponent of displacement). Then, find out the pixel whose grayscale value is closest to the middle gray value, and set the coordinates of the pixel as (Xi, Yj), ie step 204 . Then in step 205, compare the pixels whose coordinates are (Xi, Yj) and (Xi-1, Yj+1). Step 206 is the determination of the comparison value. If the comparison result of the gray scale values of the two pixels is within the error tolerance, it means that there is no disconnection between the column of Yj and the next column, and the comparison to the next column can be directly performed. But it must be checked whether Yj+1 is the last column, so step 211 is executed. If yes, then step 213 outputs the information that m lines have been dropped, and at this time, since m=0, it means that there is no drop. If Yj+1 is not the last column, then the comparison of the next column is performed, so step 210 is performed to increase the index j of the seatpost Y by 1, and the index i of the abscissa X is subtracted by 1+k, and then perform step 205 to perform Comparing in a circle until the comparison of all columns of the entire slash is completed.

If the grayscale values of the two pixels (Xi, Yj) and (Xi−1, Yj+1) are not within the error tolerance, step 207 is executed. This means that there may be a phenomenon of disconnection, so increase the value of k by 1, and compare the gray levels (Xi-k, Yj+1) of the two adjacent pixels in the column Yj+1, (Xi-( k+1), Yj+1). In step 208, it is determined whether the gray scale of the (Xi, Yj) pixel is smaller than the gray scale of the (Xi-k, Yj+1) pixel, and greater than or equal to (Xi-(k+1), Yj+1) Plain grayscale. If yes, it means that only one line is dropped, go to step 212, make m=m+1. If not, the index of i should be increased by 1 in the column of Yj+1 to compare the pixels in the column of Yj+1 one by one until the gray scale of the pixel in (Xi, Yj) is smaller than that in the column of Yj+1. The gray level of a pixel is greater than or equal to the gray level of the pixel adjacent to the pixel, that is, step 208. If not, then step 209 is performed: get (Xi-k, Yj+1), (Xi-(k+1), Yj+1) pixel frames to represent the dropped line position, and increase m by 1 , which means that the number of comparisons is equal to the number of dropped calls. Then, step 207 is executed to continue the above comparison. into another loop. Finally, after all the comparisons are completed, output the total number of dropped lines and display the lines taken by the frame. When the total number of dropped calls is greater than the preset allowable dropout value, the scanner is judged to be unqualified; and when the total number of dropped calls is within the preset dropout allowable value, the scanner is judged to be qualified.

In order to further enable the examiner to understand the above process, an embodiment is used to illustrate the method for judging disconnection. As shown in Figure 3, after completing the image reading of the 45-degree oblique line (steps 201-202) and initializing the set value (step 203), the pixels whose middle gray value is closest to the middle gray scale are read for comparison Baseline (step 204 ), find the pixel whose gray scale is 136 in the Yj column, because its gray scale value is closest to 128. Point i to where the pixel is located. Let the standard deviation be 20, m=0. Then compare the pixels whose coordinates are (Xi, Yj) and (Xi−1, Yj+1) (step 205 ), ie 136 and 139 . Since the difference is within the standard deviation, add 1 to the index of Y on the ordinate, and subtract 1 to the index i on the X of the abscissa (step 210 ), and then execute step 205 .

Then compare the difference between 139 and 138 (step 205), and judge whether it is within the error tolerance (step 206). Since it is still within the standard deviation and has not yet reached the bottom of the line (step 211), the index j of the ordinate Y is increased by 1, and the index i of the abscissa X is subtracted by 1+k (step 210). Then step 205 is repeated to compare 138 and 179 . Since the difference exceeds the standard deviation, step 207 is executed. k=k+1, compare 138 with 178, 179, and then execute step 208. Since 138<178 and 138>179, proceed to step 209; take out the pixel frames of 179 and 178, increase m by 1, and add 1 to the value of k. Then step 207 is executed. Compare 138, 178, 137. Since 138<178 and 138<179, step 212 is executed: m is increased by 1, indicating that one more line is dropped. Then, step 211 is executed to determine whether the comparison of the entire line has been completed. Since it has not been completed, step 210 is executed to add 1 to the index j, i=i-(1+k) to perform the comparison of the next column. Then, step 205 is executed to make a loop comparison. In this way, the comparison continues until the end, and step 213 is executed to output the comparison result. As a result of the calculation, there are two lines dropped in m=2, and the dropped lines are also framed for easy identification.

For the convenience of explanation, this process is illustrated by the graphs and coordinates shown in Fig. 1 and Fig. 3 . If the representation method of the coordinates is changed, or the direction of the oblique line is changed, or the order of the compared pixels is changed from left to right or from bottom to top, etc., it does not deviate from the spirit of the present invention. For example, since the black part is the center of the line, and the two sides of the center are symmetrical, the above-mentioned method flow can also be applied to the other side, but the comparison in step 208 must be changed from less than to greater than, and greater than or equal to is less than or equal to. This is because the black part is in the middle, and when compared with the left side, the gray scale becomes lighter as the abscissa moves to the left.

Claims (14)

1. An automatic detection method for the disconnection phenomenon of an optical scanner, which is used to determine whether the scanned oblique line is disconnected according to a standard deviation value after the optical scanner reads the image of the oblique line of the test chart Phenomenon, and output the number of dropped calls, characterized in that: include steps: a. Initialize an index i of an abscissa X, an index j of an ordinate Y, a displacement index k, and a counter; b. Set an error tolerance value; c. Find a pixel whose grayscale value is closest to the intermediate grayscale value among the pixels in the first column of the oblique line, and the pixel is on the right side of the pixel with the largest grayscale value, and Set the coordinate value of the pixel to (Xi, Yj); d. Compare the grayscale value of the pixel (Xi, Yj) with the pixel (Xi-1, Yj+1) in the next column of the pixel, and output the first comparison value; e, when the first comparison value is within the range of the error tolerance value, the exponent i of the abscissa X is moved to the right by k+1 bits and the exponent j of the ordinate Y is increased by 1, and step (d) is executed ); f. When the first comparison value exceeds the range of the error tolerance value, compare the picture element (Xi, Yj) with the adjacent two picture elements (Xi-k, Yj+1) of the next column of the picture element , the grayscale value of (Xi-(k+1), Yj+1), and output the second comparison value; g. Determine whether the second comparison value meets a set comparison condition; h. When the second comparison value meets the set comparison condition, increase the counter by 1, and execute step (d); i. When the second comparison value does not meet the set comparison condition, frame the pixel located at the coordinates of (Xi-k, Yj+1) and (Xi-(k+1), Yj+1), Increment the counter by 1, and execute step f; j. When it is judged that j plus 1 is equal to the total length of the oblique line, output the judgment result. 2. The automatic side detection method according to claim 1, characterized in that: the oblique line of the test chart is a 45-degree oblique line. 3. The automatic side inspection method as claimed in claim 1, wherein the error tolerance value is set according to the specification of the scanner. 4. The automatic side inspection method according to claim 1, wherein said step (a) is to set the abscissa index i to be equal to 1, the ordinate index j to be equal to 1, the shift number k to be equal to zero, and the counter is equal to zero. 5. The automatic side detection method according to claim 1, wherein the above step (a) further comprises the step of: L setting the length of the oblique line. 6. The automatic side inspection method according to claim 1, wherein the comparison condition set in the above step (g) is when the gray scale value of (Xi, Yj) is less than the gray scale value located at (Xi-k, The pixel on Yj+1), and the grayscale value of (Xi, Yj) is greater than or equal to the pixel on (Xi-(k+1), Yj+1). 7. The automatic detection method according to claim 1, characterized in that: the above j includes the steps of: m. Set an allowable number of dropped calls; n. When the number of the above-mentioned counters is within the allowable number of dropped connections, it is determined that the scanner is qualified; o. When the number of the above-mentioned counter exceeds the allowable number of dropped connections, it is determined that the scanner is unqualified. 8. An automatic detection method for the disconnection phenomenon of an optical scanner, which is used to determine whether the scanned oblique line is disconnected according to a standard deviation value after the optical scanner reads the image of the oblique line of the test chart Phenomenon, and output the number of dropped calls, characterized in that: include steps: a. Initialize an index i of an abscissa X, an index j of an ordinate Y, a displacement index k, and a counter; b. Set an error tolerance value; c. Find a pixel whose grayscale value is closest to the intermediate grayscale value among the pixels in the first column of the oblique line, and the pixel is on the right side of the pixel with the largest grayscale value, and Set the coordinate value of the pixel to (Xi, Yj); d. Compare the grayscale value of the pixel (Xi, Yj) with the pixel (Xi-1, Yj+1) in the next column of the pixel, and output the first comparison value; e, when the first comparison value is within the range of the error tolerance value, shift the index i of the abscissa X to the left by k+1 and the index j of the ordinate Y increases by 1, and execute step (d ); f. When the first comparison value exceeds the range of the error tolerance value, compare the picture element (Xi, Yj) with the adjacent two picture elements (Xi+k, Yj+1) of the next column of the picture element (Xi+k), Yj+1), (Xi+(k+1), Yj+1) gray scale value, and output the second comparison value; g. Determine whether the second comparison value meets a set comparison condition; h. When the second comparison value meets the set comparison condition, increase the counter by 1, and execute step (d); i. When the second comparison value does not meet the set comparison condition, frame the pixel located at the coordinates of (Xi+k, Yj+1), (Xi+(k+1), Yj+1), and The calculator increments by 1 and executes step f; j. When it is judged that j plus 1 is equal to the total length of the oblique line, output the judgment result. 9. The automatic side detection method according to claim 8, characterized in that: the oblique line of the test chart is a 45-degree oblique line. 10. The automatic side inspection method as claimed in claim 8, wherein the error tolerance value is set according to the specification of the scanner. 11. The automatic side detection method as claimed in claim 8, wherein said step (a) is to set the abscissa index i equal to 1, the ordinate index j equal to 1, the shift number k equal to zero, and the counter is equal to zero. 12. The automatic side detection method according to claim 8, wherein the above step (a) further comprises the step of: L, setting the length of the oblique line. 13. The automatic side inspection method according to claim 8, characterized in that: the comparison condition set in the above step (g) is when the gray scale value of (Xi, Yj) is greater than the gray scale value located at (Xi+k, The grayscale value on Yj+1), and the grayscale value of (Xi, Yj) is less than or equal to the grayscale value on (Xi+(k+1), Yj+1). 14. The automatic side inspection method according to claim 8, characterized in that: the above j includes the steps of: m. Set an allowable number of dropped calls; n. When the number of the above-mentioned counters is within the allowable number of dropped connections, it is determined that the scanner is qualified; o. When the number of the above-mentioned counter exceeds the allowable number of dropped connections, it is determined that the scanner is unqualified.

Images