Method for measuring pixel size of scribing machine

Abstract

The invention relates to the field of machine vision, in particular to a method for measuring the pixel size of a scribing machine, which comprises the following steps of: moving a camera to a position with most displayed image texture features; k + 1 different position points are selected, and the actual offset distance of K segments is recorded; identifying a pixel offset distance of image movement through a visual comparison method, constructing a relational expression between the pixel offset distance and an actual offset distance by adopting a Bezier curve, and calculating an actual distance corresponding to any pixel distance; and clicking a designated expected position of the screen, and observing whether the center of the moved image is the expected position or not. According to the method, one-key automatic acquisition of the pixel size is realized through image recognition, a person does not need to manually select a measurement position, the pixel size is acquired in different directions and at different positions, a Bezier curve is adopted to construct a relational expression between the pixel offset distance and the actual offset distance, the actual distance corresponding to any pixel distance is calculated, and the accuracy of the pixel size is improved. Therefore, the result is more accurate, and the accuracy of the pixel size can be conveniently and intuitively verified.

Description

Technical Field

[0001] The present invention relates to the field of machine vision, and particularly to a method for measuring the pixel size of a dicing machine. Background Art

[0002] A wafer cutting device, a dicing machine, is a device used to cut wafers or other sheet-shaped materials into individual chips. During the cutting process of the device, it is necessary to frequently convert the image coordinates and the actual coordinates, and reflect the actual movement process or image information to the upper computer PC for display. At the same time, the operator also needs to confirm the target position during the movement of the image, so as to realize the interaction between the PC display image and the device movement. The schematic diagram of its principle is as Figure 1 shown.

[0003] The schematic diagram of the current existing pixel measurement method is as Figure 2 shown. Usually, it moves in a single direction manually, records the movement distance and the pixel distance, and calculates the actual movement distance corresponding to a single pixel. The existing method is usually debugged and saved before the device leaves the factory. Due to the size change caused by the later device change, it is necessary for the manufacturer's personnel to perform calibration tests. The existing method for calibrating the pixel size is relatively complex in operation and limited in measurement accuracy.

[0004] In view of this, we propose a method for measuring the pixel size of a dicing machine to solve the existing problems. Summary of the Invention

[0005] The purpose of the present invention is to provide a method for measuring the pixel size of a dicing machine to solve the problems raised in the above background art.

[0006] To achieve the above purpose, the present invention provides the following technical solution: A method for measuring the pixel size of a dicing machine, the steps include:

[0007] S1: Obtain the display image: The camera moves above the wafer product and performs automatic focusing, adjusts the brightness and displays the image, and moves to the position with the most texture features of the display image;

[0008] S2: Automatically measure the position: Select K + 1 position points at different positions along the Y direction given by the PLC, select one of the position points as the origin position, and automatically move the camera by the motion mechanism, and record the K actual offset distances moved relative to the origin position;

[0009] S3: Automatically calculate the pixel size: Identify the pixel offset distance of the image movement by the visual comparison method, and use the Bezier curve to construct the relationship formula between the pixel offset distance and the actual offset distance, so as to calculate the actual distance corresponding to any pixel distance;

[0010] S4: Verify accuracy: Click on the specified desired position on the screen and observe whether the center of the moved image is the desired position.

[0011] Further, in S2, K is greater than or equal to 2.

[0012] Further, in S3, the specific steps for identifying the pixel distance of image movement include:

[0013] A1: Record the image at the origin position and the K offset position images.

[0014] A2: Move the image at the origin position in the first step and match it with the first offset position image. The degree of overlap will increase from low to high and then decrease.

[0015] A3: Select the position with the highest degree of overlap during the first-step movement for fine movement. Move according to the second step and select the position with the highest degree of overlap during the second-step movement. The offset of this position relative to the origin position is the pixel distance of the first offset position.

[0016] A4: Obtain the pixel distances of the remaining K - 1 offset positions according to the operation methods of A2 and A3. Calculate the control points of the Bezier curve based on the K pixel distances of the offset positions and the actual distances of the K offset positions, and construct the Bezier curve relationship.

[0017] Further, the first step in A2 is greater than or equal to 2pix.

[0018] Further, the second step in A3 is equal to 1pix.

[0019] Further, in A4, the input value of the Bezier curve relationship is the pixel distance of the offset position, and the output value is the actual distance of the offset position.

[0020] Further, in S3, if the pixel shape is square, there is no need to measure and calculate in the X direction, and the actual distance corresponding to the pixel distance in the X direction is the same as that in the Y direction.

[0021] Further, in S3, if the pixel shape is not square, it is necessary to measure and calculate in the X direction, and the operation method is the same as that in the Y direction.

[0022] Further, in S4, if the center of the moved image is the desired position, this step is completed.

[0023] Further, in S4, if the center of the moved image is not the desired position, repeat the operation steps from S1 to S4.

[0024] Compared with the prior art, the beneficial effects of the present invention are:

[0025] The present invention realizes one - key automatic acquisition of pixel size through image recognition, without manual selection of measurement positions by personnel, and obtains pixel sizes in different directions and at different positions. A relationship formula between pixel offset distance and actual offset distance is constructed using Bessel curves to calculate the actual distance corresponding to any pixel distance, making the result more accurate and enabling convenient and intuitive verification of the accuracy of pixel size. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 is a schematic diagram of the principle of pixel measurement in the background technology;

[0027] Figure 2 is a schematic diagram of the method of pixel measurement in the background technology;

[0028] Figure 3 is a schematic flowchart of a method for measuring the pixel size of a dicing machine according to the present invention;

[0029] Figure 4 is a schematic diagram of the measurement position according to the present invention;

[0030] Figure 5 is a schematic diagram of the principle of pixel distance calculation according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0031] The technical solutions of the present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0032] Embodiment 1

[0033] As Figure 3 shown, a method for measuring the pixel size of a dicing machine includes the following steps:

[0034] S1: Obtain the display image: The camera moves above the wafer product and performs automatic focusing, adjusts the brightness and displays the image, and moves to the position with the most texture features in the display image;

[0035] S2: Automatically measure the position: Select K + 1 position points at different positions along the Y - direction given by the PLC, select one of the position points as the origin position, and automatically move the camera through the motion mechanism, and record the K actual offset distances moved relative to the origin position, where K is greater than or equal to 2; As Figure 4 shown, select 3 position points at different positions, automatically move through the motion mechanism, and record the 2 actual offset distances moved. When measuring in the Y - direction, the actual offset distance of the first offset measurement position relative to the origin measurement position is dy1, and the actual offset distance of the second offset measurement position relative to the origin measurement position is dy2;

[0036] S3: Automatically calculate pixel size: Identify the pixel offset distance of the image movement through the visual comparison method, and use the Bezier curve to construct the relationship between the pixel offset distance and the actual offset distance, so as to calculate the actual distance corresponding to any pixel distance; if the pixel shape is square, there is no need to measure and calculate the X direction, and the actual distance corresponding to a single pixel in the X direction is the same as that in the Y direction; if the pixel shape is not square, it is necessary to measure and calculate the X direction, and the operation method is the same as that in the Y direction.

[0037] S4: Verify accuracy: Click on the specified desired position on the screen and observe whether the center of the moved image is the desired position; if the center of the moved image is the desired position, this step is completed; if the center of the moved image is not the desired position, repeat the operation steps of S1 to S4.

[0038] Among them, in S3, as Figure 5 shown, the specific steps to identify the pixel distance of the image movement include:

[0039] A1: Record the image at the origin position, the first offset position image, and the second offset position image.

[0040] A2: Move the image at the origin position in steps of 3pix and match it with the first offset position image for comparison. The overlap degree will change from low to high and then to low.

[0041] A3: Select the position with the highest overlap degree during the 3pix step movement for fine movement, move in steps of 1pix, and select the position with the highest overlap degree during the 1pix step movement. The offset of this position relative to the origin position is the pixel distance Y1 of the first offset position.

[0042] A4: Obtain the pixel distance Y2 of the second offset position according to the operation methods of A2 and A3. Calculate the control points of the Bezier curve based on the two offset position pixel distances and the two offset position actual distances, and construct the Bezier curve relationship. Among them, the input value of the Bezier curve relationship is the offset position pixel distance, and the output value is the offset position actual distance.

[0043] A method for measuring the pixel size of a dicing machine based on Embodiment 1 is:

[0044] The mathematical model of the N - th order Bezier curve is:

[0045]

[0046] In the formula: is the Bezier basis function; p i is the control point of the Bezier curve; u is the value before the transformation of the Bezier curve; P(u) is the value after the transformation of the Bezier curve.

[0047] To ensure the accuracy of the calculation, K + 1 different positions are selected to construct the relationship between the pixel offset distance and the actual offset distance using B-spline curves.

[0048] Relative to the origin measurement position, the actual offset distances of the K offset measurement positions are dy 1 , dy 2 ,..., dy K , and the corresponding pixel offset distances are Y 1 , Y 2 ,..., Y K .

[0049] Let DY S be the set of dy 1 , dy 2 ,..., dy K , and let Y S be the set of Y 1 , Y 2 ,..., Y K . Then B(Y S )·P = DY S , where B(Y S ) is the B-spline basis function of Y S , and P is the set of B-spline curve control points, from which P is obtained.

[0050] During the calculation and measurement process, let the pixel offset distance to be measured be Y, and the corresponding actual offset distance dy = B(Y)·P is obtained.

[0051] The above specific embodiments are merely several preferred embodiments of the present invention. Based on the technical solution of the present invention and the relevant revelations of the above embodiments, those skilled in the art can make various alternative improvements and combinations to the above specific embodiments.

Claims

1. A method for measuring pixel size of a dicing machine, characterized in that the steps include: S1: Acquire display image: The camera moves to the top of the wafer product and implements autofocus, adjusts the brightness and displays the image, and moves to the position with the most texture features in the display image; S2: Automatic position measurement: select K+1 different position points along the Y direction given by the PLC, select one of the position points as the origin, automatically move the camera through the motion mechanism, and record the actual offset distance of K segments moved relative to the origin; S3: Automatically calculate pixel size: Use visual comparison method to identify the pixel offset distance of image movement, use Bezier curve to construct the relationship between pixel offset distance and actual offset distance, and calculate the actual distance corresponding to any pixel distance; S4: Verify accuracy: Click the desired position specified on the screen and observe whether the center of the image after moving is the expected position.

2. The method for measuring pixel size of a dicing machine according to claim 1, characterized in that: In S2, K is greater than or equal to 2.

3. The method for measuring pixel size of a dicing machine according to claim 1, characterized in that: In S3, the specific steps of identifying the pixel distance of image movement include: A1: Record the origin position image and K offset position images; A2: Move the origin position image according to the first step and compare it with the first offset position image. The overlap will increase from low to high and then to low. A3: Select the position with the highest overlap in the first step movement for refined movement, and select the position with the highest overlap in the second step movement according to the second step movement. The offset of this position relative to the origin position is the pixel distance of the first offset position; A4: Obtain the remaining K-1 offset position pixel distances according to the operation methods of A2 and A3, obtain the Bezier curve control points according to the K offset position pixel distances and the K offset position actual distances, and construct the Bezier curve relationship.

4. The method for measuring pixel size of a dicing machine according to claim 3, characterized in that: The first step in A2 is greater than or equal to 2pix.

5. The method for measuring pixel size of a dicing machine according to claim 3, characterized in that: The second step in A3 is equal to 1pix.

6. The method for measuring pixel size of a dicing machine according to claim 3, characterized in that: In A4, the input value of the Bezier curve relationship is the pixel distance of the offset position, and the output value is the actual distance of the offset position.

7. The method for measuring pixel size of a dicing machine according to claim 1, characterized in that: In S3, if the pixel shape is a square, there is no need to perform measurement calculation in the X direction, and the actual distance corresponding to the pixel distance in the X direction is consistent with that in the Y direction.

8. The method for measuring pixel size of a dicing machine according to claim 1, characterized in that: In S3, if the pixel shape is not a square, the X direction needs to be measured and calculated, and the operation method is the same as the Y direction.

9. The method for measuring pixel size of a dicing machine according to claim 1, characterized in that: In S4, if the center of the image after the movement is the desired position, the step is completed.

10. The method for measuring pixel size of a dicing machine according to claim 1, characterized in that: In S4, if the center of the image after the movement is not at the desired position, the operation steps from S1 to S4 are repeated.

Images