A Visual Aided Positioning Method for Crystal Bars in Wafer Processing Equipment

Abstract

The present invention relates to the technical field of silicon wafer processing, and provides a method for visually assisting in positioning a crystal bar for a silicon wafer processing device, including: Step 100, arranging a crystal bar visual acquisition system; Step 200, performing error calibration on the crystal bar visual acquisition system; Step 300, using the crystal bar visual system to acquire images of the bottom and top of the crystal bar, and calculating the deflection angle and position of the crystal bar based on the images of the bottom and top of the crystal bar to obtain the actual attitude of the crystal bar. The present invention can quickly position the silicon material and accurately complete the feeding.

Description

Technical Field

[0001] The present invention relates to the technical field of silicon wafer processing, and particularly relates to a method for visually assisting in positioning a crystal bar for a silicon wafer processing device. Background Art

[0002] The slicing machine and the polishing machine are important processing devices for photovoltaic silicon wafers, and their main function is to process the finished round bar. During the processing of the round crystal bar, there are high-precision requirements for the placement position and inclination angle of the round bar. Usually, the positioning accuracy is 0.05 mm. Therefore, a dedicated sensor positioning system is required on the device to verify the position and angle.

[0003] The sensor positioning system needs to combine multiple sensors with a rotating platform for detection. Its efficiency is low, the detection is limited to the area in contact with the sensors, and at the same time, each device needs to be equipped, resulting in high costs.

[0004] With the development of the automated industry, the loading and unloading of the crystal bar grabbing robot can be seamlessly connected to the automated production line, providing feeding for multiple processing devices at the same time. If the original method is still used, the robot only serves as a conveying device, resulting in increased costs and loss of performance. Therefore, it is considered to directly position the loading and unloading by the robot. One robot corresponds to multiple workstations, improving the processing efficiency and reducing the cost. For the positioning system that cooperates with the robot for loading and unloading, a positioning system with fast detection speed and more comprehensive detection is required. The current sensor detection method has problems in terms of accuracy, speed, and comprehensive detection, and it is difficult to achieve the degree of matching with robot automation. Therefore, a new detection method is needed to replace it. Summary of the Invention

[0005] The present invention mainly solves the technical problems of the existing sensor positioning system that needs to combine multiple sensors with a rotating platform for detection, with low efficiency, limited detection to the area in contact with the sensors, and high costs for each device. A method for visually assisting in positioning a crystal bar for a silicon wafer processing device is proposed to quickly position the silicon material, facilitate cooperation with the crystal bar grabbing robot, and accurately complete the feeding.

[0006] The present invention provides a method for visually assisting in positioning a crystal bar for a silicon wafer processing device, including the following processes:

[0007] Step 100, arranging a crystal bar visual acquisition system; the crystal bar visual acquisition system includes: an upper camera and a lower camera; the upper camera can move up and down; the measured crystal bar is placed at a position between the upper camera and the lower camera.

[0008] Step 200, performing error calibration on the crystal bar visual acquisition system, including steps 201 to 206:

[0009] Step 201: Select two standard inspection rods with different lengths, namely a standard long inspection rod with a length of d and a standard short inspection rod with a length of e.

[0010] Step 202: Based on the lengths of the standard long inspection rod and the standard short inspection rod, obtain the height b between the standard long inspection rod and the upper camera and the height c between the standard short inspection rod and the upper camera, where:

[0011] c - b = d – e;

[0012] Step 203: Adjust the standard long inspection rod to the target position, and the upper camera records the position of the upper circle of the standard long inspection rod.

[0013] Step 204: Adjust the standard short inspection rod to the target position, and the upper camera records the position of the upper circle of the standard short inspection rod.

[0014] Step 205: Calculate the position error and tilt error of the ingot visual acquisition system based on the position of the upper circle of the standard long inspection rod and the position of the upper circle of the standard short inspection rod.

[0015] Step 206: Calculate the standard pixel position corresponding to the length by calibrating the system parameters.

[0016] Step 300: Use the ingot visual system to acquire the image of the bottom of the ingot and the image of the top of the ingot, and calculate the deflection angle and position of the ingot based on the image of the bottom of the ingot and the image of the top of the ingot to obtain the actual attitude of the ingot, including steps 301 to 304:

[0017] Step 301: After the ingot gripping robot places the ingot on the ingot visual acquisition system, the lower camera acquires the image of the bottom of the ingot, and determines the center position of the bottom of the ingot based on the image of the bottom of the ingot.

[0018] Step 302: The upper camera acquires the image of the top of the ingot, and determines the center position fitted by the ingot edge image in the upper camera based on the image of the top of the ingot.

[0019] Step 303: According to the distance from the upper end face of the ingot to the camera, when the length of the ingot is d, the distance from the upper end face is b, and when the length of the ingot is any length x0, calculate the actual spatial position of the upper part as b + d - x0.

[0020] Step 304: Calculate the deflection angle and position of the ingot based on the actual spatial position of the upper part of the ingot to obtain the actual attitude of the ingot.

[0021] Furthermore, for the ingot visual acquisition system, a first light source for long ingots and a second light source for long ingots are respectively arranged on both sides above the ingot placement position.

[0022] A first light source for short ingots and a second light source for short ingots are respectively arranged on both sides of the ingot placement position;

[0023] The light source for short ingots is below the corresponding light source for long ingots;

[0024] The lower camera is equipped with a lower camera light source.

[0025] Furthermore, in step 205, the position error and tilt error of the ingot vision acquisition system are calculated, including:

[0026] Calculation method of the position error f:

[0027] The pixel coordinate difference f_pixels of the centers of two circles is obtained through image processing. At the same time, according to the actual photographing distance c, the camera focal length focus, and the pixel size p, the distance f between the point on the top plane of the standard short inspection bar and the central axis of the camera axis can be calculated through the formula:

[0028] f = p * f_pixels / focus * c;

[0029] Calculation method of the tilt error angle a1 and the deviation distance a of camera 1:

[0030] According to the distance f between the point on the top plane of the standard short inspection bar and the central axis of the camera axis obtained, the deflection angle between the camera axis and the central axis of the system and the deviation distance of the upper camera are calculated:

[0031] a1 = a2 = atan(f / (d - e));

[0032] a = f / (d - e) * b;

[0033] Among them, a1 represents the deflection angle between the camera axis and the central axis of the system, the angle a2 is the apex angle of a1, and a1 = a2; a represents the deviation distance of the upper camera.

[0034] Furthermore, in step 304, according to the actual spatial position of the upper part of the ingot, the deflection angle and position of the ingot are calculated to obtain the actual attitude of the ingot, including:

[0035] When there is no error in the system and the end face positions of the ingot detected by the upper camera and the lower camera are both at the center, it indicates that the ingot is vertically upward and there is no deviation angle;

[0036] When there is an error in the system, when the lower camera detects that the end face position of the ingot is normal and the upper camera detects that the end face position is not at the center and deviates to the right by x, the ingot is tilted. At this time, the tilt angle of the ingot is:

[0037] Angle = atan(x / d);

[0038] Obtain the actual posture of the ingot according to the inclination angle of the ingot.

[0039] A visual auxiliary positioning method for an ingot in a wafer processing device provided by the present invention mainly solves the positioning of the ingot. The positioning includes the X and Y positions of the ingot and the angle with the Z axis. In the height direction, the system is based on the bottom surface of the ingot. When the positioning is completed, the spatial position and perpendicularity of the ingot can be ensured to be correct, and the posture adjusted by the robot is recorded, so as to assist the robot to ensure this posture when placing the ingot. By adopting this method, it is easy to realize a robot with a set of vision to pick up from the transfer roller table and feed multiple processing stations. Compared with the original mechanical positioning method, which requires each station to be equipped, the cost can be significantly reduced and the degree of automation can be improved. In the visual positioning method, the photographing and processing time are both within 1S, and the speed is also faster compared with the traditional mechanical detection. Therefore, the present invention replaces the existing mechanical detection method, with better detection effect, faster speed, and can match the ingot grasping robot to improve the degree of automation; low cost and good efficiency. Brief Description of the Drawings

[0040] Figure 1 is the implementation flowchart of the visual auxiliary positioning method for an ingot in a wafer processing device provided by the present invention;

[0041] Figure 2 is the identification of the ingot visual acquisition system Figure 1 (including the slide rail);

[0042] Figure 3 is the identification of the ingot visual acquisition system Figure 2 (excluding the slide rail);

[0043] Figure 4 is the coaxial schematic diagram of the upper camera and the lower camera when there is no error in the ingot visual acquisition system;

[0044] Figure 5 is the non - coaxial schematic diagram of the upper camera and the lower camera when there is an error in the ingot visual acquisition system;

[0045] Figure 6 is the schematic diagram of error calibration of the ingot visual acquisition system;

[0046] Figure 7 is the schematic diagram of the images of the top and bottom of the ingot when the position and angle of the ingot are correct;

[0047] Figure 8 is the schematic diagram of the images of the top and bottom of the ingot when the position and angle of the ingot are correct but there is an inclination angle.

[0048] Reference numerals: 1. Upper camera; 2. Moving guide rail; 3. Light source for the first long crystal bar; 4. Light source for the second long crystal bar; 5. Light source for the first short crystal bar; 6. Light source for the second short crystal bar; 7. Crystal bar; 8. Light source for the lower camera; 9. Lower camera; 10. Central axis. Detailed implementation manners

[0049] To make the technical problems solved by the present invention, the technical solutions adopted and the achieved technical effects clearer, the present invention will be further described in detail below with reference to the drawings and embodiments. It can be understood that the specific embodiments described herein are only used to explain the present invention, rather than limiting the present invention. In addition, it should be noted that for the sake of description, only parts related to the present invention are shown in the drawings, rather than all the content.

[0050] As Figure 1 shown, the crystal bar vision-assisted positioning method for a silicon wafer processing device provided by an embodiment of the present invention includes:

[0051] Step 100, arranging a crystal bar vision acquisition system.

[0052] The crystal bar vision acquisition system includes an upper camera 1 and a lower camera 9. As Figure 2 shown, the upper camera 1 can move up and down; the upper camera 1 can be installed through a moving guide rail 2 that can move up and down; or the moving guide rail 2 can be not provided (as Figure 3 shown), and the camera depth of field is used to be compatible with different heights.

[0053] The measured crystal bar 7 is placed by a crystal bar grasping robot at a position between the upper camera 1 and the lower camera 9. The measured crystal bar 7 includes long bars and short bars. In the crystal bar vision acquisition system, the placement position of the crystal bar is fixed, the lower surface of the crystal bar 7 is a standard surface, and the position heights of the upper surfaces of the crystal bars 7 with different lengths are different.

[0054] In addition, a first light source 3 for the long crystal bar and a second light source 4 for the long crystal bar are respectively arranged on both sides above the crystal bar placement position. A first light source 5 for the short crystal bar and a second light source 6 for the short crystal bar are respectively arranged on both sides of the crystal bar placement position. The light source for the short crystal bar is below the corresponding light source for the long crystal bar. The lower camera 9 is configured with a light source 8 for the lower camera.

[0055] Step 200, performing error calibration on the crystal bar vision acquisition system.

[0056] If the upper camera 1 and the lower camera 9 are on the central axis of the system and there is no offset in the crystal bar vision acquisition system (when there is no error in the system and the upper camera 1 and the lower camera 9 are coaxial), accurate detection of long and short bars can be achieved, and the compatibility effect is good. See Figure 4However, in practice, there are inevitably more or less errors in the crystal bar vision acquisition system. When there are errors in the crystal bar vision acquisition system, when the upper camera 1 deviates from the central axis by 10, theoretically, with the long bar as the center, the system needs to be calibrated. If not calibrated, it will cause the system to detect the long bar correctly, but there will be a deviation in the upper surface distance f for the short bar, resulting in a deviation in the placement angle, which has a significant impact on the placement. Therefore, it is necessary to calibrate the errors of the system. Refer to Figure 5 。

[0057] The error calibration method refers to Figure 6 : Figure 6 In it, dimension a is the distance by which the upper camera deviates; dimension b represents the distance from the upper camera to the top of the long bar; dimension c represents the distance from the upper camera to the top of the short bar; dimension d represents the length of the long bar; dimension e represents the length of the short bar; dimension f represents the distance between the point on the top plane of the short bar where the camera axis is located and the central axis; angle a1 represents the deviation angle between the camera axis and the system central axis; angle a2 represents the apex angle of a1 and is equal; among the above parameters, the distances a, angles a1, a2, and f all represent unknown parameters and need to be calibrated and calculated. The main methods are as follows:

[0058] Step 201, select two standard inspection bars with different lengths, namely the standard long inspection bar with a length of d and the standard short inspection bar with a length of e;

[0059] Step 202, according to the lengths of the standard long inspection bar and the standard short inspection bar, obtain the height b between the standard long inspection bar and the upper camera 1 and the height c between the standard short inspection bar and the upper camera 1, where:

[0060] c - b = d – e;

[0061] Step 203, adjust the standard long inspection bar to the target position, and the upper camera 1 records the upper circular position of the standard long inspection bar;

[0062] In this step, place the standard long inspection bar with a length of d below the upper camera 1. The external measuring device detection method can be used to adjust the crystal bar to the standard target position.

[0063] Step 204, adjust the standard short inspection bar to the target position, and the upper camera 1 records the upper circular position of the standard short inspection bar;

[0064] In this step, place the standard short inspection bar with a length of e below the upper camera 1 and adjust the position in the same way to ensure the accuracy of the position.

[0065] Step 205, according to the upper circular positions of the standard long inspection bar and the standard short inspection bar, calculate the position error and tilt error of the crystal bar vision acquisition system.

[0066] The calculation method of the position error f of the system:

[0067] The difference in the center pixel coordinates f_pixels of two circles is obtained through image processing. At the same time, according to the actual photographing distance c, the camera focal length focus, and the pixel size p, the distance f between the point on the top plane of the standard short calibration rod and the central axis of the camera axis can be calculated by the formula:

[0068] f = p * f_pixels / focus * c;

[0069] Calculation method for the tilt error angle a1 of the system and the deviation distance a of camera 1:

[0070] According to the obtained distance f between the point on the top plane of the standard short calibration rod and the central axis of the camera axis, calculate the deviation angle between the camera axis and the central axis of the system and the deviation distance of the upper camera 1:

[0071] a1 = a2 = atan(f / (d - e));

[0072] a = f / (d - e) * b;

[0073] Among them, a1 represents the deviation angle between the camera axis and the central axis of the system, the angle a2 is the apex angle of a1, and a1 = a2; a represents the deviation distance of the upper camera 1.

[0074] Using this method, the tilt parameters of the system can be obtained only through two sets of standard data;

[0075] Step 206, after obtaining the system tilt parameters, during the actual use process, after obtaining the length d1 of the crystal bar, the standard pixel position std corresponding to the length can be calculated by calibrating the system parameters.

[0076] Std = (d – d1) * tan(a1)

[0077] Adjust the measured position to the std pixel position, and the crystal bar of this length can be adjusted in place.

[0078] The above is a one-dimensional example. During actual use, the deviation f exists within the plane of the image. It can be decomposed into two one-dimensional systems through vector decomposition. Since the parameters are independent of each other, this method can be used separately to achieve actual detection.

[0079] Summary of the calibration process in step 200: Select a standard long calibration rod with a length of d, calibrate it to the standard position and place it in the system for image acquisition; select a standard short calibration rod with a length of e, calibrate it to the standard position and place it in the system for image acquisition; analyze to obtain the tilt error and position error, and complete the automatic calibration.

[0080] Step 300: Use the ingot visual system to collect the images of the bottom and top of the ingot 7, and calculate the deflection angle and position of the ingot 7 based on the images of the bottom and top of the ingot 7 to obtain the actual attitude of the ingot 7.

[0081] The present invention is mainly used to solve the positioning of the ingot 7, and its positioning includes the X and Y positions of the ingot and the angle with the Z axis. In the height direction, the system is based on the bottom surface of the ingot. When the positioning is completed, the spatial position and perpendicularity of the ingot can be ensured to be correct, and the attitude adjustment of the robot is recorded, so as to assist the robot to ensure this attitude when placing the ingot.

[0082] Step 301: After the ingot grasping robot places the ingot on the ingot visual acquisition system, the lower camera 9 collects the image of the bottom of the ingot 7, and determines the central position of the bottom of the ingot according to the image of the bottom of the ingot 7.

[0083] In this step, since the bottom surface of the ingot grasping robot is the reference surface and the position from the lower camera 9 to the bottom of the ingot grasping robot remains unchanged, during actual work, the lower camera 9 takes a picture, and the central position of the lower ingot can be determined by the boundary analysis method.

[0084] Step 302: The upper camera 1 collects the image of the top of the ingot 7, and determines the central position fitted by the ingot edge image in the upper camera 1 according to the image of the top of the ingot 7.

[0085] In this step, for the ingot 7 with different lengths, the size of the captured image by the upper camera 1 will be different. During actual work, image acquisition is carried out, and the central position of the circle formed by the edge of the ingot in the image captured by the upper camera is determined by the boundary analysis method.

[0086] Step 303: Calculate the actual spatial position of the upper part of the ingot 7 according to the distance from the upper end face of the ingot 7 to the camera.

[0087] When the length of the ingot 7 is the same as the standard length of the inspection bar, that is, the length of the ingot 7 is d, and the distance from the upper end face of the ingot 7 to the upper camera 1 is b, calculate the actual spatial position of the upper part when the length of the ingot is any length x0.

[0088] In this step, according to the relationship between d and b, the actual spatial position of the upper part is calculated as b + d - x0.

[0089] Step 304: Calculate the deflection angle and position of the ingot 7 according to the actual spatial position of the upper part of the ingot 7 to obtain the actual attitude of the ingot 7.

[0090] (1) As Figure 7 shown, when there is no error in the system and the end face positions of the ingot detected by the upper camera 1 and the lower camera 9 are both at the center, it indicates that the ingot is vertically upward and there is no deviation angle.

[0091] (2) As Figure 8 shown, when there is an error in the system, the lower camera 9 detects the position of the end face of the ingot 7 normally, while the upper camera 1 detects that the position of the end face is not at the center. When there is a deviation of x to the right, the ingot 7 is tilted. At this time, the tilt angle of the ingot 7 is:

[0092] Angle = atan(x / d);

[0093] According to the tilt angle of the ingot 7, the actual attitude of the ingot 7 is obtained. After the angle calculation is completed, the tilt angle value is transmitted to the ingot grasping robot for adjustment. At the same time, the bottom of the ingot 7 is used as a position reference. When there is a deviation in the position of the bottom circle, this deviation value is also transmitted to the ingot grasping robot for adjustment, ultimately ensuring that the angle and position are adjusted to the correct range.

[0094] After obtaining the actual attitude of the ingot 7, the actual attitude is transmitted to the ingot grasping robot, and the ingot grasping robot makes corresponding position and angle adjustments, so as to ensure that the spatial attitudes of the ingots before feeding are consistent.

[0095] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: modifying the technical solutions recorded in the foregoing embodiments, or equivalently replacing some or all of the technical features therein, does not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A visual assistance positioning method for a crystal bar in a silicon wafer processing device, characterized in that, It includes the following processes: Step 100, arranging a visual acquisition system for the ingot; The visual acquisition system for the ingot includes: an upper camera (1) and a lower camera (9); the upper camera (1) can move up and down; the ingot (7) to be measured is placed at a position between the upper camera (1) and the lower camera (9); Step 200, performing error calibration on the visual acquisition system for the ingot, including steps 201 to 206: Step 201, selecting two standard inspection rods with different lengths, namely a standard long inspection rod with a length of d and a standard short inspection rod with a length of e; Step 202, according to the lengths of the standard long inspection rod and the standard short inspection rod, obtaining the height b between the top of the standard long inspection rod and the upper camera (1) and the height c between the top of the standard short inspection rod and the upper camera (1), where: c - b = d – e; Step 203, adjusting the standard long inspection rod to the target position, and the upper camera (1) records the upper circular position of the standard long inspection rod; Step 204, adjusting the standard short inspection rod to the target position, and the upper camera (1) records the upper circular position of the standard short inspection rod; Step 205, according to the upper circular positions of the standard long inspection rod and the standard short inspection rod, calculating the position error and tilt error of the visual acquisition system for the ingot, including the following steps: Calculation method of the position error f: By image processing, the pixel coordinate difference f_pixels between the centers of the two is obtained. At the same time, according to the actual photographing distance c, the camera focal length focus, and the pixel size p, the distance f between the point on the top plane of the standard short inspection rod of the camera axis and the system central axis can be calculated through the formula: f = p * f_pixels / focus * c; Calculation methods of the tilt error angle a1 and the deviation distance a of camera 1: According to the obtained distance f between the point on the top plane of the standard short inspection rod of the camera axis and the system central axis, calculate the deflection angle between the camera axis and the system central axis and the deviation distance of the upper camera (1): a1 = a2 = atan(f / (d-e)); a = f / (d-e) * b; Among them, a1 represents the deflection angle between the camera axis and the system central axis, the angle a2 is the apex angle of a1, and a1 = a2; a represents the deviation distance of the upper camera (1) from the system central axis; Step 206, by calibrating the system parameters, calculating the standard pixel position corresponding to the length; Step 300, using the visual system for the ingot to collect the image of the bottom of the ingot (7) and the image of the top of the ingot (7), and according to the image of the bottom of the ingot (7) and the image of the top of the ingot (7), calculating the deflection angle and position of the ingot (7) to obtain the actual attitude of the ingot (7), including: steps 301 to 304: Step 301, after the ingot gripping robot places the ingot on the visual acquisition system for the ingot, the lower camera (9) collects the image of the bottom of the ingot (7), and determines the central position of the bottom of the ingot according to the image of the bottom of the ingot (7); Step 302: The upper camera (1) captures an image of the top of the ingot (7), and determines the center position of the fitting of the ingot edge image in the upper camera (1) based on the image of the top of the ingot (7). Step 303: According to the distance from the upper end face of the ingot (7) to the camera, when the length of the ingot (7) is the same as the standard length of the inspection bar, the length of the ingot (7) is d, and the distance between the upper end face of the ingot (7) and the upper camera (1) is b. When the length of the ingot is any length x0, calculate the actual spatial position of the upper part as b + d - x0. Step 304: According to the actual spatial position of the upper part of the ingot (7), calculate the deflection angle and position of the ingot (7) to obtain the actual attitude of the ingot (7), including the following steps: When there is no error in the system, when the end face positions of the ingot detected by the upper camera (1) and the lower camera (9) are both at the center, it indicates that the ingot is vertically upward and there is no deviation angle. When there is an error in the system, when the end face position of the ingot (7) detected by the lower camera (9) is normal, and the end face position detected by the upper camera 1 is not at the center and deviates to the right by x, the ingot (7) is tilted. At this time, the tilt angle of the ingot (7) is: Angle = atan(x / d); Obtain the actual attitude of the ingot (7) according to the tilt angle of the ingot (7).

2. The crystal bar vision-assisted positioning method for a silicon wafer processing device according to claim 1, characterized in that, In the ingot vision acquisition system, a first ingot light source (3) and a second ingot light source (4) are respectively arranged on both sides above the ingot placement position. A first short ingot light source (5) and a second short ingot light source (6) are respectively arranged on both sides of the ingot placement position. The short ingot light source is below the corresponding long ingot light source. The lower camera (9) is configured with a lower camera light source (8).

Images