A variable pitch multi-nozzle calibration method

Abstract

The present invention discloses a variable pitch multi-nozzle calibration method, aiming to provide a variable pitch multi-nozzle calibration method. The variable pitch multi-nozzle calibration method realizes obtaining the coordinates obtained from only one photographing, obtaining the offset coordinates of multiple nozzles, enabling all multiple nozzles to perform blind picking and placing without the need for visual re-guidance. The method is simple and efficient, providing technical support for actual production. The present invention realizes variable pitch multi-nozzle calibration through steps of top camera calibration, bottom camera calibration, and calibration of the relationship between the optical center of the bottom camera and the center of the nozzle marking point. The present invention is applied to the technical field of machine vision multi-nozzle calibration.

Description

Technical Field

[0001] The present invention relates to a calibration method, and particularly to a calibration method for variable pitch multi - nozzles. Background Art

[0002] In modern automated production processes, machine vision systems are widely used in various industries. Machine vision systems can greatly improve production efficiency and the degree of production automation, and machine vision has higher precision compared to human vision. With the popularization of the application of machine vision in industry, high - precision industrial production equipment has put forward higher precision requirements for machine vision. It is difficult for a single camera to achieve high precision, while multiple cameras can meet the requirements of high precision and large field of view. However, multiple cameras need to rely on a high - precision calibration board to calibrate the coordinate transformation relationship between cameras and the high - degree coupling between various components to achieve high - precision requirements. The traditional calibration method for a single camera is inefficient, complex to operate, and requires professional personnel to complete the calibration.

[0003] At present, the calibration methods for multi - camera systems mostly use manual visual inspection plus simple auxiliary objects to adjust the focal length, field - of - view registration, and collinearity of the scanning area. These methods have disadvantages such as poor accuracy, low efficiency, and limited adjustable parameters. If the perpendicular calibration of the axis of a single camera to the detection plane cannot be achieved, the accuracy of the entire system depends too much on the manufacturing and installation accuracy of the mechanical bracket.

[0004] Meanwhile, the existing manual multi - camera calibration records the respective calibration mapping matrices through the 9 - point calibration of each camera. However, this manual calibration method will cause a large amount of cumulative errors due to manual movement, and it is difficult to eliminate the errors. Manually calibrating point by point and saving data increases the complexity of the steps. Due to the problem of manual use differences, the accuracy is low, and it greatly increases the workload and fatigue of the users.

[0005] In addition, the existing calibration technologies have cumbersome steps and complex operations. It is necessary for the user to control the motion axis and the vision system simultaneously to achieve calibration. During the calibration process, the operator needs to adjust the clarity of the camera imaging, the placement position of the target, and the brightness of the light source according to the real - time image, and then through calculation, obtain the corresponding calibration parameters to complete the entire calibration process. For different operators, the evaluation of imaging clarity, image distortion degree, color, and brightness uniformity is somewhat subjective, and it is impossible to ensure the consistency of the calibration evaluation results. At the same time, the process of manual parameter calculation is complex and cumbersome, prone to errors, reducing the accuracy and efficiency of the calibration of the vision detection system. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to overcome the deficiencies of the prior art and provide a variable pitch multi-nozzle calibration method. The variable pitch multi-nozzle calibration method realizes obtaining the offset coordinates of multiple nozzles from the coordinates obtained by taking a single photo, enabling all multiple nozzles to perform blind picking and placing without the need for visual re-guidance. The method is simple and efficient, providing technical support for actual production.

[0007] The technical solution adopted by the present invention is: a variable pitch multi-nozzle calibration method, which is applied to a machine vision system. The machine vision system includes a mechanical axis motion module, a vision module, a controller, a nozzle, a top-down camera, and a bottom-up camera. The vision module is signal-connected to the controller, and the top-down camera and the bottom-up camera are signal-connected to the vision module. The nozzle and the top-down camera are arranged on the mechanical axis motion module. A variable pitch multi-nozzle calibration method includes the following steps:

[0008] a. Top-down camera calibration: The vision module notifies the controller to pick up the calibration board; the calibration board is placed on the calibration board fixing bracket above the bottom-up camera; the vision module then notifies the controller to move the top-down camera to the reference point of the nine-point calibration; the top-down camera moves to the next position of the nine-point calibration and returns the corresponding mechanical coordinates. The vision module collects an image and calculates the center pixel coordinates of the Mark point on the calibration board. Subsequently, the top-down camera moves to the other positions of the nine-point calibration in turn. The top-down camera moves a total of nine times. Each time it moves, it returns the corresponding mechanical coordinates. Each time it moves, the vision module collects an image and calculates the center pixel coordinates of the Mark point on the calibration board. Finally, an affine transformation matrix between the image coordinate system and the mechanical coordinate system of the top-down camera is calculated based on the nine groups of data. ;

[0009] b. Bottom-up camera calibration: The vision module notifies the controller to move the nozzle to the reference point of the nine-point calibration of the bottom-up camera; then the Mark point of the nozzle moves to the next position of the nine-point calibration and returns the corresponding mechanical coordinates. The vision module collects an image through the bottom-up camera and calculates the center pixel coordinates of the Mark point of the nozzle. Subsequently, the nozzle moves to the other positions of the nine-point calibration in turn. The nozzle moves a total of nine times. Each time it moves, the vision module collects an image through the bottom-up camera and calculates the center pixel coordinates of the Mark point of the nozzle. Finally, an affine transformation matrix between the image coordinate system and the mechanical coordinate system of the bottom-up camera is calculated based on the nine groups of data. ;

[0010] c. Calibration of the relationship between the optical center of the upward-facing camera and the center of the nozzle marking point: The vision module notifies the controller to fetch the calibration board; the calibration board is placed on the calibration board fixing bracket above the upward-facing camera; the vision module then notifies the controller to move the downward-facing camera to the photographing position and return the corresponding mechanical coordinates. The vision module controls the upward-facing camera and the downward-facing camera to take pictures respectively. Subsequently, the vision module notifies the controller to remove the calibration board and place it at the designated position; then the vision module notifies the controller to move the nozzle to the photographing position of the upward-facing camera. The vision module then controls the upward-facing camera to take a picture and calculates the pixel coordinates of the center of the nozzle Mark point. Finally, according to the data, the offset between the optical center of the downward-facing camera and the center of the nozzle Mark point is calculated; repeat this process multiple times to correspond the mapping relationship between the center of the marking point of each nozzle and the optical center of the upward-facing camera, obtaining multiple displacement offsets; save the multiple displacement offsets to the vision software through communication for subsequent quick use.

[0011] Further, the calculation method of the offset between the optical center of the downward-facing camera and the center of the nozzle Mark point is as follows: First, use the downward-facing camera and the upward-facing camera to respectively photograph the Mark points of the round hole calibration board. The mechanical axis motion module uses the gantry axis to calculate the pixel coordinates of the center of the calibration board Mark point. According to and , the pixel coordinates of the center of the calibration board Mark point can be mapped to the gantry axis coordinate system as and respectively; according to , the optical center coordinates of the downward-facing camera are mapped to the gantry axis coordinate system as ; then move the gantry axis a distance of so that the nozzle Mark point is within the field of view of the upward-facing camera and extends out for photographing. Calculate the pixel coordinates of the center of the nozzle Mark point. Then, according to , the center of the nozzle Mark point can be mapped to the gantry axis coordinate system as ; calculate the offset between the optical center of the downward-facing camera and the center of the nozzle Mark point : Therefore, the offset between the optical center of the downward-facing camera and the center of the nozzle Mark point: .

[0012] Further, place the round hole calibration board within the fields of view of the upward-facing camera and the downward-facing camera and photograph the round hole marking points respectively. Then remove the calibration board and move the nozzle to the field of view of the upward-facing camera in sequence and photograph the nozzle marking points. During this period, it is necessary to ensure that the calibration board plane, the Tray disk plane, and the plane when the nozzle extends out for photographing are all in the same plane.

[0013] The beneficial effects of the present invention are as follows: Since the present invention adopts a variable pitch multi-nozzle calibration method, the variable pitch multi-nozzle calibration method realizes obtaining the offset coordinates of multiple nozzles from the coordinates obtained by taking a single photo, enabling all multiple nozzles to perform blind picking and placing without the need for visual re-guidance. The method is simple and efficient, providing technical support for actual production. For the existing manual multi-camera calibration, the calibration mapping matrices of each camera are separately recorded through 9-point calibration of each camera. However, this manual calibration method will cause a large amount of cumulative errors due to manual movement, and it is difficult to eliminate the errors. Manually calibrating and saving data point by point increases the complexity of the steps. Due to the problem of manual use differences, the accuracy is low, and it greatly increases the workload and fatigue degree of the users. The calibration method of the present invention is simple and convenient to use, with one-key calibration, reducing the operation requirements for the users; the fully automatic calibration process automatically gives the required movement amount and rotation angle, automatically controls to reach the required point for taking photos, transmitting data, and performing calculations, ensuring a stable supply amount and solving the cumbersome manual steps of manual calibration. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 is the movement path of the manipulator;

[0015] Figure 2 is the movement path of the image marking point;

[0016] Figure 3 is the schematic diagram of the calibration of the overhead camera;

[0017] Figure 4 is the schematic diagram of the calibration of the bottom-up camera;

[0018] Figure 5 is the schematic diagram of the relationship between the optical center of the bottom-up camera and the center of the nozzle marking point;

[0019] Figure 6 is the calculation schematic diagram of the relationship between the optical center of the bottom-up camera and the center of the nozzle marking point;

[0020] Figure 7 is the calibration flow chart of the overhead camera;

[0021] Figure 8 is the calibration flow chart of the bottom-up camera;

[0022] Figure 9 is the calibration of the relationship between the optical center of the bottom-up camera and the center of the nozzle marking point. DETAILED DESCRIPTION OF THE INVENTION

[0023] The main objective of the research on the automatic loading and calibration system is to obtain images from industrial cameras and calculate the position information of objects in the environmental space. The camera is a mapping from image coordinates to mechanical axis coordinates, and the mapping relationship between the two is determined by the relative position between the camera and the mechanical axis, that is, the conversion relationship between the camera coordinate system and the manipulator coordinate system.Figure 1 It is the moving path of the manipulator for the nine-point calibration method, Figure 2 and it is the pixel coordinates of the center of the corresponding image marked point. The manipulator moves sequentially from the reference point according to the Figure 1 path shown. Each time it reaches a position, the camera takes a photo, and then calculates the pixel coordinates of the center of the marked point. After all 9 points are completed, the affine transformation matrix between the pixel coordinate system and the mechanical coordinate system is calculated.

[0024] 1. Calibration of the downward camera

[0025] As Figure 3 shown, move the downward camera above the round hole calibration plate and move sequentially within the camera's field of view according to the order of the nine-point calibration method. Each time it moves to a position, the camera takes a photo and calculates the pixel coordinates of the center of the round hole on the corresponding calibration plate. Finally, 9 pixel coordinates are obtained.

[0026] Finally, using the 9 pixel coordinates and their corresponding mechanical coordinates, calculate the affine transformation matrix between the image coordinate system of the downward camera and the mechanical coordinate system of the gantry axis .

[0027] 2. Calibration of the upward camera

[0028] As Figure 4 shown, move the suction nozzle to within the field of view above the upward camera, and move the suction nozzle Mark point to the specified positions sequentially according to the order of the nine-point calibration method.

[0029] Each time it moves to a position, the camera takes a photo and calculates the pixel coordinates of the center of the corresponding suction nozzle marked point. Finally, using the 9 pixel coordinates and the corresponding mechanical coordinates, calculate the affine transformation matrix between the image coordinate system of the upward camera and the mechanical coordinate system of the gantry axis .

[0030] 3. Calibration of the relationship between the optical center of the upward camera and the center of the suction nozzle marked point

[0031] As Figure 5 shown, place the round hole calibration plate within the fields of view of the upward camera and the downward camera and take photos of the round hole marked points respectively, then remove the calibration plate, move the suction nozzle sequentially into the field of view of the upward camera and take photos of the suction nozzle marked points. During this period, it is necessary to ensure that the calibration plate plane, the Tray plate plane, and the plane when the suction nozzle extends for taking photos are all in the same plane.

[0032] As Figure 6 shown, first use the downward camera and the upward camera to take photos of the Mark points on the round hole calibration plate respectively, calculate the pixel coordinates of the center of the Mark points on the calibration plate. According to and , the pixel coordinates of the center of the Mark points on the calibration plate can be mapped to the gantry axis coordinate system and are respectively and According to Mapping the coordinates of the optical center of the overhead camera to the gantry axis coordinate system gives .

[0033] Then move the gantry axis by a distance of such that the nozzle Mark point is within the field of view of the overhead camera and extends for taking a picture. Calculate the pixel coordinates of the center of the nozzle Mark point. Then, according to the center of the nozzle Mark point can be mapped to the gantry axis coordinate system as .

[0034] Now, it is necessary to calculate the relationship between the optical center of the overhead camera and the center of the nozzle Mark point, that is, the one represented by the red line . From Figure 6 it can be known that: Therefore, the offset between the optical center of the overhead camera and the center of the nozzle Mark point: .

[0035] Repeat this process multiple times to correspond the mapping relationship between the center of the marked point of each nozzle and the optical center of the overhead camera, obtaining 8 displacement offsets. Save the 8 displacement offsets to the vision software through communication for subsequent rapid use.

[0036] Although the embodiments of the present invention are described with actual solutions, they do not constitute a limitation to the meaning of the present invention. For those skilled in the art, modifications to its implementation solutions and combinations with other solutions are obvious according to this specification.

Claims

1. A variable pitch multi-nozzle calibration method, which is applied to a machine vision system. The machine vision system includes a mechanical axis motion module, a vision module, a controller, a nozzle, a top-down camera, and a bottom-up camera. The vision module is signal-connected to the controller, and the top-down camera and the bottom-up camera are signal-connected to the vision module. The nozzle and the top-down camera are arranged on the mechanical axis motion module, and it is characterized in that: A variable pitch multi-nozzle calibration method includes the following steps: a. Calibration of the downward camera: The vision module notifies the controller to fetch the calibration board; the calibration board is placed on the calibration board fixing bracket above the upward camera; the vision module then notifies the controller to move the downward camera to the reference point for nine-point calibration; the downward camera moves to the next position for nine-point calibration and returns the corresponding mechanical coordinates. The vision module collects images and calculates the pixel coordinates of the center of the Mark points on the calibration board. Subsequently, the downward camera moves to the other positions for nine-point calibration in turn. The downward camera moves a total of nine times. Each time it moves, it returns the corresponding mechanical coordinates. Each time it moves, the vision module collects images and calculates the pixel coordinates of the center of the Mark points on the calibration board. Finally, an affine transformation matrix between the image coordinate system and the mechanical coordinate system of the downward camera is calculated based on the nine groups of data. ; b. Calibration of the upward-facing camera: The vision module notifies the controller to move the nozzle to the reference point for the nine-point calibration of the upward-facing camera. Then, the nozzle Mark point is moved to the next position of the nine-point calibration, and the corresponding mechanical coordinates are returned. The vision module captures an image through the upward-facing camera and calculates the center pixel coordinates of the nozzle Mark point. Subsequently, the nozzle is moved to the other positions of the nine-point calibration in sequence. The nozzle moves a total of nine times. Each time the nozzle moves, the vision module captures an image through the upward-facing camera and calculates the center pixel coordinates of the nozzle Mark point. Finally, an affine transformation matrix between the image coordinate system and the mechanical coordinate system of the upward-facing camera is calculated based on the nine sets of data. ; c. Calibration of the relationship between the optical center of the upward-facing camera and the center of the nozzle marking point: The vision module notifies the controller to fetch the calibration plate; the calibration plate is placed on the calibration plate fixing bracket above the upward-facing camera; the vision module then notifies the controller to move the downward-facing camera to the photographing position and return the corresponding mechanical coordinates. The vision module controls the upward-facing camera and the downward-facing camera to take pictures respectively. Subsequently, the vision module notifies the controller to remove the calibration plate and place it at the designated position; then the vision module notifies the controller to move the nozzle to the photographing position of the upward-facing camera. The vision module then controls the upward-facing camera to take a picture and calculates the pixel coordinates of the center of the nozzle Mark point. Finally, the offset between the optical center of the downward-facing camera and the center of the nozzle Mark point is calculated based on the data; repeat multiple times to correspond the mapping relationship between the center of the marking point of each nozzle and the optical center of the upward-facing camera to obtain multiple displacement offsets; save the multiple displacement offsets to the vision software through communication for subsequent rapid use; The calculation method for the offset between the optical center of the downward-facing camera and the center of the nozzle Mark point is as follows: First, use the downward-facing camera and the upward-facing camera to respectively capture the Mark points on the circular hole calibration plate. The mechanical axis movement module uses a gantry axis to calculate the pixel coordinates of the center of the Mark point on the calibration plate. According to and , the pixel coordinates of the center of the Mark point on the calibration plate can be mapped to the gantry axis coordinate system as and respectively; According to , the optical center coordinates of the downward-facing camera are mapped to the gantry axis coordinate system as ; Then move the gantry axis a distance of so that the nozzle Mark point is within the field of view of the upward-facing camera and extends for taking a picture. Calculate the pixel coordinates of the center of the circle of the nozzle Mark point, and then according to , the center of the nozzle Mark point can be mapped to the gantry axis coordinate system as ; Calculate the offset between the optical center of the downward-facing camera and the center of the nozzle Mark point : Therefore, the offset between the optical center of the downward-facing camera and the center of the nozzle Mark point: .

2. A variable pitch multi-nozzle calibration method according to claim 1, characterized in that: Place the round hole calibration plate within the fields of view of the upward-facing camera and the downward-facing camera and photograph the round hole marking points respectively. Then remove the calibration plate and move the nozzles to the field of view of the upward-facing camera in sequence and photograph the nozzle marking points. During this period, it is necessary to ensure that the calibration plate plane, the Tray plate plane, and the plane when the nozzle extends for photographing are all in the same plane.

Images