An automatic detection method based on point cloud data

Abstract

The present invention discloses an automatic detection method based on point cloud data, which includes a working object, a laser scanner, a manipulator, and a probe board process parameter library, all of which are electrically connected to a programmable logic controller. It collects the cross-sectional surface profile information of a number of components to be measured for grasping a probe board matching the component to be measured, collects the three-dimensional point cloud information of the working object to determine the first motion trajectory, collects the pose information of the manipulator grasping the probe board to determine the second motion trajectory, uses the TOF method or the triangulation ranging method to measure the distance between the manipulator and the working object, accurately matches the information of the first motion trajectory and the second motion trajectory, inputs the matched information into the programmable logic controller, and drives the manipulator to perform detection operations. This method has good practicability and can realize the simultaneous detection of different products.

Description

Technical Field

[0001] The present invention belongs to the technical field of electronic device detection, and particularly relates to an automatic detection method based on point cloud data. Background Art

[0002] There are many types of electrical components, and there are also various classification methods. Classified by their operating mode, they can be divided into automatic switching electrical appliances and non-automatic switching electrical appliances; classified by their functions and positions in the control circuit, they can be divided into control electrical appliances and power distribution (protection) electrical appliances; classified by their execution functions, they can be divided into electrical appliances with contacts and without contacts, and so on. An electrical device that automatically or manually connects or disconnects a circuit according to a specific external signal to achieve control of the circuit or non-electrical object is called an electrical appliance.

[0003] Electrical components are important parts of electronic products. Before leaving the factory, the performance of electrical components must be detected to timely discover quality problems. In the prior art, during the process of detecting the performance of electrical components on an automated production line, it is necessary to manually place the finished components on the conveying track, move and position the finished components, and make the probe board contact the finished components to achieve detection.

[0004] However, during the implementation of the technical solution of the present invention by the inventors of the present application, it is found that the above technology has at least the following technical problems:

[0005] Since the selection of the probe board is determined according to the shape of the measured points of the finished component, the probe board and the finished component need to correspond one by one. Therefore, only a single finished component can be detected. When detecting different components, it is necessary to manually disassemble and replace the probe board, and then make the probe board contact the finished component to achieve detection, which affects the detection efficiency of the electrical components on the production line. Summary of the Invention

[0006] Aiming at the deficiencies of the prior art, the embodiments of the present application provide an automatic detection method based on point cloud data, which solves the technical problem in the prior art that only a single finished component can be detected, and it is necessary to manually disassemble and replace the probe board when detecting different finished components. This method has good practicability, can realize the simultaneous detection of different products, and improves the detection efficiency.

[0007] An embodiment of this application provides an automatic detection method based on point cloud data, which includes a work object, a laser scanner, a manipulator, and a probe board process parameter library, all of which are electrically connected to a programmable logic controller. It collects the cross-sectional surface profile information of several components to be measured for grasping a probe board that matches the component to be measured, collects the three-dimensional point cloud information of the work object to determine the first motion trajectory, collects the pose information of the manipulator grasping the probe board to determine the second motion trajectory, uses the TOF method to measure the distance between the manipulator and the work object, and accurately matches the information of the first motion trajectory and the second motion trajectory. The matched information is input into the programmable logic controller to drive the manipulator to perform detection operations. This method includes the following steps:

[0008] Step a: Move the work object into the area of the laser scanner, and obtain the point cloud data of the work object and the cross-sectional profile information by scanning the work object and the cross-section with green connection terminals;

[0009] Step b: Analyze and obtain the pose information of the work object, and input the point cloud data and the surface profile information of the cross-section into the programmable logic controller for filtering processing using a voxel filter;

[0010] Step c: Extract features from the point cloud segmentation data of the point cloud data by a texture-based method;

[0011] Step d: Identify the object from the extracted point cloud data by the SURF method and convert it into three-dimensional point cloud information;

[0012] Step e: Determine the first motion trajectory according to the three-dimensional point cloud information of the identified work object;

[0013] Step f: The programmable logic controller drives the manipulator to grasp the probe board in the probe board process parameter library according to the surface profile information of the cross-section, and determines the second motion trajectory according to the pose information of the probe board, the current position and current pose of the manipulator, as well as the holding position and holding pose;

[0014] Step g: Use the TOF (time of flight) method or the triangulation method to measure the distance between the manipulator and the work object;

[0015] Step h: Register the information of the first motion trajectory and the second motion trajectory by the nearest neighbor search method, and input the registered information into the programmable logic controller to drive the manipulator to perform detection operations;

[0016] Step i: The force sensor detects the contact information between the probe plate grasped by the manipulator and the connecting terminal, and feeds back its signal to the programmable logic controller.

[0017] Preferably, the embodiment of the present application further discloses that the working object in step a includes several components to be tested, and the collected point cloud data is the position and posture information of any green connection terminal of the several components to be tested.

[0018] Preferably, the embodiment of the present application further discloses that any cross section of the component to be tested is provided with a positioning groove, a plurality of green connection terminal rows are inserted or crimped into the positioning groove, and the connection terminals between the plurality of green connection terminal rows are equidistantly distributed.

[0019] Preferably, the embodiment of the present application also discloses that the programmable logic controller includes an information processing unit for preprocessing point cloud data, a storage unit for storing three-dimensional point cloud data, and a power generator, and the power generator is configured to control the movement of the manipulator via the power generated.

[0020] Preferably, the embodiment of the present application also discloses a texture-based method in step c for extraction through shape and color features.

[0021] Preferably, the embodiment of the present application also discloses that the TOF method described in step g uses the flight time of the laser pulse signal to calculate the distance. Each time the laser is emitted and received, the data of a point can be collected; the triangulation ranging method uses the light spot formed by the laser on the surface of the object to perform triangulation calculation on the position of the optical sensor inside the laser scanner to measure the distance.

[0022] Preferably, the embodiment of the present application further discloses that a clamping component is provided at the end of the manipulator in step g, a force sensor is provided on one side of the clamping component, and the force sensor is electrically connected to the programmable logic controller.

[0023] Preferably, an embodiment of the present application also discloses that when the information of the first motion trajectory and the second motion trajectory in step h do not match, it is necessary to calculate the first distance between the first motion trajectory and the second motion trajectory in the circumferential direction, and the programmable logic controller drives the manipulator to move the first distance to restore it to the first motion trajectory.

[0024] Preferably, the embodiment of the present application further discloses that the work object is placed on a mobile trolley with multi-layer shelves, and the shelves are provided with a material placement tray for fixing the work object, and the bottom of the mobile trolley is provided with universal wheels and brakes.

[0025] The technical solution provided by the embodiments of the present application has at least the following significant technical effects:

[0026] (1) Collect the cross-sectional surface profile information of several components to be measured for grasping a probe board that matches the component to be measured. Collect the three-dimensional point cloud information of the working object to determine the first motion trajectory. Collect the pose information of the manipulator grasping the probe board to determine the second motion trajectory. Use the TOF method to measure the distance between the manipulator and the working object, and register the information of the first motion trajectory and the second motion trajectory. Input the registered information into the programmable logic controller to drive the manipulator to perform the detection operation, enabling the simultaneous detection of different products and also the detection of a single item, thus improving the detection efficiency.

[0027] (2) The three-dimensional point cloud information is obtained through a series of operations such as preprocessing, segmentation, extraction, transformation, and recognition of the collected point cloud data. By registering the information of the first motion trajectory and the second motion trajectory, precise contact between the probe board and the connection terminals can be achieved, avoiding position offset or poor contact caused by vibration or other factors during other mechanical motion processes. Description of the Drawings

[0028] Figure 1 It is a flowchart of an automatic detection method based on point cloud data in an embodiment of the present application. Embodiment

[0029] The embodiment of the present application provides an automatic detection method based on point cloud data, which solves the technical problem in the prior art that only single finished component detection can be achieved, and manual disassembly and replacement of the probe board are required when detecting different finished components.

[0030] The overall idea of the technical solution in the embodiment of the present application to solve the above problems is as follows: Collect the cross-sectional surface profile information of several components to be measured for grasping a probe board that matches the component to be measured. Collect the three-dimensional point cloud information of the working object to determine the first motion trajectory. Collect the pose information of the manipulator grasping the probe board to determine the second motion trajectory. Use the TOF method to measure the distance between the manipulator and the working object, and accurately match the information of the first motion trajectory and the second motion trajectory. Input the matched information into the programmable logic controller to drive the manipulator to perform the detection operation, enabling the simultaneous detection of different products and also the detection of a single item.

[0031] As Figure 1 shown, the embodiment of the present application provides an automatic detection method based on point cloud data, including a working object, a laser scanner, a manipulator, and a probe board process parameter library, which are respectively electrically connected to a programmable logic controller. The detection method includes the following steps:

[0032] Step a: Move the workpiece into the area of the laser scanner. By scanning the workpiece and the cross-section with green connection terminals, obtain the point cloud data of the workpiece and the cross-sectional contour information. The workpiece includes several components to be measured, and the collected point cloud data is the pose information of any green connection terminal of the several components to be measured. Any cross-section of the component to be measured is provided with a positioning groove, and several green connection terminal rows are inserted or crimped in the positioning groove, and the connection terminals between the several green connection terminal rows are equally spaced. The workpiece can also be placed on a mobile trolley with multiple layers of shelves. The shelf is provided with a feeding tray for fixing the workpiece, and the bottom of the mobile trolley is provided with universal wheels and brakes to facilitate the processing of the workpiece and reduce the labor intensity of workers.

[0033] Step b: Analyze and obtain the pose information of the workpiece, and input the point cloud data and the surface contour information of the cross-section into the programmable logic controller for filtering processing through a voxel filter.

[0034] Step c: Extract the features of the point cloud segmentation data from the point cloud data through a texture-based method. The features are extracted through shape and color information.

[0035] Step d: Identify the object from the extracted point cloud data through the SURF method and convert it into three-dimensional point cloud information.

[0036] Step e: Determine the first motion trajectory according to the three-dimensional point cloud information of the identified workpiece.

[0037] Step f: The programmable logic controller drives the manipulator to grab the probe board in the probe board process parameter library according to the surface contour information of the cross-section. Determine the second motion trajectory according to the pose information of the probe board, the current position and current pose of the manipulator, and the holding position and holding pose. The end of the manipulator is provided with a clamping component, and a force sensor is provided on one side of the clamping component, and the force sensor is electrically connected to the programmable logic controller. The programmable logic controller includes an information processing unit for preprocessing point cloud data, a storage unit for storing three-dimensional point cloud data, and a power generator, and the power generator is configured to control the movement of the manipulator via the generated power.

[0038] Step g: Measure the distance between the manipulator and the workpiece by using the TOF (time of flight) method or the triangulation method. This method calculates the distance by using the flight time of the laser pulse signal. Each time the laser emits and receives, one point of data can be collected. This ranging method can also be the triangulation method, which uses the position of the light spot formed by the laser on the surface of the object on the optical sensor inside the laser scanner for triangulation calculation to measure the distance.

[0039] Step h: Register the information of the first motion trajectory and the second motion trajectory through the method of nearest neighbor search. When the information of the first motion trajectory and the second motion trajectory does not match, it is necessary to calculate the first distance between the first motion trajectory and the second motion trajectory in the circumferential direction. The programmable logic controller drives the manipulator to move the first distance to restore it to the first motion trajectory, and inputs the registered information into the programmable logic controller to drive the manipulator to perform the detection operation; and input the registered information into the programmable logic controller to drive the manipulator to perform the detection operation.

[0040] Step i: The force sensor detects the contact information between the probe board grasped by the manipulator and the connection terminal, and feeds back its signal to the programmable logic controller.

[0041] Working process:

[0042] Move the work object into the laser scanner area. By collecting the cross-sectional surface profile information of several components to be measured, it is used to grasp the probe board that matches the component to be measured. Through processes such as filtering the point cloud data, point cloud segmentation, feature extraction, and point cloud data registration, collect the three-dimensional point cloud information of the work object to determine the first motion trajectory, collect the pose information of the manipulator grasping the probe board to determine the second motion trajectory, use the TOF method to measure the distance between the manipulator and the work object, and accurately match the information of the first motion trajectory and the second motion trajectory. Input the matched information into the programmable logic controller to drive the manipulator to perform the detection operation, realizing the simultaneous detection of different products, and can also realize the detection of a single product.

[0043] Although the preferred embodiments of the present invention have been described, once those skilled in the art learn the basic creative concepts, additional changes and modifications can be made to these embodiments. Therefore, the appended claims are intended to be construed to include the preferred embodiments as well as all changes and modifications that fall within the scope of the present invention.

[0044] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention also intends to include these modifications and variations.

Claims

1. An automatic detection method based on point cloud data, wherein, it includes a working object, a laser scanner, a manipulator, and a probe board process parameter library, which are respectively electrically connected to a programmable logic controller, collects cross-sectional surface contour information of a number of components to be measured for grasping a probe board matching the components to be measured, collects three-dimensional point cloud information of the working object to determine a first motion trajectory, collects pose information of the manipulator grasping the probe board to determine a second motion trajectory, accurately matches the information of the first motion trajectory and the second motion trajectory, and inputs the matched information into the programmable logic controller to drive the manipulator to perform a detection operation. This method includes the following steps: Step a: Move the working object into the area of the laser scanner, and obtain the point cloud data of the working object and the cross-sectional contour information by scanning the working object and the cross-section provided with green connection terminals; Step b: Analyze and obtain the pose information of the working object, and input the point cloud data and the surface contour information of the cross-section into the programmable logic controller for filtering through a voxel filter; Step c: Extract features of the point cloud data by a texture-based method for the point cloud data; Step d: Identify the object from the extracted point cloud data by a SURF-based method and convert it into three-dimensional point cloud information; Step e: Determine the first motion trajectory according to the three-dimensional point cloud information of the identified working object; Step f: The programmable logic controller drives the manipulator to grasp a probe board in the probe board process parameter library according to the surface contour information of the cross-section, and determines the second motion trajectory according to the pose information of the probe board, the current position and current pose of the manipulator, as well as the grasping position and grasping pose; Step g: Measure the distance between the manipulator and the working object by the TOF method or the triangulation method; Step h: Register the information of the first motion trajectory and the second motion trajectory by the nearest neighbor search method, and input the registered information into the programmable logic controller to drive the manipulator to perform a detection operation; Wherein, when the information of the first motion trajectory and the second motion trajectory in step h does not match, it is necessary to calculate the first distance in the circumferential direction between the first motion trajectory and the second motion trajectory, and the programmable logic controller drives the manipulator to move the first distance to restore it to the first motion trajectory; Step i: The force sensor detects the contact information between the probe board grasped by the manipulator and the connection terminal, and feeds back its signal to the programmable logic controller.

2. The automatic detection method based on point cloud data according to claim 1, wherein, The working object in step a includes a number of components to be measured, and the collected point cloud data is the pose information of any green connection terminal of the number of components to be measured.

3. The automatic detection method based on point cloud data according to claim 2, wherein, Any cross-section of the component to be measured is provided with a positioning groove, and a number of green connection terminal rows are inserted or crimped in the positioning groove, and the connection terminals between the number of green connection terminal rows are equally spaced.

4. The automatic detection method based on point cloud data according to claim 1, wherein, In step b, the programmable logic controller includes an information processing unit for preprocessing point cloud data, a storage unit for storing three-dimensional point cloud data, and a power generator configured to control the movement of the manipulator via the generated power.

5. The automatic detection method based on point cloud data according to claim 1, wherein, In step c, the texture-based method is extracted through shape and color features.

6. The automatic detection method based on point cloud data according to claim 1, wherein, In step g, the TOF method calculates the distance using the flight time of the pulse signal of the laser. Each time the laser emits and receives, the data of one point can be collected; the triangulation method uses the position of the light spot formed by the laser on the surface of the object on the optical sensor inside the laser scanner for triangulation calculation to measure the distance.

7. The automatic detection method based on point cloud data according to claim 1, wherein, In step g, a clamping component is provided at the end of the manipulator, a force sensor is provided on one side of the clamping component, and the force sensor is electrically connected to the programmable logic controller.

8. An automatic detection method based on point cloud data according to claim 1, wherein, The working object is placed on a multi-layer shelf mobile trolley, and a feeding tray for fixing the working object is provided on the shelf. Universal wheels and brakes are provided at the bottom of the mobile trolley.

Images