3D Camera Laser Stripe Image Acquisition And Centerline Extraction

With the further development of industrial technology,the complexity of industrial products continues to increase,and the inspection methods of products in all walks of life have changed.The information that needs to be obtained is no longer satisfied with the two-dimensional image of the product,but puts forward requirements for its three-dimensional information.Three-dimensional measurement technology can complete the three-dimensional defect detection of products,and can greatly improve production efficiency and reduce quality control costs,which is of great significance to the improvement of industrial manufacturing technology.The product three-dimensional information acquisition system applied in industry has the characteristics of fast measurement speed and high degree of integration,and needs to adapt to the high-speed detection method in the industrial assembly line environment,which is an indispensable part of industrial manufacturing.This paper studies and builds a 3D camera system,which is mainly composed of an industrial pipeline simulation platform,a line laser generator,a processing board with an ARM+FPGA architecture ZYNQ chip as the core,and an industrial CMOS image sensor Python300,which realizes the extraction of the center line of the laser stripe image.The main work content is as follows:(1)Build an acquisition control system based on a pipeline platform to control the image sensor for imaging at equal displacement intervals.The assembly line platform moves the object under the lens through the conveyor belt,and outputs the displacement information of the conveyor belt to the ZYNQ chip through the encoder,and controls the image sensor to collect images of the complete object at equal intervals in the actual industrial production environment to ensure that the threedimensional information of the object is captured.Collect evenly.(2)On the basis of the original light stripe centerline extraction algorithm based on the gray barycenter method,its shortcomings are optimized and improved.After the image data is transmitted to the post ZYNQ chip,the ARM end configures high and low thresholds that match the environment to identify and intercept the line laser irradiation part of the image data stream,and then use dual RAM to ping-pong this part of the data.Finally,the centerline coordinates of the laser stripes on each column in a frame of image are calculated in parallel through an algorithm implemented on the FPGA,and these points are integrated and transmitted to the DDR(Double Data Rate)chip for storage in video stream format data.(3)Improve the speed of 3D information extraction,based on the ROI(Region Of Interest)function of the image sensor,complete the configuration of the specified ROI area in the entire laser fringe image acquisition system.Flexible selection of the ROI area can eliminate redundant information in the image to increase the acquisition speed,and the required acquisition range can be adjusted according to the actual scene.Finally,using the 3D camera system and the PC terminal for system debugging,the equal interval collection of laser fringe images,ROI configuration and the extraction of the center line of the line laser are completed.The experimental results show that the hardware design of the laser fringe image centerline extraction algorithm is fast,automatic,and stable.After calibration of the sampling platform and 3D reconstruction,it can reflect the 3D information of the object.