Research And Implementation Of Laser Spot Tracking System Based On Machine Vision

Large-scale measurement technology is widely used in the field of industrial manufacturing.It is an important criterion for judging the level of a country's manufacturing industry.High-precision automated large-scale measurement is the current research hotspot and difficulty in the field of manufacturing measurement.Among the many large-scale measurement technologies,the theodolite system is one of the most widely used non-contact measurement systems.It has the characteristics of high accuracy and flexible system construction.However,the traditional theodolite system relies on manual targeting,and the complicated operation and low degree of automation seriously affect the measurement efficiency.In order to realize an automated theodolite measurement system,this subject is based on an electronic theodolite and builds a vision-based theodolite tracking system.The system uses laser spots as recognition tracking objects,uses image recognition technology to replace human eye aiming,and drives the tracking motor with the deviation of the laser spot from the center of the field of view to realize automatic tracking of the spot target with aiming.In order to achieve the goal of the subject,this paper mainly designs a laser spot tracking system based on PC image processing,industrial camera image acquisition and PLC motion control.The implementation process starts from the five aspects of system hardware construction,target detection,target tracking,PTZ control and software design.The main research content and work are as follows:1.System hardware construction.According to the demand analysis and design goals of the laser spot tracking system,the overall design of the system is proposed,and the hardware design of the system is completed,including the theodolite gimbal,industrial camera,lens,stepper motor,PLC and other hardware equipment.2.Spot detection algorithm.According to the characteristics of the spot image imaging,the characteristics of several common threshold segmentation and edge extraction algorithms are analyzed,and methods based on bilateral filtering,iterative threshold segmentation,morphological filtering,edge detection,and circumscribed rectangle are proposed to achieve the detection and positioning of the target spot.3.Spot tracking algorithm.In order to meet the real-time requirements of the system,the correlation filtering tracking algorithm is used.Aiming at the problem that the traditional correlation filtering tracking algorithm will cause tracking failure in the case of target occlusion and fast motion,a tracking algorithm combining particle filtering and correlation filtering is proposed.This method optimizes the target detection mechanism of the correlation filter,improves the robustness of the algorithm,and verifies the effectiveness of the algorithm through experiments.4.PTZ control.According to the real-time requirements of the system,the image-based visual servo control algorithm is used,and the proportional relationship between the pixel miss distance of the camera and the motor control pulse quantity is obtained through the experimental method.Aiming at the problem that the light spot exits the camera's field of view and causes system tracking failure,a trajectory prediction algorithm based on Kalman filtering is proposed to increase the stability of the light spot tracking system.5.System software design: Complete the program design of target detection,target tracking,PTZ control,human-computer interaction interface,communication between upper computer and lower computer,and realize the experimental verification of the laser spot tracking system.