Research On Industrial Robot Crawling System Based On Machine Vision And Image Delay Compensation

With the upgrading of industrial technology,just a single industrial robot has been difficult to meet the needs of diversified and flexible production.The composite robotic systems which have multiple sensors can process multi-sensor data fusion are increasingly becoming the mainstream of the industry.Among them,the visual servo system which combined machine vision and industrial robots can achieve autonomous guidance of the machine,so it has broad application prospects.In this dissertation,a common industrial robot capture applications as a starting point,the visual capture system and its existing image delay problems are studied,the specific work is as follows:(1)By analyzing the grasping process of robotic binocular capture system,the work area of the system is divided into recognition area and grasping area,and the corresponding theoretical research is completed.(2)Designed the perception algorithm of the recognize zone to complete the tasks which provide initial operational data for follow-up robot control.Using Hu invariant moment to identify three different shapes included rectangle,square and triangle.Using long axis angles to recognize object's pose.Using the principle of triangulation to obtain the distance between the object and the end effector in real time.Using self-designed hybrid tracking algorithm to track the object and get it velocity.(3)Designed and programmed the PID control algorithm to control the robot completed the object capture.Aiming at the problem of image transmission and processing delay in visual servo system,a time-delay compensator based on BP neural network is proposed.(4)Designed and constructed a rectangular coordinate system robot binocular vision capture system,completed the hardware selection and interactive system programmed.Completed the physical experiments and simulation experiments,verified the validity of the algorithm and achieved good experimental results.