Research On Golf Robotic Caddie Visual Tracking System

The autonomous robot is an integrated system combined with a variety offunctions like context-aware, dynamic decision-making and planning, behavior controland perform, reflecting the level of development of information technology andartificial intelligence techniques. With the development of computers and artificialintelligence techniques, in some cases finishing jobs by autonomous robots instead ofhumans, not only can protect the personal safety, reduce labor intensity and ensure thatthe personal privacy space, but also can improve the economic benefits and bringhuman society great conveniences.In this paper, we focuses on a kind of service robot called caddy robot, As thealternative to golf caddie master tracking study of visual tracking system. In commonlyused tracking method, the vision-based robot, with considerable intelligence, canprovide rich, comprehensive environmental information, and distinguish between theowner from any other person. This is based on Golf child robot to research issuesrelated to the environment and human visual perception about Golf Courses and humanbodies, image processing, target detection, target tracking, and decision-making control,able to independently follow the owner.In this paper, taking factors such as deformation and occlusion of the human bodyin motion into consideration, we analyze the traditional moving target detection andtracking algorithm and research the golf course environment image processingalgorithms. Algorithm abandons the start of human manual intervention and developsautomatic start solutions; by comparing the traditional moving target detection methodsand analyzing the golf course environment, we develop a method, combining withhuman external moments and color and having the ability of automatic distinguishingthe objectives from the non-human interferences, to detect moving targets, and thismethod improves the detection effects; in order to solve the problems like deformationand occlusion owing to human motion, we propose an improved mean shift algorithm,combining color histogram intersection, for human tracking. The results of simulationhave shown that the algorithm has good matching accuracy, real time and stability,effective detection and tracking of human movement in the image plane.This paper utilizes SOPC system to achieve Human Movement detection andtracking functions, based on image processing algorithms for the golf courseenvironment. Compared to the composite structure, we reduce the use of chips to reduce costs in the realization and the probability of failure of the hardware in reliability; usingthe FPGA logic of SOPC system we research the design of the gray pixels of real-timeprocessing, and the parallel storage of original image and the edge image and otherhardware acceleration methods to improve system performance.After debugging, the system can achieve the desired goal, detecting and tracking ofhuman movement, operating stable, and meet the real-time requirements. This paperconcludes the methods of tracking the movement of human bodies on the golf coursesand summarizes their development prospects and outlooks.