Three-axis Pan-tilt Design And Moving Target Detection And Tracking Algorithm Research

With the rapid development of computer technology,machine vision and artificial intelligence,research on PTZ technology has become a hot topic in the fields of intelligent control and motion target detection and tracking.According to this paper mainly aimed at the improvement of traditional two-degree-of-freedom pan-tilt,increased the pan-tilt camera's own rotation and designed a three-axis pan-tilt based on the two-axis platform.Also,aimed at researching and improving the target detection and tracking algorithm,and enhancing the real-time and robustness of the PTZ's moving target tracking algorithm through the three-axis pan-tilt.The characteristics of the traditional two-degree-of-freedom platform were analyzed and designed for structure and the system hardware circuit of three-axis pan-tilt mechanical through this article.The three-axis pan-tilt adopts the main frame structure of the stainless steel cantilever beam,and the brushless DC motor was used as the motor for the PTZ.The STM32F429VG was adopted as the three-axis pan-tilt system main control(MCU).The MPU6050 gyroscope feedbacks the speed of the pan/tilt motor,the motor encoder feedbacks the absolute position of the pan/tilt motor in real time,and the 24V lithium battery was used as the system power supply.The control area network(CAN)was adopted as the bus cascade control of three-axis pan-tilt motor.The double closed-loop PID was used as the main control algorithm,while the incremental PID was used in the inner ring(rate loop)and the position PID was adopted as in the outer loop(position loop).To enhance system robustness,smooth the gyroscope data,and improve the double closed loop PID by combining the position loop output and the gyro measurement data with a Kalman filter.Based on the three-axis pan-tilt,the open source computer vision library(OpenCV3.3)was used under Visual Studio 2017 to study and compare the main moving object detection algorithms of inter-frame difference method,optical flow method and background subtraction.Focusing on the improvement of MeanShift moving target tracking algorithm,the Adaptive Window MeanShift(CamShift)algorithm was added to Harris corner detection and merged with Kalman filter.The reliability and stability of the improved algorithm were verified through experimental comparison and analysis,and the upper computer of Labview2015 was designed to visually display the results of joint debugging.Mainly from theoretical analysis to experimental comparison,based on three-axis pan-tilt and open source vision platform for target detection and tracking algorithm related research were demonstrated through this article.The progress and achievements made in this thesis have certain theoretical support and practical reference value for the study of PTZ control technology and motion target detection and tracking in machine vision...