Font Size: a A A

Research On Key Technologies Of Intelligent Follower Robot

Posted on:2022-05-25Degree:MasterType:Thesis
Country:ChinaCandidate:Y Y ZhengFull Text:PDF
GTID:2518306557471534Subject:Logistics Engineering
Abstract/Summary:PDF Full Text Request
In recent years,with the improvement of computer vision technology,intelligent follow-up robots have also been developed rapidly,greatly improving people's lives.At present,intelligent follow-up robot systems mainly include various follow-up systems based on infrared or ultrasonic waves.In the infrared-based tracking system,it has strong adaptability to complex and changeable environments,but the measurement accuracy is low;the ultrasonic-based tracking system can quickly and easily process data,and has strong stain resistance,but its research cost is relatively high,The accuracy is low.The realization of the intelligent follower robot studied in this article relies on good visual target detection technology.However,many target detection networks have problems such as large model scale and long calculation process when combined with embedded devices.The real-time performance of the following system is severely restricted.In order to solve the above problems,this paper studies an intelligent follower robot system based on the YOLO-LITE algorithm and the SC stereo matching algorithm,combined with hardware devices such as binocular cameras,to realize the intelligent follower of the robot and ensure the real-time performance of the system.The main work of this paper is as follows:1.The software and hardware design of the intelligent follower robot system.The system is divided into four functional modules: camera processing,target detection,camera ranging and vehicle body movement.In order to realize the intelligent follow-up of the robot,firstly,after each frame of video image is acquired by the binocular camera,it is calibrated and stereo-corrected.And so on;secondly,the system uses the YOLO-LITE convolutional neural network model combined with the raspberry party to identify and locate the target in the image;thirdly,the SC-based stereo matching algorithm is proposed to calculate the distance measurement between the camera and the target.Finally,the L298 N chip is used to drive the robot to perform corresponding movements in combination with the center offset and steering indicators such as steering indicators.Through the transfer of related parameters,each module can run in an orderly manner.2.The realization and optimization of the intelligent follower robot system.In the implementation process,when judging the steering of the following car,the single steering index is improved,and a steering judgment method suitable for the following system studied in this paper is proposed,and then from the three aspects of the real-time performance of the car,the detection accuracy of the car and the following distance of the car Measure and analyze the following performance.The experimental results show that in indoor scenes,the detection accuracy of the following system is more than 80%,the delay time is only 20 ms,and the detection speed also reaches 20 fps,which maintains a good balance between the real-time performance of the car and the detection accuracy of the car.
Keywords/Search Tags:YOLO-LITE algorithm, binocular camera, stereo matching, turning indicators, robot follow
PDF Full Text Request
Related items