Remote Patrol Alarm Robot Based On Machine Vision

With the rapid development of artificial intelligence,robot research has made great progress.Many different types of automatic robots are gradually replacing human,which not only improves work efficiency,but also reduces production and labor costs.However,in the process of practical application,due to the limitations of the terrain environment,the interference of the external environment,such as illumination changes,occlusion,video background jitter and so on,the detection and tracking of the target in dynamic background is still a significant research topic.In order to solve the above problems,this paper designs an intelligent patrol robot which combines automatic patrol and artificial remote control.Aiming at the complex terrain environment,the robot moves in the way of crawler.Due to the increase of the contact area with the ground,the crawler robot has stronger trafficability in the process of movement.In view of the slow processing speed and low correct recognition rate of video in dynamic background,this paper proposes a target detection algorithm combining super-pixel algorithm and dynamic compensation algorithm.Firstly,the super pixel algorithm is used to process the video image frame by frame to classify the content in the video image.This process greatly reduces the number of pixels to be processed in the image,while retaining the key information in the video image.Then,the feature points of the image are extracted and matched.After finding the key information of the feature points in the image,the two frame image is registered In order to eliminate the influence of background motion on target detection.Finally,the target contour is described by frame difference operation and image morphology operation.In order to solve the problem of target loss caused by occlusion in dynamic tracking,this paper proposes an adaptive threshold tracking algorithm which combines Kalman filter and CAMSHIFT algorithm.The algorithm detects the occlusion of the tracking target and calculates the occlusion coefficient,then adjusts and predicts the target tracking position according to the calculated occlusion coefficient,thus improving the target tracking position Tracking success rate,achieve better results.In order to facilitate the use ofoperators,this paper developed a set of video monitoring system and mobile phone remote control software.The actual operators can observe the real-time images transmitted by the robot at the PC end,and remote control the robot at the same time,which has better practicability.Finally,the performance of each part of the robot is tested through a series of experiments,including the stability of video transmission,the feasibility of remote control operation,and the anti-interference of video detection and tracking algorithm under the influence of external environment.The experimental results show that the design of the robot has better practicability,and the detection and tracking of the target have better stability in the face of external interference.