Research And Design Of Marine Ranch Anti-theft System Based On Improved DPM Algorithm

With the continuous deterioration of the marine environment,traditional marine fisheries have gradually developed towards the cultivation of marine pasture technology.Marine pasture farming is generally carried out by cage culture in a sea area about 10 km from the coast.Breeding in this way can solve the bottleneck problem of traditional fisheries,but because the cage is far away from the coast,farming in this way is easy to be stolen.To solve this problem,this paper designs an ocean pasture based on image processing and recognition.surveillance system.The main research contents and innovations are as follows:Design marine ranch monitoring system.The system mainly consists of three parts: the front-end information acquisition part,the image processing and identification part,and the remote communication part.The front-end information acquisition part is completed by the camera,the image processing and recognition part is completed by the DPM algorithm,and the remote communication part is completed by the GPRS wireless module.The pedestrian detection algorithm based on DPM algorithm is studied.Considering that the original DPM algorithm can not meet the real-time performance of the anti-theft system,this paper studies the improved DPM algorithm.The improved algorithm mainly uses the single-point computing model algorithm instead of the dynamic programming algorithm,the binary layer fast layer localization algorithm to search the pyramid,and the intermediate threshold to optimize the algorithm,and the improved DPM algorithm is applied to the campus.The detection of the human body and the sea body in different environments has realized the automatic monitoring and anti-theft function of the marine pasture.The marine pasture monitoring system was tested.Firstly,in order to analyze the performance of the method,the human body database in different environments of sea body and campus is established.The pedestrian detection algorithm is tested with VOC data set and self-built data set respectively and compared with the original algorithm.The test results show that the improved results show that the improved The algorithm greatly improves the detection efficiency of the algorithm.Finally,the system was tested as a whole.The test results proved that the system basically met the real-time requirements of the marine ranch monitoring system,and the system was stable and reliable.This paper innovatively applies image processing and recognition technology to the marine ranch anti-theft system,realizing the automatic monitoring and anti-theft function of the marine ranch.The research results of this thesis lay a certain foundation for the further realization of the intelligent monitoring of the marine pasture,and have a good application prospect in the field of intelligent monitoring and theft.