Application Of Machine Vision In UAV Power Inspection

Nowadays,in the environment of stable and rapid economic development in China,all sectors of the market are also constantly developing,among which the development of the power industry is particularly significant,making our country's demand for power more and more.For the power industry,in order to ensure the safe operation of the circuit,the circuit inspection is very important.Because the traditional inspection method has some defects,specifically,the work efficiency is low,unsafe and so on.Therefore,in order to solve this problem,a new inspection method came into being,specifically,UAV power inspection,which not only improves the efficiency of work,but also the safety of the whole inspection work.Based on this advantage,it has been widely used,and its further research and development is also in progress.In this paper,the insulator tracking technology is studied under the environment of the rapid development of power inspection.Its main solution is the limitations of the new patrol technology,that is,obstacle avoidance and ranging of UAV.In this paper,image tracking of insulator is studied as followsThis paper introduces the theory of compressed sensing,including three aspects:signal reconstruction method,signal sparse representation and measurement matrix design.And the third aspect of the constraints of isometric conditions related analysis,the common design methods are described,and then the signal reconstruction method related analysis,mainly summarizes its advantages and disadvantages.It also analyzes the application of compressed sensing in the future development and video tracking,so as to provide theoretical basis for related research.This paper analyzes the insulator image processing and tracking algorithm used in UAV power inspection,and gives the method used in this paper.Finally,the algorithm test platform is built,including software and hardware,and the specific test environment is described.In the test of the algorithm,different video sequences are used,and the adaptability of the algorithm to the scale of insulator change is analyzed.