With the increase of service life,tower crane is prone to crack,oxidation corrosion and other problems,resulting in insufficient bearing capacity and serious harmful accidents.The traditional tower crane inspection relies on manual subjective observation,which has great limitations.Therefore,a more intelligent inspection method of tower crane is needed.Aiming at the problem of using UAV to inspect tower crane,this paper presents a method of local inspection path planning for tower crane.First,the improved YOLOV9 target detection algorithm was trained to realize the recognition of the tower crane by the UAV,and the identified tower crane was located by using the monocular distance de-tection algorithm.Then the skeleton of the tower crane is extracted,and the cluster fitting of the skele-ton after discretization is obtained,and the local inspection path of the UAV is finally obtained.The ex-perimental results show that this method can realize the path planning of tower crane and tower crane arm,and the average running time is 1.14s,which meets the real-time detection requirements.
维普
万方数据