Research And Development Of Cable Surface Damage Detection System Based On Multi-channel Video Acquisition And Image Analysis

With the large amount of domestic investment in infrastructure,there are more and more long-span bridges and super-span bridges.Among them,mostly cable-stayed bridges and suspension bridges play an important role in transportation.In cable-stayed bridges and suspension bridges,the cable is the major structural members.The service-life and safety of bridges mainly depend on the degree of cable damage.If the cable protection layer is damaged,the internal steel cable will be exposed to the severe environment.Corrosion of the internal steel cable can affect the safety of the whole cable.Therefore,regular inspection of the cable is necessary.At present in China,cable surfaces are detected mostly by hand.This detection method is inefficient,costly,and poorly safe.In order to solve this problem,a lightweight cable crawler robot equipped with multiple video system for surface detection was designed.The main works and results of this paper are as follows:1)A lightweight cable crawler robot was developed,and it is equipped with a multiple video system for surface detection.The cable crawler robot designed in this paper adopts a modular assembly structure and is similar to a screen when it is deployed.The robot enclose the cable when used.It has the characteristics of smart structure and small body.The screen type structure makes it easy to install and disassemble.2)A real-time acquisition and transmission system for multichannel video was realized.Cable surface detection system is equipped with four cameras on the crawler robot,and it transmits real-time video data to PC client by WIFI wireless transmission technology.Based on the UDP transmission protocol,this paper designed a high quality multi-channel video transmission method.It can ensure the real-time of video data while transmitting multiple video images.The application software design of video detection system for mobile platform and PC client were detailed described in this paper.3)An extraction method of defect target in cable images was designed.At first,the cable surface image is preprocessed.Then the image segmentation,the target detection,and target determination on the defect area of the cable surface are performed respectively.By this method,it is more accurately and quickly to identify the damaged image on the cable surface.