| Robust transmission of electrical energy is a strong guarantee of the orderly development of modern society.However,factors which are the redundant sites,large distribution areas and permanent exposure to outdoor environment of the transmission lines and tower accessories cause the lines to vibrate,expand or contract.Those deformations lead to loosened bolts as the most common potential security problem in power transmission.Most UTV transmission lines are located far away any city in complex terrain and various natural conditions.It is difficult for manual inspection and operation.In order to repair the transmission lines effectively as well as reducing the labor intensity and risks of the maintenance staff,this paper introduced the identification method of the hot-line robots in fastening the bolts,designed and developed an experimental prototype of the bolt fastening robot.Besides,this paper also showed the tests of bolt identification algorithm and fastening analog line bolt based on this prototype.The hot-line bolt fastening robot includes the walking body,wheels,grip,equipotential wheel,mechanical arm and bolt fastening operation mechanism,etc.Those parts ensure the robot to move safely and stably on the transmission lines and to complete the hot-line bolt fastening tasks.The bolt fastening operation mechanism which is located at the end of the mechanical arm,mainly includes two wrench sockets,two cameras,micro motor,a number of supporting several transmission structure,etc.Both cameras are close-focusing cameras,installed in the center of the wrench sockets.They can obtain relatively clear images of the objects in close distance and blur the background at the meantime.As long as to adjust the mechanical arm to keep the bolt in the center of the images,the registration work is completed and then you can continue the following fastening operations.There are numerous bolts in overhead transmission lines with relative small volumes.Besides the complex background and linear structures of the transmission lines bring a big challenge to the bolt-hexagonal recognition method based on Hough Transform.It is hard to design the identification algorithm.Most drainage lines on the transmission lines are made of multiple twisted stranded wires with a unique texture feature.Those drainage lines become an obvious feature around the bolts.This paper simplified a bolt searching process by using drainage lines as reference.In other words,the method is to get the trend of the drainage lines by locating them first and find the bolts along the drainage lines.The majority of structures around the bolts are made from grayish metal.Both these structures and metal bolts are prone to mirror reflection.There will be a highlight interference area especially under the strong sunlight and rains.Thus,a simple threshold region extraction method is not applicable.To avoid this interference on the accuracy of the bolt recognition,this paper will improve the classical Hough straight line detection by using the image edge information and achieve the hexagonal bolt identification and location by implementing the hexagonal peak voting strategy.Meanwhile,this paper will classify objects in view preliminarily by combining the HOG and the SVM method and improve the accuracy of identification.Algorithm is proposed in this paper in the MATLAB platform,the design and development in the Visual Studio 2013 platform testing experiment.The experimental results show that this algorithm can efficiently complete the bolt on the transmission line,search,identification and localization work,improve the level of the bolt fastening of live working robot intelligence,improve the efficiency of the live working. |
PDF Full Text Request