The Application Of3D Reconstruction Technique In Power Equipment Image Monitoring

China is one of the countries suffered serious icing problems on transmission lines.Icing may cause several serious problems, such as tower collapse and insulator stringflashover, which are danger to the power system and cause huge economic losses aswell as certain social influence. Therefore, it is significant to monitor the icingcondition of the power transmission as well as transformation equipment and protectthe power system from icing failure for the safe operation of the power system.According to the shortcomings of the existing methods for detecting the claddingice thickness, an online monitoring method for insulators and transmission lines isproposed in this paper based on3D reconstruction of images. The images of theinsulators and transmission lines before and after the icing process are got byembedded image acquisition system, transported to the computer control center byZigBee wireless network and then processed by the computer. The icing models can bereconstructed and the cladding ice thickness as well as volume can all be obtained bycalculating the reconstructed icing models of the insulators and transmission lines.The computer binocular vision technology is adopted in this paper. The two icingimages of a single insulator or a single conductor can be got by two same type camerasin similar position with different point of view. The mapping relationship of the plainpoints and space points can be obtained by the parallax of these two images and thecorrelation coefficients can be gained by camera calibration calculation. Cladding iceoutline can be obtained by feature point extraction and the depth of the points on thecladding ice outline in3dimensions space can be calculated through matching points,by which their3D models can be reconstructed. At last, the3D models are presented inpoints cloud and gridding image. The icing thickness and weight can be calculated bytheir gridding models.Several groups of images are reconstructed and analysed comparatively using themethods proposed in this paper. The3D model can highly restore the3D icing contourand reflect the cladding ice forms intuitively, by which the thickness and the volume ofthe ice wrapped in the surface of the insulator and conductor can be calculated withhigh precision. The relative error is below10%compared with artificial measurementresults. The position of the largest ice thickness can be given for insulator monitoringand the ice cover distribution along the lines can also be given for transmission lines monitoring. Compared with the existing methods, more icing information of icing statecan be obtained using the methods proposed in this paper, which can provide basis forice cover degree obtaining and icing locating. The results also show that the accuracymeets the requirement of engineering application, so it has huge developmentprospects in the cladding ice online monitoring on power transmission andtransformation equipment.