Study On Construction Defects Detection Technology Of 10kV Power Cable Joint Based On Image Recognition

Power cables play an important role in the construction of smart grid.With the acceleration of urbanization,10 k V power cables are increasingly applied to transmission lines.The cable joint is an essential part of the power cable line,and it is also the most vulnerable part.According to statistics,about 30% of the faults in the cable fault occur in the cable joint,and the construction defect of the cable joint is the most important cause of the failure in the cable joint.At present,the research on the typical construction defects of power cable joints is only in terms of hazards.There is no effective method to detect the typical construction defects and it is impossible to control the construction defects from the source.In this paper,the cable joint of 10 kV power cable is made,and the typical types of defects which are easily produced in the construction process are studied,and the key nodes of the inspection of the construction defects in the cable joint are obtained.According to the shape and size characteristics of 10 k V power cable,designing a device for identifying typical defects in the construction of cable joints and taking pictures of the construction process.The image is processed to determine whether there are typical defects and types of defects.It provides a new idea for detecting typical construction defects of intermediate joints.Through the analysis of the construction process,it is concluded that the typical construction defects of the cable joints are the main insulation surface stains,the main insulation scratches,and the outer semiconducting layer peeling irregularity.According to the color difference between the stain and the main insulation,it is proposed that the gray color difference is used as an effective feature to determine the defect.In view of the problem that the area of scratch defect and the color difference of the main insulation are not obvious,the purpose of effectively distinguishing the main insulation stains and scratch defects is achieved by extracting the texture features.For the main insulation defects,a defect recognition method based on neural network is proposed,which effectively identifies two kinds of defects: stain and scratch on main insulation.For the outer semiconducting layer defects,the characteristics of the outer semiconducting layer were analyzed.The position of the outer semiconducting layer was adjacent to the main insulation.The color difference with the main insulation was large,and there was a prominent boundary at the occurrence of semiconducting layer peeling irregularity.It is concluded that detecting the edge of outer semiconducting layer is the key to identify such defects.In this paper,the advantages and disadvantages of the traditional edge detection algorithm are compared,and the method of Kirsch operator edge detection is obtained to extract the edge of the outer semiconducting layer,which can effectively reduce the detection of the pseudo edge.The minimum circumscribed moment is calculated through the extracted edges,and the rectangular degree can be used to effectively judge the irregularity of the semiconducting layer.The actual cable joints images taken by the identification device do not only contain the main insulation or the outer semiconducting layer,so the two parts need to be separated from the entire image.Through image preprocessing,the color space of the image is transformed,the region of interest is highlighted,the image is de-noised,the main insulation and the outer semiconducting layer are segmented,and the detection of the cable joint defect of the complex image is solved.For the divided main insulation and outer semiconducting layers,defect detection was performed using the determination method of each component,and the defect of the entire cable joint was judged.The overall detection efficiency reached 75%,and the undetected defect image was analysed.The analysis pointed out the direction for later research.