Research On Defect Detection Technology Of Pipeline Robot Based On Adaptive Image Enhancement

As a commonly used medium conveying device,pipelines play an important role in the life of residents and industrial development.However,due to the influence of natural conditions,transportation media,and human damage,the pipes are prone to damage such as corrosion,cracks,holes,etc.Therefore,it is necessary to carry out regular inspection,maintenance and timely repair of pipelines.Pipeline visual inspection technology has its unique advantages in the field of nondestructive inspection of pipelines due to its advantages such as fast detection speed,high efficiency,and visualization.Under the influence of natural weathering corrosion and artificial construction,crackings and holes occur inevitably,causing environmental pollution and resource waste which seriously affects the normal operation of pipeline,so it is necessary to carry out pipeline periodically maintenance repair in time.Compared with other detection methods,pipeline visual inspection speed is quick and efficient,and defects visualization can be realized.This paper studies the visual defect detection method of pipeline detection robots,and a visual inspection robot for pipeline defects is designed.According to the image features collected in the closed pipeline,an adaptive image enhancement method for detecting defects in pipeline robots was proposed.Experiments show that this method is highly adaptable and accurate,and can realize the effective detection of typical pipeline defects.The main work of this article is as follows:(1)The actual working conditions of pipeline detection was analyzed,then the overall scheme of the pipeline detection robot was designed,and the module design of the robot's key mechanical structure,image acquisition and processing was completed.(2)Aiming at the problem of uneven illumination of image acquisition during the pipeline inspection process by a pipeline robot,a single-scale Retinex adaptive image enhancement algorithm based on guided filtering is proposed.After analyzing the shortcomings of traditional single-scale Retinex image enhancement,the improvement of image enhancement effect is illustrated,and adaptive enhancement is implemented for different illumination images.In order to accurately separate the image brightness information,the collected uneven illuminance RGB images are converted into HSV images;then guided filtering is used to accurately estimate the illuminance components of the image brightness according to Retinex theory;finally,according to the brightness distribution of the image pixels,the illumination is balanced by adaptive image enhancement algorithm.(3)After performing the image enhancement pre-processing operation on the pipeline image,the defect area of the pipeline image is extracted.The traditional Canny edge detection operator is improved,and a defect detection algorithm based on bilateral filtering and adaptive threshold is proposed.The bilateral filtering method is used to replace the traditional Gaussian filtering,and the image gradient template is enlarged to fully represent the edge characteristics of the image.After iterative threshold segmentation and edge pixel connection,the effective extraction of defective edges is realized.(4)A pipeline robot defect detection experiment system based on image enhancement was established,and the development of interactive software for visual inspection of the pipeline was completed,thus the comprehensive detection,transmission,and display of the image of the inner wall of the pipeline was realized.The experimental results show that the detection method in this paper can adaptively correct the brightness of the pipeline image,the brightness unevenness is significantly improved,and the holes and crack defects of the pipe can be effectively detected,with the recognition accuracy close to 97%.