Research On Intelligent Detection Technology Of Deep Pipe Inner Morphology

In modern industrial production,pipelines are widely used in all walks of life.Due to longterm use,scratches and corrosion on the inner wall of the pipeline may cause irreversible damage,for example,the gun bore bursts or the transmission pipeline leaks,so it is necessary to regularly detect the inner morphology of the deep pipe.However,the complex internal conditions of pipelines,traditional detection methods require high operators and are inefficient.Therefore,it is particularly important to study the intelligent detection technology of deep pipe internal morphology.tubeIn this paper,based on the working principle of the inner morphology detection and control system and the collection system of deep pipe,an intelligent inner morphology detection system is designed by using STM32 and FPGA,which to achieve intelligent and real-time collection of the inner morphology of the pipeline.At the same time,the image processing technology which combines the FPGA and the host computer is used to filter,correct the distortion and stitch the image of the inner wall of the deep pipe,which improves the image quality and detection efficiency.The main contents are as follows:(1)Designing of the inner morphology detection and control system for deep pipe.System hardware designed includes STM32 main control unit,power circuit,serial port circuit,wireless communication module circuit.The software design of the control system is carried out around the hardware framework,Serial communication is used between the host computer and the underlying hardware structure to achieve intelligent control of the deep pipe morphology detection device according to instructions.(2)Designing deep pipe internal morphology acquisition system.By using FPGA,IIC interface module is designed to configurate CMOS camera,DDR3＿control module is designed to control the AXI bus to cache the acquired images into DDR3 SDRAM memory and image transmission module which based on Ethernet protocol is designed to package and transmit image data to the host computer for real-time display.The logic design of each module for deep pipe image acquisition and transmission is implemented using hardware description language Verilog,Used the simulation software Questa Sim to simulate each module,establish an experimental platform under the condition that the simulation waveform is correct,and verify the deep pipe image acquisition and transmission module.The results show that under the working clock of the system,image data transmission has good real-time performance,no packet loss,and meets the requirements for resolution and frame rate.(3)Research on image processing technology for collected deep pipe images.Firstly,the mean filtering algorithm is implemented using FPGA to denoise deep pipe images.Then,image distortion correction algorithm is designed using matlab,and interpolation is used to correct the distortion of deep pipe images,effectively eliminating the distortion phenomenon that exists in the process of deep pipe image acquisition;Based on the advantages of high registration accuracy and good stitching effect of SIFT algorithm in deep pipe image stitching,this algorithm is selected to provide specific feature point extraction,key point location,and matching results,and feature point fitting is performed using RANSAC algorithm,then image stitching module is designed by matlab to complete the design and verification,the field of view after the image stitching is expanded to 2.7 times the field of view of the deep pipe image to be spliced,and the image stitching quality is high.