Research On Pipeline Auxiliary Detection Based On Swarm Intelligent Optimization Algorithm

The fast development of pipeline industry in our country has become an important part of national infrastructure construction,and the safety and reliability of pipeline property safety and stability.However,pipeline accidents happen from time to time.On the one hand,it is because of the process quality problems during production,on the other hand,it is caused by internal and external corrosion in daily use.Therefore,the daily maintenance and detection of pipelines has become an important task.Generally,there are external and internal detection methods for pipeline detection.The external data collection usually only contains digital data but does not include image acquisition,or collection by complex equipment costs a lot of manpower and material resources.Internal detection usually collects image data,but the quality of pipeline image is affected by various objective factors,it can not reach the ideal situation,machine recognition and manual recognition time consumingIn view of the above detection status,this paper takes a natural gas pipeline in a city of Anhui province as the research object.Firstly,small sample data of the inner diameter of the pipeline are collected outside the pipeline as the data set to predict the pipeline corrosion trend,set key measurement points,predict the pipeline corrosion situation according to the monthly data changes,and implement image collection of the inner side of the fixed pipeline according to the situation.Secondly,the pipeline interior image acquisition robot is built,and the appropriate acquisition equipment is selected for image acquisition.The collected images are uploaded to the supreme machine,and the image is nonlinear enhanced by the improved swarm intelligence algorithm.After enhancing the readability of the enhanced images,Otsu multi-threshold segmentation technology is used to segment the defects of the enhanced images,so as to facilitate the artificial judgment of defects in the later stage and provide feature materials for machine learning.The main research is as follows:First,pipeline loss prediction based on improved bee colony algorithm to optimize grey model parameters is studied.Firstly,the pipeline inner diameter data collected by ultrasonic were summarized,and the gray model was improved from two parts: 1.The summarized data was preprocessed to eliminate error parameters;2.Improve the construction of grey model background value;And the swarm algorithm is improved.Handle data using the grey model,and the improved colony algorithm optimization ?value as the important parameters of the changes in the grey model.Second,build a pipeline image collection system.Look up the information from the Internet,draw the basic circuit diagram,buy accessories to build the robot platform,write the robot movement system,can achieve the effect of smooth and fast walking in the pipeline,the collected image data is uploaded to the supreme computer.Third,the nonlinear enhancement of pipeline defect image based on swarm intelligence algorithm is studied.There is a lack of light inside the pipeline,and LED bulbs are used as point light source for brightness support,but the image acquisition is still poor readability.The improved swarm intelligence algorithm is used to find the parameter values of gray image curve adaptively.The contents of this chapter include the improvement of firefly algorithm and classical particle swarm optimization algorithm,as well as the comparison of the original image,the image enhanced by two nonlinear image enhancement methods and the image enhanced by histogram linear enhancement algorithm.It is found that the nonlinear enhancement algorithm based on swarm intelligence algorithm has certain feasibility and application value in industrial production detection.Finally,Aiming at the problem of poor readability of the original image with pipeline defects,an image segmentation technique based on improved particle swarm optimization algorithm was proposed to optimize the threshold parameters of Otsu multi-threshold segmentation.The multi-threshold segmentation of the enhanced image will better display the defect edge,which is convenient for the artificial later study and judgment of the defects in the image,and take the next maintenance measures for the pipeline.The experimental results show that the mean variance of the prediction results of pipeline internal wear optimized by the modified ABC algorithm is less than the prediction results of the original GM(1,1)model,and the prediction results are more accurate.The nonlinear enhancement techniques of improved FA algorithm and improved PSO algorithm in optimizing gray curve parameters of gray image are better than those of other control groups in terms of information entropy.In the image segmentation strategy of improving the PSO algorithm to optimize the parameters of Otsu segmentation algorithm,it can be seen intuitively that the boundary segmentation between the image defect and the background is obvious by comparing the original image,which has a good segmentation effect.In conclusion,the prediction of pipeline wear in the early stage,image enhancement in the late stage and segmentation strategy based on swarm intelligence algorithm can well assist the manual judgment of pipeline defects and play a role in auxiliary detection.Figure28 table7 parameter119...