Font Size: a A A

Research On Mattress Laying Process And Afterwards Overlapping Detection Based On Sonar Image Analysis

Posted on:2018-11-10Degree:MasterType:Thesis
Country:ChinaCandidate:F Y HuFull Text:PDF
GTID:2382330596954789Subject:Software engineering
Abstract/Summary:PDF Full Text Request
Mattress is the basis of waterway regulation engineering,which can effectively prevent river bed from soil erosion and the damage due to the water seepage.Because of the difficulties to obtain the information from the underwater,there is no effective method for mattress layout and the after detection means.The mattress overlapping detection in the process of mattress layout is mainly to detect the overlap width.The afterward detection is mainly to check whether the overlapping area is unqualified.So the two stages of the detection method require different implementations.Two stages of the overlap detection means missing,resulting in high cost and unstable quality of the layout process,and the mattress can not be timely cleaning and maintenance after layout.The main purpose of the thesis is to realize the precise guidance of the layout and the quality detection of mattress overlap.Through the analysis of the two stage's mattress sonar image.The main research work and achievements are as follows:(1)Designed the mattress layout process and afterward overlapping detection.In the thesis,the detection scheme of overlap width at mattress laying is realized by three steps: firstly,the edge detection of mattress laying sonar image;and then the four main edges are fitted in a straight line,finally the lap width is calculated by the intersection of the straight line edges.The method of quality inspection scheme of the sinking row by image segmentation after arranging is used to detect whether there is unqualified overlapping area in the mattress sonar image.(2)Research on edge detection algorithm and optimization algorithm of layout image.The edge detection method of mattress layout sonar image needs to meet the requirements of CFAR and single pixel edge.The thesis mainly studies ROEWA and Canny edge detection operators.In the thesis,an improved ROEWA edge detection operator is designed based on the Canny criterion,and the improved method can obtain the single pixel edge and inherit the ROEWA operator.Aiming at the shortcomings of the Canny operator,the edge strength calculation method of the Canny operator is modified.The method obtains the precise edge detection of sinking sonar and can overcome most of the multiplicative noise.Experiments show that the improved ROEWA operator is more suitable for edge detection of the layout in low frequency and high opening angle.The improved Canny operator is more suitable for edge detection of the layout high frequency and low opening angle.(3)Research on segmentation method and optimization algorithm of mattress sonar Image.Analyzing the limitations of the traditional image segmentation method in sonar image segmentation and the superiority of the method based on MRF model segmentation.In the thesis,we choose the belief propagation(BP)algorithm which is better for segmentation and anti-noise performance in MRF model.Due to the large amount of BP algorithm,we designed a region BP algorithm.Through the experimental analysis,the improved BP algorithm remains the high precision of the original algorithm,and improves the running speed of the algorithm by reducing the number of iterations and each iteration time.The experimental results show that the method of mattress overlapping detection can be more accurately judge whether the overlapping area is unqualified.
Keywords/Search Tags:Mattress Sonar Image, Edge Detection, ROEWA Operator, Canny Operator, Image Segmentation, Region BP Algorithm
PDF Full Text Request
Related items