Font Size: a A A

Research On Analysis And Extraction Of Inland Lake Stream Texture

Posted on:2022-10-28Degree:MasterType:Thesis
Country:ChinaCandidate:W J PengFull Text:PDF
GTID:2480306524996349Subject:Cartography and Geographic Information System
Abstract/Summary:PDF Full Text Request
Lake is an important part of nature,maintains the ecological balance of various regions and provides a wide range of resources and space for human survival and development.Lake current,as an important geographical phenomenon in the lake,directly affects the exchange of materials in the lake.,It always affects the ecological environment and social development of the lake and its surrounding areas.If you can grasp the flow direction information of the lake,it can provide greater help for environmental monitoring and waterway transportation.The traditional field measurement method is limited by the complex terrain and landforms in the field,poor traffic accessibility,and high cost of manpower and material resources;the model construction and running time of the hydrodynamic simulation method is long and the timeliness is insufficient;while the current remote sensing monitoring method is mostly The hydrodynamic simulation based on the tracer lacks in-depth analysis of the lake current texture in the image;for this,this paper proposes a method for extracting lake current information based on the image texture.This paper takes Poyang Lake,Dongting Lake,and Tai Lake as the research area,combines multi-temporal data and analyzes the lake current texture through Tamura texture eigenvalues and gray-level co-occurrence matrix(GLCM).The analysis results show that the lake currents of inland lakes reflect certain visual characteristics in the images.After calculating the Tamura characteristic values of the lake current in images of different resolutions,it is found that roughness,linearity and resolution have a certain correlation,while contrast and directionality are less affected by the image resolution.Landsat8 is more suitable for lakes.Stream texture extraction.For different lakes,the Tamura texture feature values of Dongting Lake and Poyang Lake are similar,while the roughness,direction and linearity of Lake Tai are lower than the former two,and the significance of the lake current texture is lower.Secondly,this paper analyzes the directionality of the texture through GLCM,and analyzes the variation of GLCM characteristic values in different regions with quarters as the time unit.The results show that the lake texture has certain seasonal characteristics,mainly manifested in water regimes,topography,etc.Changes in environmental factors lead to changes in the texture of the lake.Combining the results of texture analysis above,this article uses Laws algorithm,LBP algorithm and Gabor filter to enhance the image texture,and analyzes the characteristics of each method based on GLCM feature value.The results show that the three methods can enhance the image in different aspects.The Gabor filter can increase the energy and reduce the entropy value of the texture information;the Laws algorithm significantly improves the gray-scale correlation;the LBP operator can increase the entropy value of the image.After texture enhancement,this paper uses Hough transform to extract lake flow information,and compares the effects of the above three methods with multi-source data and measured flow direction.The results show that Gabor filtering has relatively stable characteristics for different data sources.Laws algorithm It is suitable for medium-resolution images and the LBP algorithm is suitable for high-resolution images.Based on the flow directions extracted from Landsat8 images,the average relative errors of the Gabor filter,Laws,and LBP extraction flow directions are 0.0279,0.0387,0.0592,and the root mean square errors are 17.58°,29.64°,and 38.89°,respectively.It can be concluded that the Gabor filter texture After the enhancement,the accuracy of the extracted flow direction is higher,followed by the Laws algorithm,and the LBP operator is the lowest.Finally,Gabor filtering is used to process the image data of Lake Taihu,Poyang Lake,and Dongting Lake and extract the flow direction of the lake,and analyze the flow direction distribution of each lake in the four seasons of the year.The results show that the flow direction of each lake has certain seasonal characteristics.Poyang Lake and Dongting Lake are mainly reflected in the change of water level due to seasonal changes,which affects the flow direction.The flow direction of Taihu Lake is greatly affected by the change of seasonal wind direction.In addition,this article also analyzes the changes in the flow direction of the lake under the action of different factors.The results show that natural effects will change the flow direction of the lake,which is mainly reflected in the changes in topography and water level and aquatic plants.Greatly disrupt the hydrodynamic state of the original lake.In summary,it is shown that for inland lakes,the method in this paper can better extract the flow direction of the water body,and has certain theoretical and practical value.
Keywords/Search Tags:Remote sensing, inland lakes, Texture analysis, Lake currents, Flow direction extraction
PDF Full Text Request
Related items