Algae Image Segmentation Based On The Level Set Of Regional Information

At present, the monitoring of red tides on ocean and lake blue greenalgae in our country is mainly rely on satellite remote sensing technology,which can not be forecasting and early warning. In view of the above question,the forecasting, early warning, monitoring and control so as to reduce theinfluence of the environmental disasters is important. And the imagerecognition of algae species and form is the important link of the monitoring.According to the micrograph, identify the red tide or water bloom types,advantages algae, population density and growth trend with the imagerecognition software. However, due to the algae cell is complicated andmicroscopic image is susceptible to the light and the influence of the algalcells color, the traditional image segmentation algorithm on algae contourextraction is difficult to obtain satisfactory results. In this paper a methodbased on a local Chan—Vese(LCV)model, used in algae cell microscopicimage segmentation.First of all, this paper introduces the significance of algae research andthe current situation of the development of algae analysis technology both athome and abroad, mainly to the level set method and its application in imagesegmentation and further expand in-depth study. Various methods areanalyzed and compared, summarizes the advantages and disadvantages ofvarious methods, performance and problems. According to several commonMarine red tide organisms and lakes, reservoirs blue green algae biologicalcell structure characteristics, analyzed the traditional C-V model the existingproblems. Using a new based on Local information of the Local Chan-Vese(LCV) model, can be in less iterations internal division gray uneven image, applied to algae cell microscopic image segmentation, and good results havebeen achieved.In the MATLAB environment, through a large number of imageprocessing and analysis, retain the original information, algae eventuallyreturn to identify only contain algae bitmap. Choose the best distinguish algaecell characteristics combination, for later analysis using algae characteristic.Through the contrast, show LCV model compared with the traditionalsegmentation method can be partitioned gray uniform or non-uniform algaeimage.