Research On Water Blooms Detecting Algorithm Based On Image Processing

Water blooms is a very serious environmental disaster, it can make water hypoxia and produce water algal toxins, result in massive death of aquatic animals, cause serious ecological damage and incalculable economic losses,threat human survival and health, Therefore, real-time monitoring of the occurrence of algal blooms'changing and trending in the disaster, is a very interesting topic.After elaborating on the traditional water blooms monitoring methods and analyzing the current monitoring methods, this paper points out the problems existing in these methods, such as over-reliance on manual monitoring instruments, real poor, high cost and popularity of the poor, the subjectivity of large, labor-intensive and other problem, then propose an intelligent, low-cost, easy to promote Image-based monitoring method of water bloom.According to the problems above, After discussing target detecting and tracking algorithms based on image processing, this paper presents a ideal method of water blooms detecting and tracking which is based on color feature matching algorithm, can adapt to all kinds of weather, environment, lighting changes, can inhibit the background noise of the water complex interference with non-rigid target detection scenarios, this ideal algorithm of water blooms detecting and tracking is Camshift algorithm.After analyzing the mechanism of Camshift algorithm, summarizing the defect of Camshift algorithm in detect and the application context of water blooms detecting and tracking, the paper propose a method based on color prior models and multi-model Camshift algorithm to detect water blooms, and propose that filtering interference color by Morphological image processing with the area threshold, the approach of calculating the actual area of the water blooms region, establishing water bloom early warning model similar to meteorological disasters model by the link between the area of the concentration of algae, and then it can early warn water blooms in order to notify the relevant departments to take appropriate control measures.Using improved Camshift algorithm and water blooms early warning model above, the paper developed and implemented based on image processing of algal bloom monitoring system. Experimental results show that the system can accurately detect the water blooms region, and real-time computing the actual area of bloom and then early warn water blooms. The algorithm has less calculation and quick speed. Its adaptability to the water surface complex background and the changes in lighting conditions is very good. Especially, multi-model Camshift algorithm has a good robustness in detecting water blooms of the color changing of target area.