The Classificate And Retrieval For Red Tide Algae And Image Segmentation For No Chaetoceros Algae And Applied Research Based On SIFT

In recent years, due to seawater contamination, excess nutrients get plenty oforganic matter, plus sunlight, temperature, and other appropriate conditions, some ofthe algae suddenly blooms, make sea water discoloration, the red tides occurrence.more and more frequent HABs (harmful algae blooms) posed serious threats oncoastal environment, marine resources and public health. It is a serious global marinedisaster causing billions of economic loss every year in China, which caused moreattention by the governments and public. Artificial observation and analysis cannotsatisfy the prediction of red tide due to the limitation of their efforts and biologicalknowledge. It is thus very urgent to study the methods of warning, forecasting red tideand establish operational monitoring system. It should be noted that identifying thedominant species of red tides plays an important role in automatic monitoring of redtides.Based on the traditional biological morphological theory, this paper divides41kinds of red tide algae in China’s coastal waters, and designs a software for red tidealgae classification retrieval. And according to the no chaetoceros algae, proposed agoal segmentation method based on SIFT. The experimental results has demonstratedthis method can better segment the target image and is applicable to all the nochaetoceros.The main work of this paper include:1、 Based on the traditional biological morphological theory, this paper designs asoftware for red tide algae classification retrieval. The design purpose of theclassification retrieval software is to complete man-machine interaction. Tothe common40algal red tides, this software on the basis of lifestyle, cellsize, cell shape, the cytochrome body color, in citro characteristics, groupcharacteristics, form the outside to inside, from the simple to the complex way to indentify. Through the way of man-machine dialogue, use thepictures and texts assisted monitoring staff to determine the detection algaespecies. And displays detailed information of the target species, as well as inthe microscopic image of the light microscope and electron microscope.2、 According to the no chaetoceros algae, this paper proposed a goalsegmentation method based on SIFT. First, according to the local invariantfeature extraction algorithm to extract key points, the feature information ofthe key points on behalf of the characteristics of the original image. The keypoints and then closing operation, corrosion, expansion, and othermorphological processing to remove noise, and the after the step of contourextraction image fill, we can get a good target image cells segmentation.Finally, a large number of experimental data have proved the segmentationresults are applicable to the target image feature extraction and imagerecognition mudule, in turn, can get a better image recognition results.