Font Size: a A A

Study On Northeast Tiger Skin Texture Extraction And Recognition Based On BP Network

Posted on:2009-01-07Degree:MasterType:Thesis
Country:ChinaCandidate:P ZhangFull Text:PDF
GTID:2178360275466941Subject:Biophysics
Abstract/Summary:PDF Full Text Request
Northeast tiger skin texture as a biological characteristics shown genetic characteristics appear to significant different in tiger populations. It is inextricably linked to genes. Effectively extracting and identifying tiger skin texture feature make sense of northeast tiger individual recognition and researching on tiger texture inheritance. Color images of one hundred and three northeast tigers in Heilongjiang northeast tiger park were collected by digital photographic equipment. One hundred tiger skin texture color images of both sides of each northeast tiger trunk were collected.This paper presents three-layer BP neural network can be used in northeast tiger skin texture feature recognition, because three-layer BP neural network has advantages in pattern recognition. The images of northeast tiger skin texture were processed by computer image processing technology, according to the color and the shape characteristics of northeast tiger skin texture. This paper presents several common image processing methods that can highlight the tiger skin texture feature. Some methods like equalization and filtering enhance texture image contrast, remove the noise of image, highlight tiger skin texture feature. These methods make the tiger texture feature clearer, easily recognized and observed. Tiger skin texture feature was separated from the texture image by binary segmentation. The tiger skin texture ending positions can be easily observed and measured after the image was processed by mathematical morphology method. Ten distances of eleven adjacent texture endings which started from the spine of tiger were in turn selected from the hip to the head of each tiger can be used as feature information. These feature information can be used in northeast tiger individual recognition by BP neural network. This paper presents the best number of neurons in the hidden layer of tiger recognition BP neural network by comparing network performance. The experiment show the northeast tiger individual recognition results tally with the actual circumstances. This paper presents a new method in recognizing northeast tiger individuals according to tiger skin texture image.
Keywords/Search Tags:Tiger skin texture, Image processing, Artificial neural network, Pattern recognition
PDF Full Text Request
Related items