Study On Common Shellfish Feature Recognition Technology

Abstract: With the popularity of IT and the rapid development of informationtechnology, image recognition technology plays an increasingly important role in people’slives. For example, this technology has been widely used in character recognition (the postoffice letters of automatic sorting machine), medical image analysis and recognition, notes ofanti-counterfeiting, commodity bar code identification, missile and other military projects.Image recognition technology represents different meanings for different peoples, but it isgenerally accepted that this technology refers to the image processing by computer, i.e., thecomputer is enabled to read and analyze images, and eventually the images are identified bycomputer instead of human brain.It can be said that image recognition technology is an extension of the human visualperception. So far the image recognition technique has been applied to the identification ofsea foods such as fishes, sea cucumbers and so on, but few applications of this technology inrecognizing shellfish is seen. In this paper, the discussion and research were performed toinvestigate the capability of image recognition technology in identifying the morphology andcharacteristics of shellfish.In this paper, image recognition is carried out to identify those shellfish that are on theproduction line or those shellfish that have been simply treated after salvage. Under suchenvironment, the image recognition can clearly identify the shellfish without any attachments.Based on image acquisition, image preprocessing, feature extraction, classification andrecognition techniques, the shellfish image recognition method is proposed, which is used bycombining the Gabor transform,2-D image principal component analysis(2D PCD) withextreme leaning machine(ELM). Using this method, the shellfish classification method isachieved based on the complicated calculations such as image Gabor transform, imagefeatures extraction, and image characteristics dimension reduction for the shellfishes.Based on the comparison between the experiments and the calculating results ofBP(Black Propagation) neural network and support vector machine(SVM), it has beenfound that the extreme leaning machine in the shellfish classification recognition not only hasa satisfactory speed, but also has a wide range of applications. And a kind of new algorithm isproposed in this paper. Through the contrast experiments between BP neural network and support vectormachine, it has found that the extreme learning machine in the shellfish classificationrecognition has good generalization, and the experiments show that the algorithm has higherprecision in this paper.