MACHINE VISION BASED LOCAL FISH RECOGNITION

Abstract

Every single country has its own tradition which help to identify its self differently in front of the whole universe. Like others our county has some abundance too. River dimensional, this country is the state of sweet water fishes. Because of natural climate this is the only birth place of those local fishes and available here. Which fish are local? Where they belong? When they can be found? So many questions are related with them as we hardly know our local fishes. Branch of them can identify easily, but most of are lost in the absence of saving. Without knowing, we cannot save those fishes from being extinct. A lot of work has been done in the field of research on recognition which has been came in front of us. There is minimum work on fish, and if we focus on the word local, still any research has not been done on Bangladeshi local fish. This is why it make us more interested to work on this section. With the help of this research some local fish can be recognize by capturing image. For research, we collect data from local fish market and then with the help of machine learning the rest of work has been done. We use some feature extraction like Color Analysis, Texture Analysis, Segmentation and three type of classifiers for training our data. After classification various experimental test have done which can determine whether it’s enough hard to recognize our local fishes. This work of research is related with Machine- vision, in that case image processing is used to recognize our local fishes. Machine Vision is the one kind of technology and methods which is used to provide imaging-based automatic inspection and also analysis for such applications like automatic inspection, process control, and robot guidance, usually in industry.