Fish Detection by Machine learning Approach

Abstract

In Bangladesh fish farming is already a major source of employment and many working facilities can be created through high-tech commercial fish farming systems. Even the unemployed educated people can also contribute to this business and create a lucrative business and earning opportunity for them. This “Fish Detection by Machine learning Approach” makes our life easy and increases our knowledge about fish. This project represents a model to detect and recognize local fishes of Bangladesh implementing image processing and neural networking approaches. The project work aims to apply computer vision and AI techniques so that people of the next generation can recognize Bangladeshi fishes as most of the young people in the city, have less idea to classify traditional and desi fishes. We implemented our custom Dataset consisting of 1250 sample images for the experiment method to measure out its credibility. In the proposed, model a sequential grassfire algorithm is used along with pre-processing techniques like noise cancelation, gray scaling, flood-fill method, binarization to detect and analyze the shape of fish. Then Further, to do classification and recognition of the detected fishes, convolutional neural network (CNN) and method of Visual Geometry Group (VGG-16) had been applied.