Fuzzy Logic Approximation and Deep Learning Neural Network for Fish Concentration Maps

Abstract

This paper proposes an algorithm to obtain topographic maps of lakes, maps of fish concentration and a map of predator location based on the results of an intelligent sonar data processing. The algorithm is based on the following steps: input frame separation into overlapping blocks, blocks-processing using convolutional neural networks (CNN) YOLO v2, and merging extracted bounding boxes around one object. To construct maps of the distribution of features along the lake, we propose a novel method for constructing the approximation of GPSreferenced CNN results based on the original implementation of fuzzy logic. Keywords— sonar data; fish concentration; maps of lakes; fuzzy logic; convolutional neural networks