A Fish Detection Approach Based on BAT Algorithm

摘要: Fish detection and identification are important steps towards monitoring fish behavior. The importance of such step comes from the need for better understanding ecology issuing conservative actions keeping safety this vital food resource. recent advances in machine learning approaches allow many applications to easily analyze detect a number species. main competence between these is based on two parameters: time accuracy measurements. Therefore, paper proposes approach BAT optimization algorithm (BA). This aims reduce classification within process. performance system was evaluated by well-known classifiers, KNN, ANN, SVM. tested with 151 images Nile Tilapia species results showed that k-NN can achieve high 90 %, feature reduction ratio close 61 % along noticeable decrease time.