Building a fish disease detection application based on Smart Fishery with Image Processing and Deep Learning

Indonesia has great potential in the fisheries sector, especially in catfish cultivation which is one of the main commodities. Catfish has advantages in fast growth, high durability to the environment, and high market demand. However, the main challenge in catfish farming is disease attacks that can lead to mass deaths and economic losses for farmers. This disease often arises due to poor water quality and lack of knowledge of farmers regarding its identification and handling. Deep Learning-based artificial intelligence technology with the Convolutional Neural Network (CNN) method offers automatic fish disease detection solutions through digital image analysis. The CNN model is able to recognize patterns and classify images with high accuracy, so it can be a quick and efficient diagnostic tool. This study developed a Smart Fishery-based fish disease identification system with Image Processing techniques to improve image quality before being analyzed using CNN. The implementation of this method is expected to increase the efficiency of detecting catfish diseases and provide appropriate treatment recommendations. The study results indicate that the developed system effectively achieves high accuracy in classifying fish diseases. Moreover, the system includes additional features, such as virtual fish growth measurement and texture analysis utilizing the Gray Level Co-occurrence Matrix (GLCM). This technology is anticipated to assist fish farmers in enhancing productivity and minimizing the risk of fish mortality due to diseases.