Translate 
Abstract
Estimating the biomass volume of fish in a pond is one of the most common and relevant practices in the cultivation of biological organisms. Regular real-time data collection on fish biomass is required by researchers to optimize daily feeding, control stocking density, and determine the optimal time for collecting biological samples. Unfortunately, assessing fish biomass volume without human intervention is quite challenging, as fish move freely in an aquatic environment where visibility, lighting, and environmental stability factors are beyond control. To date, the most common method for assessing fish biomass has primarily been manual sampling, which is typically invasive, labor-intensive, and time-consuming. In this context, it is necessary and desirable to develop non-invasive, fast, and economical methods. Machine vision and video stream analysis from cameras help develop non-invasive, faster, and cheaper methods for assessing fish biomass in ponds. This article summarizes the experience of developing such methods for assessing fish biomass and presents their main concepts and principles. The strengths and weaknesses of each method are analyzed, and future research directions are presented. Research shows that the application of information technologies, such as advanced sensors and communication technologies, plays a significant role in accelerating the development of new tools and methods for more effective biomass assessment. The main goal of this study was to develop an automatic system for assessing fish volume using machine vision, collecting visual data on the geometric characteristics of fish extracted from the video stream, and machine learning algorithms exemplified by the Haar cascade. However, the accuracy of these methods still needs improvement to meet the requirements of intensive aquaculture.
Published Version
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
