Water transparency analysis in fish farming environment through unmanned aerial vehicles

Abstract

Since water transparency in a fish farming environment is an important parameter in fish development, and it is important that the fish farmers be attentive to the control of the same. However, the process of measuring water transparency in fish ponds is usually done manually, which makes it expensive and inefficient. In order to solve this problem, a prototype was developed containing 2 optical sensors capable of capturing images in the NIR, R, G and B bands, connected to a microcomputer coupled to a Remotely Piloted Aircraft (RPAS), in order to capture aerial images of water and estimate water transparency through an image processing algorithm. The prototype was tested by an experiment conducted in a rural estate of the city of Toledo – PR, where 66 images were captured on a conical target, constructed to serve as a reflective reference for the images, comparing the results with the measurements collected by a Secchi disk, using Pearson's correlation analysis. The results showed a strong correlation (r = 0.98) in the estimates made by the NIR band processing, followed by the RGB, R, G and B bands. The developed system proved to be effective in data collection, processing and analysis, optimizing the fish farming management.