Investigation on Correlation of Water Quality Data with Aerial Images

Maintaining good water quality is fundamental for successful and sustainable pond fish farming. It directly affects the health, growth, and overall well-being of the fish population and contributes to the overall success of the aquaculture operation. Therefore, periodic control of water quality is highly recommended. The forthcoming methodology aims to assess water quality using aerial imagery. Nevertheless, to implement this approach, an initial inquiry is required to examine the correlation between aerial imagery and sensor data. In this scenario, data comprising TDS, pH, and temperature is acquired using manual sensors. Simultaneously, aerial images are captured using drones at different altitudes. Two approaches are employed to extract features from these aerial images: one involves determining the dominant eigenvalues of the image pixel matrix, and the other employs matrix convolution techniques, incorporating kernel matrices and max pooling matrix techniques. Subsequent to extracting the image feature data, correlation analysis is performed using the Spearman method. The result shows that there is a weak correlation between grayscale image and TDS level at 0.1 from eigenvalues method, and −0.30 from convolution method.