Fish Freshness Detection Using UV Light Based On Deep Neural Network

Fish contains a lot of nutrients that are very beneficial for the body, but fish are often traded both live and dead. To observe the freshness of the fish, it is done by looking at the changes that occur in the fish's eyes and gills using UV imaging and image processing. Indicators of fish freshness quality can be seen through the color and smell. By using UV imaging can be seen the reaction that occurs in fish. Using UV light means freeing electrons from atoms or molecules, which can cause chemical reactions. After collecting data using UV imaging, the freshness level of the fish will be determined using deep learning. Deep learning is used to determine the level of segmentation of fish freshness. These results indicate that using DNN successfully assesses fish freshness non-destructively only by capturing fish images with UV light with a wavelength of 365-400 nm. By using segmentation based on RGB, CIE L∗a∗b, and HSV it is proven that this technique can accurately reveal changes in fish freshness with an R of 0.993. So that with this technique it is possible to evaluate the freshness level of fish non-destructively and classification of fish can be done online.