<![CDATA[The factor of fish quality can be affected by storage procedure and processing treatment is still done manually. Of course, it can make the fish quality will decrease and the sorting process will be wrong. From this problem, it is needed some research and system that can reduce errors to classify fish quality. On this research, we are using image processing and Bayesian method to classify fish quality. Fish will be placed on a styrofoam box that has been equipped with a webcam camera and lamp as lighting. Image processing is used to convert an image from RGB space to HSV space, and we crop the image to get the head section. And after that, we use the hue histogram colour information for the parameter to classify. so the value of bin1, bin 2, and bin 3 and also the standard deviation from histogram value are using as input for classification using Naive Bayes and will process in Raspberry Pi 3 and finally, we can get the fish quality. We are doing some testing. From testing how to implement image processing for this system we get some conclusion that the image which uses the lighting from 5 Watt lamp with white fabric clothes has a good image result, and the result for hue value information from images has to be added. And from testing Naive Bayes methods accuracy was 72.727% and the computation time was 468.864 ms.]]>
Copyrights © 2019
