Application of backpropagation method for quality sorting classification system on white dragon fruit (Hylocereus undatus)

Abstract Several problems related to determining the quality of dragon fruit are: disease, harvest time selection, sorting process and post-harvest grading. Determination by observing appearance fruit, smoothness, presence or absence defects size. However, this determination has disadvantages such as longer different perceptions farmers about fruit. To solve problem, we need a system that is able determine effectively efficiently without damaging In study, white using digital image processing techniques intelligent systems. The output technique five morphological features area, perimeter, length, diameter metric. This feature input backpropagation method so divided into 3 classes class A, B C. results showed best network architecture model was 5,8,5,3 with testing accuracy rate 86.67%.