This work aims at grading the oil palm crop bunch in to three categories unripe, ripe and overripe. Different color feature models like color histogram, color moments, color correlogram and color coherence vector are used to extract the color features of the crop bunch. Oil palm crop bunches are classified into above mentioned grades using Probabilistic Neural Network. Experimentation is carried out using image dataset of 300 RGB images across three categories. An accuracy of 98.33 is achieved with 70 training, 10 validation and 20 testing for Color Coherence Vector features.
展开▼
