IDENTIFIKASI JENIS KAYU BERBASIS CITRA MENGGUNAKAN PROBABILISTIC NEURAL NETWORK (PNN)

Abstract

The purpose of this research is to identify the types of wood based on imagery by using PNN method. In this paper, gray level co-occurence matrix (GLCM) is used as texture classification techniques. The GLCMs are generated to obtain three features: autocorrelation, cluster shade and sum variance. The classification technique used to classify the wood species is a probabilistic neural network (PNN). This research was carried out using 12 different types of wood. For each type of wood, 6 images were collected. The images of wood were divided in two sets: training set and test set. The leave-one-out cross-validation technique was applied for model validation. Our experimental results showed that the proposed method can increase the recognition rate up to 80.55%. The result of this research indicated that three features of GLCM are accurate to distinguish types of wood. This research used only a small-size dataset, so for further research is needed to use more feature extract methods and types of wood