Application of artificial neural network for detecting Phalaenopsis seedling diseases using color and texture features

In this study, we present an application of neural network and image processing techniques for detecting and classifying Phalaenopsis seedling diseases, including bacterial soft rot (BSR), bacterial brown spot (BBS), and Phytophthora black rot (PBR). The lesion areas with BSR, PBR, and BBS of Phalaenopsis seedlings were segmented by an exponential transform with an adjustable parameter and image processing techniques. The gray level co-occurrence matrix (GLCM) was further used to evaluate the texture features of the lesion area. These texture features and three color features (the mean gray level of lesion area on the R, G, and B bands) were used in the classification procedure. A back-propagation neural network classifier was employed to classify BSR, BBS, PBR, and OK (uninfected area of leaf) of Phalaenopsis seedlings. The methodology presented herein effectively detected and classified these Phalaenopsis seedling lesions to an accuracy of 89.6%. The detection capability of the system, without classifying the disease type, is as high as 97.2%.