Non-destructive Grading of Pattavia Pineapple using Texture Analysis

Pineapple is a tropical fruit with many cultivars in Thailand such as Pattavia, Intarachit, Phuket, Nang-Lea and Phu-Lea. Pattavia pineapple is a popular cultivar in Thailand because of its good taste which can be planted in many areas in Thailand. The fresh Pattavia pineapple can be classified into two grades: (i) Keaw 1 and (ii) Keaw 2 depending on their sweetness and juiciness. In this paper, an automated mechanism to classify the grade of Pattavia pineapple by non-destructive process is proposed. The fundamental idea is to construct the classification model to which it is able to distinguish between Keaw 1 or Keaw 2 images. The texture analysis mechanism is suggested to analysis the pineapple skin. More specifically, Local Binary Pattern (LBP) texture descriptor is suggested to obtain key information of pineapple texture image. The best model produced the excellent results with recorded AUC value of 0.979.