Classification of Ripe and Unripe White Berry Fruit Using Thermal Image Processing

Since white berry has much medical and edible benefit, it is known as a popular fruit. This research is carried out in order to classify ripe and unripe berries from each other using active thermography method. In this research 70 berries have been used as experimental samples randomly selected from a tree. They were sorted to “ripe” and “unripe” class by 5 expert people with attention to the color and texture of fruits. The temperature changes of samples by heat shock induction were recorded with a thermal camera and coefficients of first order, second order and logarithmic equations fitted to temperature-time graphs were employed for classification in MATLAB software. Using Principal Component Analysis (PCA) method, in MATLAB software, causes to use only first order equation coefficients, with an accuracy of 90 percent. Using the results of this research, the speed and accuracy of classification can be effectively increased in berry classification lines.