Fruit Quality Evaluation in The Maturation Process of Blueberries Using Image Processing

Abstract

Blueberries' quality does not change uniformly during ripeness. Blueberries should be harvested fully ripened at the post-climacteric stage with an excellent indicator including consistent color, taste, and ease of removal from plant as excellent indicators. Therefore, the blueberries are not harvested until it has the desired blue color. The reliance on human perception on the fruit's taste and appearance might cause inconsistency and inaccurate judgment of the fruit maturation. This study aimed to develop an image processing algorithm capable of classifying blueberry maturity stages. The Bluecrop Northern highbush blueberry was harvested at five different stages of maturity based on visual grading of the fruit color (green, green-red, red, red-blue, and blue) from various fruit positions on the tree. Image processing with discriminant analysis accurately classified maturity stages at 98.3% accuracy. The image quality attributes of blueberries changed significantly at different maturity stages. Overall, most image quality attributes correlated strongly with well-performed blueberry physicochemical properties. This study showed that image processing during the blueberry maturation process could be a reliable and comprehensible method for estimating changes in color, shape, weight, and ultimately changes in specific physicochemical properties. This study also provided a practical evaluation of the maturity stages and physicochemical properties, which were predicted using image processing.