Vision-based strawberry feature extraction and quality evaluation model.

In order to improve the efficiency of strawberry sorting and increase the data differentiation of strawberry quality evaluation methods, a vision-based strawberry quality evaluation model was proposed. Based on the traditional evaluation method of strawberry quality of the industry standard, the color, size and shape of strawberries were extracted by image processing methods, so as to quantify the ripeness, defects, quality and shape features of strawberries, and realize the grade evaluation and specification evaluation of strawberries. Furthermore, the entropy weight method was adopted, the weight factor was introduced, and a comprehensive evaluation model of strawberry quality was established to realize the automatic grading of strawberry quality. A strawberry quality sorting device was designed to verify the effectiveness of the model. The test results show that the accuracy of automatic grading of strawberry grade and specification was 96.7% and 98.4%, respectively, the accuracy of comprehensive quality grading reaches 96.3%, the success rate of strawberry sorting reaches 95.3%, the average time of quality grading of each strawberry image is 37 ms, and the average time of each strawberry sorting is 0.87 s. This study provides a reference for manipulator-based automatic strawberry sorting. [ABSTRACT FROM AUTHOR]