关键词: 在线自动分级, 特征提取, 图像处理, 权重分析法

Abstract: Aimed at the problem that the automatic level of fruit grading and all fruit grading systems work in the way of off-line at present, one kind of apple on-line automatic grading system is designed for apple picking robots based on multi-information fusion. Weight and color characteristic parameters of the apple are extracted, the data are fused through weight analysis to achieve comprehensive grading based on multi-feature. To improve the accuracy of grading, the weight coefficients are optimized by experiments. Experimental results show that the designed system can achieve grading based on double-feature（weight and color） with the correct rate of over 87%.

Key words: on-line automatic grading, feature extraction, image processing, weight analysis method