Date maturity and quality evaluation using color distribution analysis and back projection

A new color grading method is proposed in this paper to provide an automatic and intuitive way of evaluating the maturity and quality of harvested dates. Different from other existing methods that rely on complicated machine learning or artificial intelligent algorithms, this method uses 2D histograms of colors in each grading category to determine the co-occurrence frequency. Based on this color analysis result, a mapping matrix is generated for back projecting the input colors of the fruit to their designated color indexes. The date color, which strongly correlates to maturity and quality, can then be graded by analyzing the resulting back projected color indexes. With the proposed method, the operator is allowed to set or adjust the cutoff points of different maturity or quality groups of date for automatic inspection. This paper uses Medjool date grading as an example to demonstrate the performance of the proposed algorithm. It is suitable and can be easily adapted for other fruit or vegetable grading applications. The proposed method is proven accurate and user-friendly and has been implemented and used for commercial production for date maturity evaluation.