Development of high-precision measurement method of fruit color based on digital image analysis

在果实颜色的育种和采前采后研究中，需要一种有效的颜色评价方法来代替肉眼的主观评价。我们使用基于内容的图像检索中使用的一系列数据处理和统计分析来评估甜瓜和草莓果肉颜色，并评估了该方法的有效性。该方法将图像中的颜色信息汇总为颜色特征，计算颜色特征之间的土方距离（EMD），并基于EMD矩阵进行多维缩放。该方法揭示了重要的颜色特征，如绿瓤瓜的黄绿色强度，红瓤瓜的绿色和红色部分的相对大小，草莓果肉的白色和红色部分的相对大小，而不需要明确定义这些特征。与传统的评价方法相比，该方法能更好地检测品种间的颜色差异。该方法提供了有效、客观的果实颜色指标，在研究和育种计划中显示出相当大的前景。

In breeding and in pre- and postharvest studies of fruit color, an effective method for evaluating color is needed to replace subjective evaluations by eye. We used a series of data processing and statistical analyses used in content-based image retrieval to evaluate melon and strawberry flesh color, and assessed the efficacy of this approach. This method relies on summarizing color information from images into color signatures, calculating the earth mover's distance (EMD) between color signatures, and multi-dimensional scaling based on an EMD matrix. This method revealed important color features, such as the yellowish-green strength in green-fleshed melons, the relative size of the green and red parts in red-fleshed melons, and the relative size of the white and red parts in strawberry flesh, without the need for an explicit definition of these features. Moreover, this new method performed better at detecting color differences among cultivars than traditional evaluation methods. The method provides effective, objective indicators of fruit color, and shows considerable promise for use in research and breeding programs.