中文摘要：

为更有效地估测苹果产量，分析自然场景下3种颜色模型提取成熟苹果果实的效果。根据成熟苹果的颜色特征，利用RGB、L＊a＊ b＊和YCbCr颜色模型中的分量对彩色图像进行灰度化处理，采用OTSU法进行二值化处理和形态学运算。对逆光和顺光下的苹果图像进行苹果果实面积和噪声面积比较分析，比较结果表明，3种分量均能分离出苹果，a＊分量在保留最多苹果面积的情况下噪声较小， r‐g色差分量会削弱部分苹果面积， Cr分量噪声较大，给出了各分量的适用范围。

英文摘要：

To estimate the yield of apple more effectively ,the effects of three color models to extract ripe apples in natural scenes were analyzed .Based on the color features of ripe apples ,color image was grayed using the component of RGB ,L＊ a＊ b＊ and YCbCr color models ,binaryzation image was obtained through adaptive threshold segmentation and morphological operations . Finally ,apple fruits area and noise area were compared and analyzed under direct sunlight and backlight ,and experiment results show that three kinds of components can separate the apples .Most apple areas are retained while guaranteeing minimum noise using a＊component ,a part of apple areas are weakened using r‐g color component ,and larger noise emerges and the scope of ap‐plication of each component is presented using Cr component .