针对自然生长状态下的成熟苹果图像的识别问题，提出选用R-G色差模型，改进的Ostu最大类间方差法分割图像。 然后用面积阈值法消除噪声，获取成熟苹果果实的目标区域。然后采用基于OpenCV的圆形Hough变换方法对多个粘连苹果果实进行分离。-For image recognition problem ripe apples natural growth state, the proposed selection of R-G color model between improved Ostu Ot

使用超绿特征分割作物图像。具体来说，首先使用超绿特征将图像灰度化，然后经过中值滤波后，使用Otsu最大类内方差方法将图像二值化，得到分割图像。(The crop image is segmented by super green feature. Specifically, firstly, the image is grayed by Super-Green feature, and then after median filtering, the Otsu maximum intra-cla