Introduction - If you have any usage issues, please Google them yourself
The use of K-means image segmentation, without the use of " red apple" This a priori knowledge, and simply use the characteristics of the image itself- the more obvious differences between the foreground and background