相似背景下的苹果图像分割方法仿真研究

Similar Background of Apple Image Segmentation Method Simulation

研究基于图像的苹果分割问题。针对传统基于背景颜色像素的区别,利用像素相减的方法进行苹果图像的分割,当图像中,苹果颜色和树叶颜色相近的时候,边缘像素相减的结果变化不明显,算法根据结果不能完整的分割苹果图像的问题。本文提出基于改进的主成分分析算法的苹果图像分割方法。通过像素概率知识,将苹果图像的颜色特征进行归类,运用边缘像素排序的方法,描述苹果图像边沿的细节信息。排除相似边沿像素的干扰。实验证明,这种改进后的方法实现了在与自身颜色相似情况下苹果图像的完整分割,取得了满意的效果。

Research apple image segmentation problem.The traditional method for apple image segmentation is based on the subtraction of color pixels of background,when apple colors close to the leave colors,the result of edge pixel subtraction does not change significantly.This paper put forward the principal component analysis based on the improved algorithm of image segmentation method.Through the pixel probability knowledge,the color characteristics of apple images were classified,and the details of th apple image were described using method of edge pixels sorting.The interference of similar edge pixels was eliminated.Experiments show that the improved method can realize the complete apple segmentation when apple colors are similar to the leaves or bakchground.