一种基于二维直方图的图像分割改进算法

针对基于二维直方图阈值斜分分割算法中斜分区域不定的问题，提出了一种新的二维直方图区域斜分方法，仅通过两条平行线将直方图分成内点区和噪声点区，然后根据图像自身的直方图自适应地确定斜分区域宽度，通过坐标变换，使得待处罂区域减少为原来的3％左右，大大减少了无用计算量，提高了运算速度，并且不影响分割精度。经过和其他算法比较表明，该算法有明显的优越性，可以运用于基于二维直方图的阁值分割中。

An Improved Image Segmentation Algorithm Based on Two-dimensional Histogram

Due to the oblique sub-regional uncertainty in two-dimensional histogram oblique segmentation method,a new two-dimensional histogram oblique segmentation method is proposed. The 2D histogram is divided into interior point area and noise area by two parallel lines. Based on the image histogram in accordance with their own adaptive sub-region, the width of ramp is determined. Then through the coordinate transformation, the region to be processed is 3% of the original value. This method significantly reduces the amount of useless calculation and improves the computation speed. At the same time,it does not affect the segmentation accuracy. Compared with other methods, the algorithm has obvious advantages and it can be applied to all two-dimensional histogram based threshold segmentation methods.