基于改进的四叉树Sobel算子图像分割研究

研究图像分割精度提高问题,由于图像噪声的干扰,造成图像模糊。传统的Sobel图像分割方法容易造成图像分割不清晰和对比度不明显。为解决上述问题,应用Sobel算子检测出数字图像真正的边缘,并进行边缘提取;最后运用四叉树算法对去噪图像进行多区域目标分割,并对分割的不同区域进行合并,有效地解决了因局部图像的叠加和分割不准确的难题,并且通过仿真实验显示,本文提出的改进型算法,也就是利用四叉树Sobel算子,能够有效地去除图像分割所带来的去除噪声,大大地提高升图像边缘的清晰度,是一种有效地图像分割方法。

Image Segmentation Method Based on Sobel and Quadtree

Improve the accuracy of image segmentation optimization problem.Because image noise interference,the traditional Sobel operator segmentation method likely to cause segmentation is not clear,the contrast is not obvious,low accuracy problem of image segmentation defects,an improved Sobel operator quadtree Digital Image Segmentation.Firstly,the improved median filter to remove noise,then the nature of image features on the early image segmentation,applied Sobel operator to detect the true edge of the digital image and the edge extraction;Finally,we use the quadtree algorithm denoising multi-zone target image segmentation,and segmentation of different regions to merge,effective solution to the partial image overlay as a result of the segmentation problem is not accurate.Simulation results show that the improved Sobel operator quadtree segmentation method can effectively remove noise and improve the clarity of the image edge segmentation,is an effective method of image segmentation.