大数据环境下异质模糊图像角点特征提取仿真

Simulation of Corner Feature Extraction in Heterogeneous Fuzzy Images in Big Data Environment

运用传统方法在大数据环境下进行异质模糊图像角点特征提取时,存在角点遗漏、提取不精确等问题。为解决上述问题,提出一种基于改进SUSAN的异质图像角点特征提取方法。在大数据环境下,采用高斯滤波对异质模糊图像进行去噪处理,通过改进SUSAN角点检测方法建立圆形模板,采用圆形模板邻域内像素灰度值中值代替模板中心像素灰度值作为模板"核"来检测区域目标角点。对检测出的图像角点所在边界进行细化;采用多角点结合的提取方法对经过处理的异质模糊图像角点特征进行精确提取。仿真结果证明,所提方法对图像角点特征提取的精度较高,大大缩短了提取用时,并且不会对图像的内容及细节造成破坏;面对噪声及其它环境的干扰,该方法的运算效率没有受到太大的影响,说明该方法的抗噪性比较好,为图像处理工作提供了更好的依据。

Traditional method leads to the missing corner and inaccurate extraction during the extraction of corner features from heterogeneous fuzzy image in big data environment. Therefore, a method to extract corner features from heterogeneous images based on improved SUSAN was proposed. In big data environment, Gaussian filter was used to remove the noise from the heterogeneous fuzzy image, and the circular template was established by improving SUSAN corner detection method. Moreover, the median grayscale value of pixel in the neighborhood of circular template was used to replace the gray value of central pixel of template. As the "core" of the template, it was used to detect the corner points of regional target. After that, the boundary of the detected corner points was refined. The corner feature of the heterogeneous fuzzy image after the treatment was extracted accurately by the extraction method combining many corner points. Simulation results verify that the proposed method has high accuracy in extracting corner features. Meanwhile, this method greatly shortens the extraction time without damaging the content and details of image. Facing to noise and other environmental interference, the operation efficiency of the proposed method is not greatly affected. Consequently, the proposed method has better anti-noise performance, which provides a better basis for the image processing.