【摘要】 在人脸检测、火灾识别等应用中,应用物体的轮廓特征来检测和定位目标物体是一种有效的方法。本文提出一种基于连通区域面积阈值化的实现算法,可以同时实现噪声消除与轮廓提取,并据此定位图像中的物体目标。实验结果显示,只要噪声面积没有超过物体面积,应用该算法即可实现噪声完全消除,并能够得到连续的无交叉的单像素宽度的物体轮廓,且轮廓不变形。更多还原

【Abstract】 It is an efficient way to detect and locate an object by its contour features in an image in some applications such as face detection or fire recognition. This paper gives an algorithm and its implement based connected regions and their size threshold to not only extract the contour of the object but also erase the noise at the same step in an image processing. Furthermore, the object could be located according to the contour. As shown in the experiments, the noises could be erased clearly when the area of each noise is less than that of the object, and especially a single- pixel- width contour of the object could be presented which is also no- interval, no- cross, and not distorted from the object. The good precision of detection and capability of noise erasing are exhibited.更多还原