分割线扫描鱼眼图像有效区域提取算法  

Algorithm for Dividing Line Scan Fisheye Image Region Extraction

作　　者：程德强[1] 刘洋[1] 柳雪[1] 赵国[1] 张佳琳[1] 

机构地区：[1]中国矿业大学信息与电气工程学院,江苏徐州221000

出　　处：《电视技术》2015年第6期1-4,42,共5页Video Engineering

基　　金：国家自然科学基金委员会与神华集团有限责任公司联合资助项目(U1261105)

摘　　要：对于超大广角鱼眼图像的校正,关键是必须准确地提取出鱼眼图像的有效区域,并且同时确定有效区域的中心和半径。现有的鱼眼镜头校正算法针对高清鱼眼视频监控有一定的不足,因此为满足高清鱼眼视频监控的高分辨率需求,在现有算法的基础上提出一种分割线扫描算法。该算法首先划定一条特定水平分割线,由此寻找最接近边界线的范围区域,减少不必要像素点的检测,然后在此区域内精确寻找有效区域边界线。经实验表明,该算法在不降低提取精度的同时,可以显著降低鱼眼图像有效区域提取的时间复杂度和计算量,将以往算法的200万计算量降低到万量级以下。Abstract： For correcting the ultra wide angle of fish-eye image,the key is accurately extracting the effective coverage of fish-eye image, and determining the center and radius of effective coverage. In view of video surveillance of high definition fish-eye, the existing correction algorithms of fish-eye lens have some insufficient. In order to satisfy the high definitionof fish-eye video surveillance,the cut line scanning algorithm basing on existing algorithms is proposed in the paper. Aiming at reducing unnecessary detections of pixel points,a specific level cut line must be determined to find the range and region which most closing boundary, line. Then the boundary line of effective coverage can be precisely tot, nd in the region. The experimental results show that the algorithm can significantly reduce the time complexity and computational complexity when the algorithm extracts efficient coverage of fish-eye image, meanwhile maintain the original extraction precision, and reduce the computational complexity from 2 000 000 to 10 000 or below.

关 键 词：鱼眼镜头 轮廓提取 分割线扫描 

分 类 号：TN919.81[电子电信—通信与信息系统]