灰度图像矩的一种快速算法  

AN ALGORITHM TO FAST COMPUTATION GEOMETRIC MOMENTS FOR GREY LEVELIMAGE

作　　者：张仲选[1] 王冰[1] 职秦川[1] 周明全[1] 

机构地区：[1]西北大学计算机科学系,西安710069

出　　处：《计算机应用与软件》2004年第7期3-5,共3页Computer Applications and Software

基　　金：国家自然科学基金项目(60 2 71 0 32 )资助

摘　　要：大多数矩的快速算法仅适用于二值图像。本文给出了一种适用于灰度图像的矩快速算法。对N×N大小的灰度图像 ,引入 6个大小为N的数组 ,将求矩过程中重复计算的量一次计算后存入数组 ,减少了重复和冗余计算。虽然增加了 6个数组 ,内存的相对增加仅为 6/N。该算法原理简单 ,求矩结果精确 ,计算量大大减少。和直接法相比 ,加法和乘法运算次数仅为其 40 4%和2 2 6%。和Yang s算法相比 ,加法次数相同 ,而乘法次数约为其 67 6%。This paper propose a method to fast computation geome￣tric moments for gray level image while the most literature to computation geometric moments for binary image.6 arrays are defined to store the value which are repeatly used during the process of computation moments,so the computational complexity decreased greatly.Although the array is used the memory increasing is only 6/ N percent.The principle of the algorithm is simple and the result is accurate by the test.The computational complexity is also given.Compared with Direct method,which shows that our method decreases additions and multiplications to 40.4 percent and 22.6 percent respectively.Compared with Yangs method in Ref.[10],which shows that the addi￣tions is the same while multiplications decreases to 67.6 percent.

关 键 词：计算机视觉 灰度图像矩 图像变换 数学形态学 图像分析 模式识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]