用遗传算法提取基元

Extracting Geometric Primitives Using a Generic Algorithm

基元提取是基于模型的计算机视觉的一项重要任务.Hough变换是基元提取的最常用的方法,然而,在许多情况下,它的存贮开销太大而难以让人接受.近些年来,有些人用统计学方法来提取基元,但如何构造合适的代价函数仍是一个困难问题.基元提取等同于寻找具有多个局部极小值的代价函数的最优解.遗传算法(Genetic algorithms)能够有效地在搜索空间中找出全局最优解.为实现有效的基元提取,作者从几何数据点中随机地选择一组最小子集,然后用遗传算法对几何数据点进行动态划分,经过若干次进化将得到一个最优划分,与之对应的基元和基元所对应的数据点将被提取出来.这种算法可用于多种基元和多个基元的提取.

Extracting geometric primitives is an important task in model－based computer vision.The Hough transform is the most common method of extracting geometric primitives,however,its space requirements are too large,and the number is an exponential function of the dimension of the parameter space.Recently,methods derived from the field of robust statistics RS have been used for this purpose,however,the most serious difficulty with the RS approach is its actual robustness.The extracting geometric primitives is equivalent to finding the optimum value of a cost function which has potentially many local minimum value.GA can be designed to efficiently locate an approximate global maximum in a search space.In order to extract the geometric primitives,Authors choose a number of minimal subsets randomly from the geometric data.Then the geometric data are partitioned dynamically by GA.The generic process converges on this ideal partitioning result though successive iterations,subsequently,the associated geometric primitives which are taken as a description of the geometric data are extracted.The resulting extraction algorithm can be used with a wide variety of geometric primitives and geometric data．