根据迁移策略并行遗传算法的模板匹配研究

Template Matching Research Basing on Migration Strategy Parallel Genetic Algorithm

模板匹配是提高智能识别与实物提取的有效途径。为了提高传统模板匹配方法的效率和精度,提出了一种遗传算法协同迁移策略的模板匹配算法,研究本算法在图像匹配过程方面的优选性及实现。遍历式搜索是匹配精度最高的传统算法,其时间复杂度以平方的规模增长。该算法在保证精准匹配的同时有效地减少了每个搜索位置的计算量,且仿真结果表明匹配结果基本稳定,准确率达99.3%以上,高于一般算法,优化后能满足工业实时性的要求,模板匹配时间同模板大小成反比,受软硬件影响。

Template matching is an effective way to improve intelligent identification and physical extraction. In order to improve the efficiency and accuracy of the traditional template matching method, a template matching algorithm based on the cooperative migration strategy of the genetic algorithm was proposed. The optimization and implementation of the algorithm in the image matching process were studied. Traversal search is the traditional algorithm with the highest matching accuracy, and its time complexity grows on a square scale.The algorithm can effectively reduce the amount of calculation for each search position while ensuring precise matching. The simulation results show that the matching result is basically stable with an accuracy rate of 99.3% or more. It is higher than the general algorithm and can meet the requirements of industrial real-time performance after optimization. The template matching time is inversely proportional to the template size and is affected by hardware and software.