基于SIFT算法的物流货品图像识别系统研究

摘要

In thispaper, we improved the traditional SIFT algorithm and developed a modified multi-angle matchmakingmethod which could cope with changes in size but also soh e the slope and rotation of images. Besides the calculation timeliness and photographing angle could satisfy the requirement of the label identification system and greatly promote the improvement of logistics management technologies.%随着各行业对物流管理技术要求的增长,针对物流技术中的货品图像识别系统进行研究,传统的SIF算法在商品类别控制过程中暴露的现实问题使得物流管理技术停滞不前,本文在此算法上进行改造,研制了一种改良后的多角度匹配方法,该方法不仅能应对图像的尺寸变化,还能解决图像倾角和旋转等问题,且计算的及时性和拍摄角度都能满足货品标签代码图像识别系统要求,对物流管理技术的提高有很大的促进作用.