摘要
针对当前医院外科手术器械管理流程中人工清点费力、低效、易出错的问题,提出一种基于机器视觉的器械清点分类方法。该方法利用形态学骨架提取与像素滑动检索方法确定器械的数量与位置,并将数量与位置作为先验信息预先建立待测物搜索框,再结合模板匹配方法对器械进行分类。实验结果表明,比较传统匹配方法,文章提出的方法对于多数量、多种类器械清点的准确率得到明显提升,清点速度可以匹配人工清点水平。
Manual inventory is laborious,inefficient and error-prone in the management of surgical instrument in hospital,currently.Method aiming to realize auto-inventory of surgical instrument,based on machine vision,was proposed.By extracting the skeleton of instrument based on morphology and retrieving the key pixels,the number and location of the instrument was obtained.The search boxes,in which the object to be detected existed,were established according to the number and location as priori information in advance.And then the instrument was classified successfully by the improved template matching method.The results show that,the method proposed performs well in the accuracy of inventory,compared with the traditional matching method,and the inventory speed matches that of the manual inventory.
出处
《机电一体化》
2022年第4期51-57,共7页
Mechatronics
关键词
手术器械
机器视觉
清点
骨架提取
模板匹配
surgical instrument
machine vision
inventory
skeleton extraction
template matching