基于特征匹配的微生物弱小目标分割识别方法

摘要

Aiming at the problems of microorganism bacilli that their small objects is hard to be recognised in complex background and that low real-time property in detection,we presented the OpenCV-based small object segmentation operator,introduced feature matching,and designed an automatic image recognition mechanism integrating the above two.The solution,using H-S segmentation and image fusion,and combining morphology filtering and feature matching,solves the problem of difficulty recognition of small microorganism objects.The solution has small computation load and is implemented based on OpenCV with assembly optimisation.Experimental results showed that compared with traditional method, the solution achieved good effect in automatic microorganism image recognition, and was feasible in engineering application.%针对微生物杆菌在复杂背景下弱小目标识别困难，检测实时性不高等难题，提出基于 OpenCV 下的弱小目标分割算子，并引入特征匹配，设计二者相融合的自动图像识别机制。该方案通过 H-S 分割与图像融合，再结合形态学滤波与特征匹配，解决了微生物弱小目标难识别的问题，计算量小。并基于经过汇编优化的 OpenCV 实现，解决了实时性要求高等难题。实验结果表明，相对于传统方法，该方案应用在自动微生物图像识别上可以达到较好的效果，在工程上应用是可行的。