基于机器视觉图像的多目标提取识别

Recognition and Extraction of Multiple Targets Based on Machine Vision Image

为了实现农作物图像多目标的准确识别,提出了四连通或八连通的连通域判别法。利用MATLAB的数值计算、图形处理和可视化建模以及动态仿真等功能,对黄瓜植株花朵图像预处理,去除干扰目标,对目标处理结果进行整合,进而得到多目标处理结果。结果表明该方法依据图像信息即可有效地实现多目标的分离和识别。

The discriminance of 4-connected or 8-connected domain was presented to realize accurate identification of the target crop images. Using MATLAB functions, such as numerical computation, graphic processing, visual modeling and dy- namic simulation, the images of cucumber flowers were preprocessed and the jamming images were removed, and then single images were integrated to obtain the multiple images. The results showed that this algorithm could separation and recognization of the multiple objects effectively with machine vision.