温室大棚中杂草识别系统的应用研究

介绍了一种基于计算机视觉的快速杂草识别算法，能够快速正确地将杂草叶面从复杂的土壤背景中识别出来，从而满足实时地为后续变最控制提供信号的要求。利用AOI测试工具，采集了不同土壤、作物残留物以及各种光照条件下温室大棚内的杂草的图像进行处理，通过提取图像中每一个像素R、G、B的3个分量值计算出(2G-R—B)颜色特征值，将彩色图像转变成灰度图像显示。

Study in Applying the System of Recognizing Weed in Greenhouses

A arithmetic for recognizing weed rapidly based on computer vision is introduced in this paper, which can recognize leaf of weed on a complex soil background rapidly and exactly, and accordingly meets the need of Real-timely giving signals to control successive variables. Using the testing tool of AOI, images of various soil, remnants of crops and weed in greenhouses in diversified illumination conditions are acquired and processed, and the conversion from colorful images to gray images can be effectively realized through picking up the three values of one pixel R, G, B in these images to calculate the 2G-R-B eigenvalue of color.