色差边缘检测方法在杂草识别中的研究

田间作物与杂草的识别是实现变量喷洒除草药剂的关键.为此,提出了田间作物和杂草的叶子颜色信息以及HSI颜色模型研究的基础上,提出了一种与HSI空间3分量相互分离的特点相结合的色差边缘检测方法,并将实验结果同传统的Sobel方法进行了比较.结果表明,该方法充分利用了田间作物与杂草的彩色信息,能够快速准确地检测到图像的边缘,边缘连续性好,能很好地与背景分离,而且能够满足实时检测的要求.

Study of Weed Recognition from Field Crops Based on Color Difference Edge Detection Method

Weed identification technology by computer vision is the key to the selectively spraying herbicide.In order to do so,the first thing is to realize the edge detection of the weed and the crop.On the base of studying the HSI color space,a new method is proposed in this paper,which is a color deference method taking full advantage of the character of HSI color space.The experimental results revealed that this method can detect color image edge fast and accurately with low error to the old Sobel based on gray image.