论文部分内容阅读
牧草自动识别是对普通数码相机获取的牧草数字图像进行预处理、特征提取与特征匹配等环节处理,达到利用计算机实现牧草分类的目的。牧草自动识别具有成本低,易于采集,准确性高等优点,是实现草地数字化的基础。图像预处理是保证识别精度的关键环节,本文以典型草原优质牧草禾本科种子为研究对象,研究图像的预处理方法,获取感兴趣区域(Region of Interest,ROI)。主要步骤包括:首先对图像进行去噪、灰度化、二值化处理,然后对二值图像进行形态学腐蚀、膨胀运算,确定种子边缘,最后根据种子主体位置建立坐标系,分割原始图像,获取ROI。为验证预处理方法的有效性,本文利用主成分分析(Principal Components Analysis,PCA)提取特征,对20个样本的禾本科牧草种子1000幅图像进行识别,平均识别率达到94.6%。
Automatic recognition of forage grasses is to pretreat the forage digital images obtained by a common digital camera, and to extract and match features, and to achieve the purpose of forage classification by computer. Automatic grass forage recognition has the advantages of low cost, easy collection and high accuracy, and is the basis for realizing grassland digitalization. Image preprocessing is the key step to ensure the accuracy of recognition. In this paper, grass grasses of typical grazing grassland are taken as research objects, the image preprocessing method is studied, and the region of interest (ROI) is obtained. The main steps include: Firstly, the image is denoised, grayscale and binarized; then the binary image is subjected to morphological erosion and expansion operation to determine the edge of the seed; finally, the coordinate system is established according to the position of the seed main body to segment the original image, Get ROI. In order to verify the effectiveness of the pretreatment method, we use principal component analysis (PCA) feature extraction to identify 1000 images of gramineous forage seeds from 20 samples with the average recognition rate of 94.6%.