Affiliation:
1. Northeast Forestry University
Abstract
The paper proposes a digital image extraction and segmentation algorithm based on color features. The traditional transformation from RGB model to HSI model is improved, meanwhile the leaf color information is extracted by similarity distance between pixels. The green component of leaf image in the RGB model is strengthened, and then the digital image is transformed to the HSI model by the improved method. Finally the image is divided by similarity distance of pixels’ H weight which determines whether the pixel belongs to the blade. The results of simulation experiment shows that this algorithm can achieve a good image segmentation effect, and it has a high degree of accuracy as well as a clearly distinguish degree and many other advantages such as good consistency with human visual system. It completely meets the effectiveness and clarity requirements of image segmentation.
Publisher
Trans Tech Publications, Ltd.
Subject
Mechanical Engineering,Mechanics of Materials,General Materials Science
Reference12 articles.
1. WANG, Fang-Yong. , LI, Shao-Kun. : Obtaining Information of Cotton Population Chlorophyll by Using Machine Vision Technology. J. ACTA AGRONOMICA SINICA, 33(12), 2041-2046 (2007).
2. Clement, A., Vigouroux, B.: Unsupervised segmentation of scenes containing vegetation and soil by hierarchical analysis of bidimensional histograms. J. Pattern Recognition Letters, 24: 1951—1957(2003).
3. Aitkenhead, M. J., Dalgetty, I. A., Mullins, C. E, et al.: Weed and crop discrimination using image analysis and artificial intelligence methods. J. Computers and Electronics in Agriculture, 39: 157-171(2003).
4. Tang, L., Tian, L., Steward, B. L.: Classification of broadleaf and grass weeds using global wavelets and an artificial neural network. J. Transactions of the ASAE, 46(4): 595-596(2003).
5. Kavdir, I.: Discrimination of sunflower, weed and soil by artificial neural networks. J. Computers and Electronics in Agriculture, 44: 153-160(2004).
Cited by
4 articles.
订阅此论文施引文献
订阅此论文施引文献，注册后可以免费订阅5篇论文的施引文献，订阅后可以查看论文全部施引文献