80%。A method of identifying corn and weed using fractal dimension was developed. The obtained original images in the field had to be transformed to the HSI space at first. Image transformation was done using the non-variety of H channel in the HSI space in order to reduce effects from illumination changes,which was in favor of image segmentation. In order to identify corn and weed, three computational formulas of fractal dimension were proposed and compared. Mean fractal dimensions of corn and weed were obtained by fractal software which was programmed using Matlab software. The result shows that the Bouligand-Minkowski method was more effective than other methods, and mean fractal dimension of corn and weed was equal to 1.204 and 1.079 respectively. The identify accuracies of method based on SVM reach 80%.
参考文献
相似文献
引证文献
引用本文
吴兰兰,刘俭英,文友先.基于分形维数的玉米和杂草图像识[J].农业机械学报,2009,40(3):176-179. Identification of Corn and Weed Based on Fractal Dimension[J]. Transactions of the Chinese Society for Agricultural Machinery,2009,40(3):176-179.