Water content of fresh tobacco leaves was judged with the computer vision technology. Images of fresh tobacco leaves were gathered under sunlight and image segmentation algorithm was performed with the Matlab software. The parameters of r, g, H, acreage, width, shrinkage rate, mean value, consistency and entropy were extracted and used as the indexes of comprehensive evaluation so as to reduce the inaccuracy of water content estimation based on a single parameter, and they were ranked with the three-demarcation AHP method. The results showed that leaf width, leaf area and mean value had relatively great weights and, therefore, are recommended as the indicators in rapid determination of water content of fluecure tobacco leaves.
NETL
NSTL
万方数据
维普
