Extracting rubber plant in Quickbird imagery based on Support Vector Machines

In recent years high-resolution space borne images have disclosed a large number of new opportunities for medium and large-scale rubber plant mapping. Some traditional algorithms used for hyper spectral remote sensing image classification have some problems such as low computing rate, low accuracy. According to SVM theory, the Rubber plant classification model based on SVM was constructed, by experimenting with Quickbird imagery, the classification accuracy of SVM using four different kernel functions were analyzed, the results indicate that the four types of kernels for training and classification of SVM can be used for rubber plant classification.