The characterization of digital surface model from stereo imagery over vegetated areas

Some researches in the field of surveying and mapping showed that it held potentials to derive vegetation height from stereo imagery. However, most of current researches were conducted on aerial images or spaceborne images with very high resolutions (about 0.5 m). The resolution of stereo sensors with global coverage is always not so high. The characteristics of digital surface models (DSMs) should be affected by image resolutions because the DSM from stereo imagery was directly determined by the cloud of common points from the pair of stereo images while the automatic recognition of common points depended on the image textures. More stereo imagery has been and will be acquired by ALOS/PRISM and Chinese ZY03 with the resolution of about 2~4 m. Therefore, the characteristics of DSM from high stereo imagery at the resolution of 2-4 m over vegetated areas should be investigated. In this study, an automatic procedure for the extraction of DSM from stereo imagery and a method for the estimation of vegetation height from DSM were proposed. The experiments results showed that the method for the automatic extraction of DSM worked well and the estimation of vegetation height was valid at sparse forest.