Research On Traffic State Estimation From Satellite Remote Sensing Images

With the advance of high-resolution commercial remote sensing system, detecting traffic flow data from remote sensing images at large regions has become an attractive option in the fields of traffic information collection technology. The main goal of the research is to extract the vehicle velocity from high-resolution satellite remote sensing images, and to further determine the traffic status.The study includes the following aspects:(1) The time gap between the acquisition of the panchromatic and multispectral data can be used to derive velocity information. In tIHS paper, I will analysis how to use the time gap to derive velocity, while key technological process will be build and introduced;(2) Using the Proposed method of extracting vehicle velocity, the sample image will be processed, including Gaussian smoothing and enhancement for Panchromatic image, IHS Transforming for multi-spectral images, and gray/feature-based image registration. Then coordinates of vehicles in the panchromatic and multi-spectral images will be extracted. Further, the speed of vehicles will be estimate;(3) Eventually, two experiments and results are described and assessed. Experiment 1:Integrate the panchromatic and multispectral images of a main road in Beijing, using the matching method based on gray-scale; Experiment 2:Matching the panchromatic and multispectral images of one section of Badaling high-way based on the characteristics. Experimental results coincide with the actual traffic conditions.As a new technology for the traffic estimation, the proposed method in tIHS research will be valuable for exploration in traffic field.