The Method Research Of Pattern Recognition About Satellite Remote Sensing Vehicle Based On Morphological

QUICKBIRD satellite is a hi-resolution optical satellite for business in Digital Globe Corporation in Unite States; the resolution is 61 centimeter high. This kind hi-resolution satellite can work day and night, it has higher space resolution, and it is a real-work system over time and space. Hi-resolution satellite is used in military, remote sensing, ocean, landform, hydrology, zoology, communication, layout, mapping and so on. It has been an important tool in people's live and work; meanwhile, it is used in traffic area deeper. Because vehicle can be seen clearly in the QUICKBIRD satellite image, so people put more attention in the traffic detection assisted the QUICKBIRD satellite image. This paper uses the QUICKBIRD satellite image as research object.The paper research the main things is to deal and recognize the traffic road image from hi-resolution satellite, to pick up the pixel set which contain the vehicle characters, which contain the vehicle characters pixel set and the alike vehicle characters pixel set. Next, according to set an adaptive Artificial Neural Network to process these pixels set with selection, confirming and classification. Finally, it reaches to recognize and classify vehicle in the hi-resolution satellite ground traffic road image. The application of Morphological is main technology in the paper, by morphologically processed image; it makes the pixel matrix more distinction and standardization, and is good for next step on the pattern recognition. According to create adaptive Artificial Neural Network system, it can calculate the texture characters after images processed; finally, it will reach a satisfying outcome.