The Road Extraction And Representation Model Based On High Resolution Remote Sensing Image

The object extraction form remote sensing image is the research focus of remote sensing, GIS, computer vision and so on. The applications of high resolution remote sensing images have provided more possibility and higher accuracy for the object extraction from remote image. In the research of the object extraction from remote sensing image, road extrxction is an important one and is far-reaching significance for renovation of GIS data, image matching, object detection, and digital mapping automatization.In the paper, the road extraction and representation based on high reaolution remote sensing image is studied. A total matching model for road extraction is developed especially. The main results are as follow:(1) The characteristic of high resolution remote image especially the characteristic of the road in the high resolution image is analysed, which is the theory base for the road extraction.(2) Based on the analysis for the extracted line object, a vectorization road representation model by three spline function is obtained. The model can reduce the data redundancy effectively and retain continuity and smooth of the line object.(3) Based on the characteristic of line road of high resolution remote sensing image and hit idea of mathematical morphology, a rectangle inside-outside total matching model for extracting line road is proposed. And the model has good automatization level.(4) Based on spline function and total matching idea, a polygon inside-outside total matching model is developed for curve road in the high resolution remote sensing image.(5) Using object-orient ideal, a software system has been designed an realized for the road extraction for high resolution remote sensing image based on the proposed models.