To accurately extract buildings from remote sensing operations

In recent years , Domestic and foreign scholars have proposed a variety of shadow based building extraction side Law . H . Liu , and so on 3 uses the shaded area in the His in the model I value Changes small , Value large attribute , the creates an exponent for detecting shadows , and successfully detect shaded areas with histogram thresholds ; D . Mehmet , and so on [4 on image segmentation based on shaded area Tint / Brightness ( h/i ) high value and detect shaded area ;C . Jaynes , and so on 5 and G . D . Finlayson , and so on 0 the observes shaded areas like The element has a low brightness feature , present the corresponding shadow detection party law . Most of the above methods pass the threshold after extracting the corresponding spectral features Values the precision of the extracted Shadow area is mainly the result of, depend on thresholds and extract building directly from Shadow area, instead of to consider the height of the building . when encountering more tidy tree shadows can be to create shape features similar to building shadows , some special material When the roof can be mistakenly divided into shadows , will result in a building target. False check for information . In view of this , This article combines DSMData , The presents the A Color Remote sensing image based on DSM and Shadow law , In order to reduce the false inspection of the building area brought by the non shaded area , Extract the outline of the Building more accurately .