Retrieving urban objects using a wavelet transform approach

The multi-scale framework of the wavelet transform is used to represent the edge information for the retrieval of traditional industrial complexes from digital aerial photographs in an urban area. A set of retrieval parameters is used to extract and enhance the edge information for the target objects. These parameters include summary statistical indices for the detail sub-images of the wavelet transform, an edge-pixel index, weighting schemes for these indices, the level of the wavelet transform, and the search window size. All of these retrieval parameters are effective, except for the weighting schemes, in retrieving the target objects. As a result, a majority of retrieval rates reach 60 percent to 90 percent.