A new parallel ortho-rectification algorithm in a cluster environment

Ortho-rectification of remotely sensed image is one of the most important processing steps in photogrammetry. As the data of remotely sensed image become larger and larger, the computation of ortho-rectification increases greatly. Parallel processing technology can quicken the processing speed of image ortho-rectification. Based on the analysis of ortho-rectification algorithm and intuitionistic parallel ortho-rectification algorithm, a parallel algorithm oriented to massive remotely sensed ortho-rectification was given. Through the overlap of computation and data transmission, processing speed was improved, and by local caching and prior searching, redundant data transmission was avoided. It is proved by experiment that, in a fast network, this algorithm can achieve good performance and is well scalable, and is suitable for fast processing of massive image in photogrammetry.