Lingxia Chen and Michael Bushell Pages 179 - 183 ( 5 )
Background: During the process of extracting three-dimension (3D) dynamic characteristics of remote sensing image, the extraction effect is vulnerable to be influenced by gray value and gradient of remote sensing image. When the traditional algorithm is used in extracting 3D dynamic characteristics of remote sensing image, due to the changes of pixel gray value and gradient in remote sensing image, pixel weights are interfered, and the traditional algorithm is unable to tract the change of picture element, resulting in poor extraction effect of 3D characteristics.Method: For this, a method for 3D dynamic characteristics extraction of remote sensing images is proposed based on optical flow analysis. 3D space information of reflection point is extracted from remote sensing image sequence as the basis of 3D dynamic feature extraction. The change of picture element in every frame of remote sensing image is tracking, to establish the corresponding relation of picture element and remote sensing images, so as to achieve the extraction of target image 3D dynamic feature. Result: Simulation experimental results show that the improved algorithm can extract 3D dynamic characteristics from remote sensing images accurately, and the effect is satisfactory.
Remote sensing images, three-dimensional dynamic characteristics, feature extraction, 3D dynamic, flow analogs, image process.
College of Tourism and Resources Environment, Xianyang Normal University, Xianyang, 712000, Department of Computer Science, Rutgers University, New Brunswick, NJ