A Study And Implementation Of Photogrammetry Technology Based On Web Service

Photogrammetry is the art, science, and technology of obtaining reliable information about physical objects and the environment through the processes of recording, measuring, and interpreting photographic images and patterns of electromagnetic radiant energy and other phenomena. This technology has wide range of applications including aerial remote sensing, 3D scene reconstruction, traffic accidents and crime scenes reconstruction, archeological virtual reality, digital museum etc.Web Service is a new internet technology that has seen rapid development in recent years. Taking the advantage of standard XML, it can describe any data in a platform independent way. So that inter-platforms and inter-programming languages communication can be achieved.In this article, we introduce a photogrammetry architecture based on Web Service. This architecture builds function modules according to the key photogrammetry methods, ideas and theories such as projective geometry, epipolar geometry and so on. Then it organizes the function modules, which is closely related to photogrammetry methods, in a Web Service technique fashion. These modules can be combined into specific application procedures like single view, two views, etc. Using this architecture, a 2D or 3D reconstruction of the scene with satisfactory accuracy can be retained. Different from traditional systems, modules in this architecture are capsulated by Web Service and broadcasted on net using SOAP protocol. Programs can invoke function modules on other system platforms in terms of relevant Web Service standards; therefore a build-up on local platform will be easily achieved. So a great deal of time and efforts will be saved from duplicate development of basic function modules.Functionally, we achieve 2D reconstruction through obtaining transformation relations between real reference object and reference object in image and gaining orthograph image. Referring to 3D reconstruction, firstly we calibrate camera which takes pictures of the scene to obtain intrinsic matrix and distortion parameters; secondly, take pictures of the scene using the camera just calibrated from two different view points; thirdly, choose corresponding points from two images and calculate fundamental matrix from the points, then decompose the fundamental matrix to retrieve a pair of camera matrix that only a projective ambiguity to the real camera matrix; fourthly, using absolute dual conic(ADC) to obtain that projective relation; fifthly, employ the optimal triangulation method to get 3D model in projective space then use the projective relation mentioned above to update projective 3D model to similarity 3D model; after that compare the model to the references in real world to get accurate measurable 3D reconstruction.