Research And Implementation Of Massive Terrain Surface Image Processing Method

In recent years,the size of terrain images which are applied in the field of petroleum exploration is rapidly growing.So efficient image processing technology is required to research and analyze those “big” image data resources.However,traditional image processing method based on memory can't satisfy the requirement of image data analysis and processing in the field of petroleum exploration at present.In this paper,the main goal is to research and implement efficient processing methods which can deal with those “big” terrain image.These methods include pretreatment,display,move,mosaic,clipping and relocation.At the same time,these methods meet the exploration design system requirements.The major research content is outlined as follows:(1)Design and implementation of pyramid structure are used to store massive terrain surface image.Pyramid supports multi-resolution display and real-time interaction because of structural features of multi-resolution,layers and blocks.At the same time,combining pyramid with effective memory buffer management benefits to manage massive image and access data rapidly,and processes massive image smoothly while zooming or moving image.(2)Research and implementation of scan-line algorithm which rely on “Active Edge Table” is used to cut image.The scan-line can achieve image cropping of random polygon area,and optimize performance by less intersections and flexibility of operation.(3)Research of three image mosaic algorithms include geographic coordinates,feature points and massive satellite tile.The geographic coordinates algorithm uses block technology in order to reduce I/O operations and optimize performance.The massive satellite tile algorithm uses buffer scheduling mechanism to optimize performance.(4)Research of multi-point relocation algorithm is used to correct image and assign relevant geography information.When images perform geometric conversion,transformation model is important.So,two transformation models about geometric polynomial and thin plate spline are studied to complete image geometric transform.Relevant tests and the actual system tests show that correlation theories and algorithms which are put forward above research can process massive terrain surface image data efficiently.In addition,those theories and algorithms can satisfy performance requirement of image data processing in the seismic exploration design system.