Font Size: a A A

Study On Technology Of Real-time Electronic Image Stabilization

Posted on:2016-05-03Degree:MasterType:Thesis
Country:ChinaCandidate:X D WangFull Text:PDF
GTID:2308330476954749Subject:Optical Engineering
Abstract/Summary:PDF Full Text Request
With the development of multi-media and computer network technology, the digital image, as one of the most intuitive media, has been more and more widely used. The high quality of the video is also demand increasingly. The video is not only required high resolution and clear, but also is required stable without jitter. Therefore, the research on technology of real-time electronic image stabilization(EIS) has become an important content of digital image processing.Firstly, it analyzes the imaging principles of camera and the reason of the video jitter. It introduces the principles of EIS and the main algorithm of motion compute in detail. It compares and analyzes the advantages and disadvantages of the projection algorithm, the block algorithm and the feature matching algorithm.Secondly, it proposes a fast image stabilization algorithm based on projection. It realizes the accurate matching through the interpolation in the projection of reference frame and current frame. It designed a motion filter based on kalman filter, which separates the normal movement and random jitter of the camera in complex environment effectively.Thirdly, it proposes a full frame compensate algorithm to compensate the missing edge area in the motion compensation. It improves the quality of the compensated images by using the reference frame to compensate the edge area of the current frame through fusion mosaic.Finally, it designed a real-time electronic image stabilization system based on TMS320DM642 chip, and realized real-time electronic image stabilization. It tests the communication module, the performance of stabilization and real-time of the instrument, through a test experiment. The result shows that the algorithm presented in this paper can stabilize jitter video effectively and meet the real-time requirements of the image stability when dealing with the 720*576 PAL standard video. It has some value in the practice.
Keywords/Search Tags:Real-time stabilization, DM642, Gray projection, Kalman filter, Full frame compensation
PDF Full Text Request
Related items