| In the image processing field, wide-angle lenses are widely used. Such as medical electronic endoscope systems, vehicle monitoring systems and dedicated monitoring systems for banks, classrooms and supermarkets.However, in most of the image processing systems, image distortion will be brought in due to the using of wide-angle lenses, which makes the image difficult to identify. So, in order to restore the original image accurately, the digital image processing techniques for the image distortion correction should be used.The research of corrections for digital image distortion is a hotspot. But, the existing correction algorithms are mostly based on the PC platform. The PC platform leads to limitations of practical application.Based on the research and analysis of the image distortion correction techniques, and appropriate DSP (digital signal processing) processor-Blackfin561, this paper designs a real-time correction system for wide-angle lens video image, which uses the current relatively advanced digital signal processing technology. According to specific system requirements, analysis and improvement of the existing distortion correction algorithms, the system eventually proposes a new distortion correction model (spherical distortion model), and designs one calibration algorithm, that is innovative, real-time, versatile and simple. Eventually, the system is realized.The paper mainly discusses the principle of the whole system design and implementation of some techniques, and the key point of the system contains:1. Analyze the principle of wide-angle lens and image distortion.2. Design models for wide-angle lens image distortion correction algorithm.3. Choose the sensor and DSP platform4. Implement the system on DSP platformExperimental results and testing data have showed that the system presented in this project can accomplish distortion correction of NTSC format video image (30f/s 720*480) and have achieved the expected goal, real-time, accuracy, and reliability. |
PDF Full Text Request