基于模拟视频信号差分和射影畸变校正的图像系统

视觉传感器越来越多地应用到嵌入式系统中，通常对视频信号的处理是先对模拟信号数字化，然后再利用嵌入式处理器进行计算。大量的数据和处理器处理能力的差异尤为显著，以至于限制了视觉传感器的广泛使用。本文提出一种基于硬件的模拟差分处理以及射影畸变校正的图像系统，通过像素模拟信号的差分实现边缘检测，同时通过查表校正在摄像头的像平面与地面不平行时产生的图像射影畸变，便于后续的图像处理和模式识别。该方法提高了信号处理的实时性，减轻了视觉传感器对处理能力的依赖。

Image Processing System Based on Analog Video Signal Differential and Projective Distortion Rectification

The video camera is widely applied to embedded systems .Before further processed by the embedded systems , the ana-log video signal is usually digitized first .The large amount of data has made the difference of the processing capability much sig-nificant, and that limits the widespread use of vision sensor .An analog video signal processing system with differential and image rectification is proposed , in which the differential between pixel in analog signal is used to detect the edge .Meanwhile, the pro-jective distortion is corrected according to lookup table .It is easier for further processing after the rectification of projective distor-tion.The solution improves the real time of signal processing and reduces the dependence on processing capability .