CAJ | 학술논문

针对圆形鱼眼图像成像畸变大的特点，在现有数字图像处理算法的基础上，引入自适应技术开展了对鱼眼图像进行直接处理的：相关关键算法研究．自适应技术通过对采集的图像画面进行分析，自动获取最佳的处理参数，从而针对不同的图像获得最佳的处理效果．论文首先从理论上对鱼眼图像直接处理的几个关键算法即模糊自适应中值滤波算法、自适应动态更新背景差运动目标检测算法和二维改进Otsu自适应鱼眼图像分割算法进行了详细的研究，然后对各个算法进行了实验仿真．实验结果表明，基于自适应技术的鱼眼图像直接处理算法能比较理想的对鱼眼图像进行去噪、运动目标提取以及目标分割，相对于传统的先校正后处理方法，具有运行速度快，信息无损的特点，在鱼眼超大范围智能视频监控中有着重要的实际意义．
침대원형어안도상성상기변대적특점，재현유수자도상처리산법적기출상，인입자괄응기술개전료대어안도상진행직접처리적：상관관건산법연구．자괄응기술통과대채집적도상화면진행분석，자동획취최가적처리삼수，종이침대불동적도상획득최가적처리효과．논문수선종이론상대어안도상직접처리적궤개관건산법즉모호자괄응중치려파산법、자괄응동태경신배경차운동목표검측산법화이유개진Otsu자괄응어안도상분할산법진행료상세적연구，연후대각개산법진행료실험방진．실험결과표명，기우자괄응기술적어안도상직접처리산법능비교이상적대어안도상진행거조、운동목표제취이급목표분할，상대우전통적선교정후처리방법，구유운행속도쾌，신식무손적특점，재어안초대범위지능시빈감공중유착중요적실제의의．
The circular fish-eye image is imaging by fish-eye lens and it has higher distortion because of its more than 180 degree field of view, so that it has bad processing effect if used some normal digital processing methods. In this paper, the auto-adapt technology was applied in the fish-eye image directness processing algorithm. Through analysis the image with automation, the best parameter would be obtained and the image would have the best processing effect to every image. The theory of key algorithm has been detailed researched and simulated for fish-eye image directness processing like Fuzzy Auto-adapted Median Filter （FAMF）, Auto-adapted Renewal Background Subtraction （ARBS）, and improved 2D Otsu algorithm for auto-adapted threshold image segmentation. The simulated result of these algorithms indicated that the directness processing algorithm for fish-eye image has higher processing speed and lossless information than traditional method of emend the fish-eye image to normal plane image. The processing effect is achieved the applied request practically and could be used in the super range video intelligent surveillance system.