基于成像畸变模型的鱼眼图像直接处理算法研究

项目名称： 基于成像畸变模型的鱼眼图像直接处理算法研究

项目编号： No.61201435

项目类型： 青年科学基金项目

立项/批准年度： 2013

项目学科： 电子学与信息系统

项目作者： 吴健辉

作者单位： 湖南理工学院

项目金额： 24万元

中文摘要： 本项目以鱼眼成像畸变模型为基础开展鱼眼图像直接处理算法研究，实现对鱼眼图像直接预处理以及鱼眼图像中目标检测与识别功能，并建立鱼眼图像直接处理算法包，推动鱼眼镜头的实用化进程。本项目的具体内容包括：①以鱼眼镜头成像原理为基础，建立鱼眼图像成像畸变模型，解决鱼眼图像中目标在图像任意位置的检测稳定性，以及目标的形状和形态畸变问题；②以畸变模型为基础，建立鱼眼图像直接处理算法模型，实现鱼眼图像信息无损和高效处理；③以鱼眼图像直接处理模型为基础，提出或改进适合鱼眼图像的直接预处理专用算法，研究鱼眼图像中目标检测、跟踪和识别算法，并输出目标轨迹和相关属性，实现对鱼眼图像的高效、可靠处理；④对各个算法首先进行Matlab仿真，然后进行实验平台验证，最后根据仿真和实验的结果对算法进行优化，提高其稳定性和可靠性；⑤对优化设计后的算法进行打包，形成鱼眼图像直接处理算法软件包，方便移植和进一步应用。

中文关键词： 鱼眼图像；畸变模型；直接处理；目标检测；目标识别

英文摘要： The fisheye image is imaging by fisheye lens which has super wide-angle and the field of view generally at 180 degree to 230 degree. The circular fisheye image has ultra big view field which can imaging for a hemisphere space. However, the fisheye image has a big distortion so that it is not use widely in application. This project will be studied the directness processing model to realize the fish-eye image processing which arrived the directness preprocessing of distortion fish-eye image and the function of object detection and recognition.At last, the directness processing algorithms package of fish-eye image will be established to use in the application system. The studied content of this project includes: (1) The fisheye image distortion model will be established based on the fish-eye lens imaging principle and the robustness will be solved of the object detection and extraction in the different locations of fisheye image. (2) The direct processing algorithm model of fish-eye image will be established based on the distortion model to realeze the high efficiency and information lossless processing. (3) The algorithm of object detection will be studied which based on the fisheye distortion model to extract any object in the monitoring area. The segmentation algorithm of the object will be studied. The feature

英文关键词： fisheye image；distortion model；direct processing；object detection；object recognition