Ricolfe-Viala Carlos, Esparza Alicia
Instituto de Automática e Informática Industrial, Universitat Politècnica de València, Camino de Vera s/n, 46022 Valencia, Spain.
Department of Systems Engineering and Automatic Control, Universitat Politècnica de València, Camino de Vera s/n, 46022 Valencia, Spain.
Sensors (Basel). 2020 Jul 1;20(13):3695. doi: 10.3390/s20133695.
Accurate correction of high distorted images is a very complex problem. Several lens distortion models exist that are adjusted using different techniques. Usually, regardless of the chosen model, a unique distortion model is adjusted to undistort images and the camera-calibration template distance is not considered. Several authors have presented the depth dependency of lens distortion but none of them have treated it with highly distorted images. This paper presents an analysis of the distortion depth dependency in strongly distorted images. The division model that is able to represent high distortion with only one parameter is modified to represent a depth-dependent high distortion lens model. The proposed calibration method obtains more accurate results when compared to existing calibration methods.
精确校正高度失真的图像是一个非常复杂的问题。存在几种使用不同技术进行调整的镜头失真模型。通常，无论选择哪种模型，都会调整一个唯一的失真模型来对图像进行去失真处理，而不考虑相机校准模板距离。几位作者已经提出了镜头失真的深度依赖性，但他们中没有一个人处理过高度失真的图像。本文对严重失真图像中的失真深度依赖性进行了分析。能够仅用一个参数表示高失真的划分模型被修改为表示深度依赖的高失真镜头模型。与现有的校准方法相比，所提出的校准方法获得了更准确的结果。
