基于模糊核均值聚类优化的光流测速方法

Optimized Optical Flow Velocity Measurement Method Based on Fuzzy Kernel Clustering Algorithm

光流算法作为一种载体速度测量方法,易受光照变化、物体相对移动影响,最终导致速度信息获取不准确。为了提高载体运动速度解算精度,提出了一种基于模糊核均值聚类算法优化的金字塔Lucas-Kanade光流测速方法(FKCM-金字塔LK)。该方法首先通过金字塔LK光流算法得到稀疏光流,然后使用模糊核均值聚类算法对稀疏光流数据进行聚类处理,最终建立光流与实验平台运动参数之间的关系,并得到准确的速度信息。实验结果表明,提出的基于模糊核均值聚类算法优化的金字塔LK光流测速方法不仅可以有效地减少物体相对移动对光流测速的影响,而且与现有其他光流测速方法相比,速度测量精度得到显著提高。

The optical flow algorithm,as a method of carrier velocity measurement,is susceptible to changes in illumination and relative motion of objects,resulting in inaccurate acquisition of speed information.In order to improve the accuracy of the carrier velocity solution,this paper presents an optimized pyramidal Lucas-Kanade optical flow velocimetry method(FKCM-Pyramidal LK)based on fuzzy kernel clustering algorithm.The method first obtains a sparse optical flow by means of Pyramidal LK optical flow algorithm,and then uses FKCM to cluster the sparse optical flow data,and finally establishes the relationship between the optical flow and the motion parameters of the experimental platform and obtains accurate optical flow information.The experimental results show that the proposed optimized optical flow velocity measurement method can not only effectively reduce the influence of the relative motion of the object on optical flow velocity measurement,but also significantly improve the accuracy of the measurement compared to other existing optical velocity measurement methods.