大数据挖掘技术的光流场图像匹配方法设计

Design of image matching method of optical flow field based on big data mining technology

光流场图像匹配可以提高图像的质量,便于后继的光流场图像处理,而当前光流场图像匹配方法存在一些不足,如匹配耗时间长,错误率较高等,为了获得更优的光流场图像匹配结果,提出了大数据挖掘技术的光流场图像匹配方法。首先采集大量的光流场图像,并对图像预处理,消除干扰因素,从中提取边缘信息,幅值特征,角度特征等特征向量,并引入大数据挖掘技术根据对提取到的特征向量进行光流场图像匹配,最后对比其它图像匹配方法。实验结果表明,大数据挖掘技术的光流场图像匹配方法用时较短,提高光流场图像匹配效率,而且光流场图像匹配精度超过95%,大幅度减少了光流场图像匹配错误率,错误率仅为3%,具有广泛的应用范围。

Image matching of optical flow field can improve the image quality and facilitate the subsequent image processing of optical flow field. However,there are some shortcomings in the current image matching methods of the optical flow field,such as long matching time and high error rate. In order to obtain better image matching results of the optical flow field,an image matching method of optical flow field based on big data mining technology is proposed.Firstly,many optical flow field images are collected,and image preprocessing is carried out to eliminate interference factors,and feature vectors such as edge information,amplitude feature and angle feature are extracted from them. Big data mining technology is then introduced to match optical flow field images according to the multi-extracted feature vectors. Finally,other image matching methods are compared. The experimental results show that the optical flow field image matching method of big data mining technology in this paper takes less time,improves the efficiency of optical flow field image matching,and the accuracy of optical flow field image matching is more than 95%. It dramatically reduces the error rate of optical flow field image matching. The error rate is only 3% and has a wide range of applications.