低分辨率模糊车辆的人工智能识别研究

Research on Artificial Intelligence Recognition of Low Resolution Blurred Vehicles

伴随着我国经济的飞速迅猛发展,国内各种汽车的数量增长飞快,道路交通治理将面临严峻的挑战。采用人工智能应对交通拥挤等各种状况是当前研究的热点。根据人工智能中经典的卷积神经网络,对远距离的低分辨率模糊车辆进行训练学习,提取图像的特点,得到图像分层的特征值。进一步采用区域卷积神经网络,从待识别的图片中定位识别出来车辆,实验结果表明识别准确率可以达到99%以上。该识别技术给智能交通系统提供了便捷的方案。

With the rapid development of economy and the rapid growth of domestic automobiles, road traffic management will face severe challenges. Using artificial intelligence to deal with traffic congestion and other conditions is the current research hotspot. According to the classical convolutional neural network in artificial intelligence, the long-distance low-resolution blurred vehicles are trained and learned to extract the characteristics of the image. The feature of the image hierarchy is obtained.Furthermore, the area convolution neural network is used to locate and recognize the vehicles. The experimental results show that the recognition accuracy can reach more than 99%. This recognition technology provides a convenient scheme for Intelligent Transport System.