基于深度学习的雾天车标识别

Vehicle Logo Recognition on Fog Weather Based on Deep Learning

车标作为车辆的一个重要信息,在智能交通系统中对于车辆的识别起到一个非常重要的辅助作用,然而雾天摄像机采集到的图片模糊不清,使得雾天的车标识别成为现阶段智能交通系统的一个重要问题。因此,提出一种深度学习与图像去雾相结合的方法,该方法加入图像去雾算法,具有图像增强、降低噪声等优点。实验表明,这种方法正确率较高,在大雾天气情况下准确性和稳定性都较好,很好地解决雾天车标识别的问题。

The logo as an important information of the vehicle plays an important auxiliary role in recognition of vehicle in Intelligent Traffic System (ITS), however the images captured by camera in fog weather are blurred, which makes the recognition of vehicle in fog weather became an important issue in current Intelligent Traffic System. Therefore, proposes a method combined deep learning with image defog, which adds image defog algorithm, and has the advantage of image enhancement and noise reduction and so on. Experimental results demonstrate that the proposed method has a high accuracy, and even in the condition of dense fog, it still has a better accuracy and stability, and it's also a good solution of the problem of vehicle logo recognition in fog weather.