谷歌Tensorflow框架在自动驾驶汽车车辆识别功能中的应用

Application of Google Tensorflow Framework in Automatic Driving Vehicle Identification Function

文章研究了基于谷歌深度学习框架Tensorflow的图像识别在智能驾驶汽车领域里应用的可能性。使用卷积神经网络模型,将车辆图片作为训练集输入神经网络,通过多次训练校正神经网络参数,最终可以得到可以识别汽车图片的神经网络模型。通过学习Tensorflow建模,编程的完整流程,为进一步使用Tensorflow构建图像识别应用打下了基础。

This paper make the research on the Google deep learning Framework, Tensorflow, and study the image recognition utilization possibility in the intelligent drive area. The model used in this paper is based on Convolutional Neural Network （CNN）, we collect auto images for the image data set, which used as training set. Through multiple training tests we set the parameters of the Convolutional Neural Network and set up one CNN model used for auto image recognition. During the tensorflow model programming process, we set up the whole process for deep learning training, which are useful for the later tensorflow image recognition applications