无人驾驶车辆视觉导航中几种典型道路场景的识别方法研究
毕业论文关键词 道路场景识别 颜色直方图 BP神经网络 样本 学习 检测与识别
Title Study on Recognition Method of Unmanned Vehicle Visual Navigation in Several Typical Road Scene
Abstract Accurate, real-time road scene recognition is a key technology for intelligent vehicle plays an important role in the unmanned autonomous vehicle navigation. In view of the persity and complexity of the external environment of the road, real-time and reliable identification is the focus of current research in the field of road scene recognition of the road scene. In this paper, four types of road scene classification method based on BP neural network model, include: asphalt, cement road, gravel roads and dirt roads. Color histogram feature extracting road region by the front of the vehicle, and the use of BP neural network for training and classification, finally to evaluate the recognition effect through the classification boundaries. The experiments show that the algorithm has higher recognition rate, good real-time ability in running speed.
Keywords Road scene recognition Color histogram BP neural network Sample Study Detection and recognition
目次
1 引言 .. 1
1.1 课题研究的背景及意义 1
1.1.1 研究背景 .. 1
1.1.1 研究目的 .. 1
1.3 课题的研究内容 2
1.4 论文的结构和安排 . 3
2 道路特征提取 4
2.1 典型的颜色空间 4
2.1.1 RGB颜色空间 .. 4
2.1.2 YUV和 YIQ颜色空间 .. 5
2.1.3 HSV颜色空间 .. 6
2.2 颜色直方图 . 7
2.2.1 全局直方图 7
2.2.2 局部直方图 8
2.3 直方图的优点和构造 . 11
3.4 道路颜色直方图提取 . 11
3 基于BP神经网络的道路场景识别 14
3.1 神经网络 .. 14
3.1.1 人工神经网络的研究内容 .. 14
3.1.2 人工神经网络的特点 15
3.1.3 人工神经网络的基本形态 .. 15
3.2 BP神经网络 16
3.2.1 BP网络的基本结构 .. 16
3.2.2 BP算法的基本过程 .. 17
3.3 使用BP网络进行道路类型的识别 .. 19
2.4 识别结果 .. 21
4 程序设计 24
4.1 程序开发平台 . 24
4.2 程序功能介绍 . 25
5 结论与展望 . 27
5.1 工作总结 .. 27
5.2 下一步工作展望 .. 28