| 基于神经网络的PNN算法在交通标志检测中的应用 |
| |
| 引用本文： | 王江涛,;石红岩.基于神经网络的PNN算法在交通标志检测中的应用[J].贵阳金筑大学学报,2014(4):20-24. |
| |
| 作者姓名： | 王江涛 ;石红岩 |
| |
| 作者单位： | [1]仰恩大学计算机学院,福建泉州362014; [2]仰恩大学数学系,福建泉州362014 |
| |
| 基金项目： | 福建省教育厅科技项目：“彩色图像分割技术在辅助驾驶系统研究中的应用”（项目编号：JA12361） |
| |
| 摘 要： | 交通标志检测技术是实现智能交通系统的关键。查阅诸多文献,文章阐述模式识别的概念,并且给出计算机模式识别的抽象过程。通过对现有交通标志检测框架进行研究,根据交通标志设计之规则,结合概率神经网络,设计了一种基于多层决策树的PNN分类算法模型,并对神经元高斯函数的参数进行改进,最终建立交通标志检测的算法流程。通过实验,对60个交通标志进行晴天、多云以及阴雨三种天气背景下的检测。最后整理实验数据,通过建立图表进行分析和比较,证明该分类器能够实现交通标志检测的功能,达到预期的检测效果。最后,分析了该算法模型还存在的不足之处,也指出了将来研究的方向。
|
| 关 键 词： | 模式识别 边缘检测 PNN 决策树 神经网络 |
The PNN neural network algorithm in the application of traffic sign detection based on |
| |
| Institution: | WANG Jiang-tao ,SHI Hong-yan ( 1. Computer College of Yangn University, Quanzhou Fujian 362014, China ; 2. mathematics department of Yangbn University, Quanzhou Fujian 362014, China) |
| |
| Abstract: | Traffic sign detection technology is the key to realize the intelligent traffic system,access to a lot of literature,the author expounds the concept of pattern recognition,and give the abstract process of computer pattern recognition.Then,through research on the current traffic sign detection framework. According to the traffic sign design rules,combined with probabilistic neural network,the design of a base layer of the PNN decision tree classification algorithm model,and the parameters of neuron Gauss function is improved,the eventual establishment of traffic sign detection algorithm flow. Finally,to detect the 60 traffic signs are sunny,cloudy and rainy weather background three. Finally,the experimental data,comparison and analysis through the establishment of the chart,shows that the classifier can achieve traffic sign detection function,to achieve the desired effect of detection. Finally,the author analyzes the deficiency of the algorithm model still exist,also pointed out the direction of future research. |
| |
| Keywords: | Pattern recognition Edge detection PNN Decision tree Neural network |
| 本文献已被 维普 等数据库收录！ |
|
