论文标题:路面病害自动检测系统的体系结构改进及检测算法研究
Research on Some Key Techniques of Passive Radar Based on FM-broadcast Signals
论文作者 滕井玉
论文导师 唐振民,论文学位 硕士,论文专业 计算机应用技术
论文单位 南京理工大学,点击次数 90,论文页数 70页File Size6694k
2004-06-01论文网 http://www.lw23.com/lunwen_93506947/ 体系结构改进,分级处理模型,路面病害自动检测,图像分割,边缘检测
hierarchy improvement, hierarchical processing model ,road surface disease automatic inspection, image segmentation , edge detection
本文的工程背景是南京理工大学计算机系模式识别与人工智能实验室的N—1型道路状况智能检测车项目，目的是对系统的体系结构进行改进，对路面病害自动检测算法进行研究。 高速公路路面自动检测系统是一个复杂的系统。本文首先介绍了系统的软硬件环境、总体结构、系统各模块及其功能，在此基础上对系统的体系结构进行改进，给出了路面病害定位更有效更合理的解决方案。接下来分析了路面图像的特点，并对原始路面图像进行了必要的预处理，进行灰度校正和几何校正。 本文将检测重点放在裂缝病害的检测上，回顾了传统的边缘检测算法，对各种算子进行评价，分析了传统的检测方法在路面病害检测中的局限性，然后给出了基于分级处理模型的路面病害自动检测方法。分级处理模型灵活高效并能有效地适应复杂的路面图像情况，克服了传统的边缘检测方法的不足。通过实验，证明了基于此模型的检测方法对路面裂缝病害检测的有效性。
This paper is written on the background of the project of N-1 Style Intelligent Road Surface Disease Inspection Vehicle of Department of Computer Science of Nanjing University of Science and Technology. It aims to improve the system architecture and to study the automatic inspection algorithm of diseases in road surface images.The automatic inspection system of highway road surface is a complicated system. Firstly, the software and hardware environment, the system architecture, modules of the system and their functions are introduced. On this basis the system architecture is improved by providing a more effective and rational locating solution of road surface diseases. Then by analyzing the characteristics of the road surface images, we carry out some necessary preprocesses such as gray scale calibration and geometry calibration to these images.This paper is focused on crack disease detection. After reviewing and evaluating the traditional edge detection methods and pointing out their limitation in the detection of the road surface diseases, a new method based on hierarchical processing model is implemented. The hierarchical processing model is flexible and efficient and can meet the complicated situation of the road surface. The validity of the model is proved through experiments.
|
