地下管线位置分布的图像检测技术研究

【作者】 童飞；

【导师】 汪贵华；

【作者基本信息】 南京理工大学 ， 物理电子学， 2014， 硕士

【摘要】 随着城市规划建设步伐的加快,地下管线的建设和维护正扮演着不可或缺的角色。施工人员在对管道维护施工之前,必须充分弄清施工地点周围的地下管线的铺设情况、地理位置和埋设深度。目前,工程应用中主要使用管线探测仪、探地雷达等设备获取管线位置分布的相关参数并绘制管线分布图,但这些方法在实际测量中存在抗干扰能力弱、工作量大、定位精度不高等问题。本文根据瑞典Reflex Maxibor钻孔测斜仪的光学测量原理,提出一种将计算机视觉检测技术和ARM嵌入式处理技术相结合的管线位置检测新方法,该方法通过计算CMOS图像传感器前方的目标圆环的三维位置信息来实现地下管线的位置检测,装置在管内自动检测,有效地降低了工作量和检测成本。地下管线位置分布检测系统主要包括图像采集子系统、摄像机标定子系统、目标检测与定位子系统、数据处理子系统。图像采集子系统以OV9650摄像头作为图像采集设备,设计和编写了OV9650设备驱动程序,并利用Linux中提供的中断处理、信号量、等待队列等机制解决了设备的并发控制问题,用户应用程序完成了目标圆环图像数据的采集。由于摄像头一般存在透镜畸变效应,本文在摄像机标定子系统中分析了摄像机的成像模型并采用张氏标定法对摄像机进行标定,利用计算得出的内参数和畸变参数对采集到的图像进行校正,提高了系统的测量精度。对校正后的图像利用图像滤波、阈值分割、高光消除、轮廓提取等图像处理算法计算出目标圆环的位置信息,数据处理子系统中可以利用这些位置信息重建出管线的形状与方位。更多还原

【Abstract】 As the step of urban planning and construction speeds up, construction and maintenance of underground pipelines is playing an indispensable role. Before the construction of the pipeline, constructors must fully ascertain the laying, position and burial depth of the underground pipelines around the construction site. Currently, in engineering applications, constructors mainly use the pipeline detectors, ground penetrating radar and other equipment to get the relevant parameters of the pipeline’s positional distribution and draw pipeline distribution maps. However, in actual measurement, these methods have some problems such as weak anti-interference ability, workload and low positioning accuracy. This paper presents a new method which combines computer vision detection technology with ARM embedded processing technology to detect the position of the pipeline based on the optical measurement principle of the Sweden Reflex Maxibor borehole inclinometer. This method achieves the position detection of the underground pipeline by calculating the three-dimensional position of the ring which is in the front of the CMOS image sensor. It can achieve automatic detection which reduces the workload and testing costs.The underground pipeline position detection system mainly includes image acquisition subsystem, camera calibration subsystem, target detection and position subsystem and data processing subsystem. Image acquisition subsystem uses OV9650camera as the capture device, this paper designs and writes the OV9650device drivers and uses interrupt, semaphores and wait queue mechanism which Linux provides to solve the problem of the concurrency control of equipment. User application completes the image data acquisition of the circle. Generally, the camera has lens distortion effect, so this paper analyzes the image model of the camera and uses Zhang calibration method to calibrate the OV9650camera in camera calibration subsystem. Then, this paper uses the intrinsic parameters and distortion parameters to correct these images. This operation can improve the measurement accuracy of our system. After correction these images, this paper uses image filter, threshold segmentation, high light elimination, contour extraction and other image processing algorithms to calculate the position information of the target circle. Data processing subsystem can use the position information to reconstruct the shape and orientation of the pipeline.更多还原