【作者】 王会峰；

【导师】 刘上乾；

【作者基本信息】 西安电子科技大学 ， 物理电子学， 2009， 博士

【摘要】 高精度的特殊管道在兵器、航天、核能等诸多领域内有着广泛应用。其静态参数的检测是特殊管道设计、研究和使用中一个必需且重要的技术环节。随着特殊管道研制、生产水平和使用要求的不断提高,原来的基于光学望远系统的人工测量、人工处理数据的方法已经不能满足需要,因此研究开发高精度、高效率和自动化的特殊管道静态参数测量系统迫在眉睫。作者在对特殊管道静态参数的测量原理、方法和检测装备的性能及其局限性进行深入分析的基础上,研究并设计了一种高精度的光机电一体化系统。论文以研究过程中遇到的特殊管道复杂环境的高精度光电成像测量为切入点,对测量过程中的图像获取、图像处理、图像测量以及测量数据的处理等多个关键技术进行了深入研究,提出了相应的技术方案,实践结果验证了其合理性和有效性。首先,鉴于成像环境特殊又复杂,而获取图像像质要求较高,故必须借助人工照明系统,因此提出了相应的场景照明设计理念,并且设计了适合特殊环境图像获取的分布式LED光电成像照明光源,实践证明该光源能够保证在特殊复杂环境下获取较高质量测量需求的图像。其次,LED照明光源在一定程度上使成像质量得到了提高,但是仍存在较多的噪声,还不能满足高精度测量要求,为此作者设计了基于小波和中值滤波相结合的自适应图像混合去噪方法和基于人眼视觉特性和模糊算子的图像增强算法,提出了基于MRF规整化非均匀照明背景的抑制算法,并进行了试验仿真,结果表明按上述方法预处理后图像的信噪比显著提高,适合于对靶标图像的预处理。同时由于传统的图像分割方法很难满足特殊管道高精度图像测量的要求,本文又分别提出了基于蚁群算法和最大熵相结合的多阈值图像分割算法、基于局部直径Hough变换的测量靶面同心圆自适应检测方法。完成了相应的算法仿真试验,结果表明图像经过上述方法处理后能满足测量要求。再次,为了满足特殊管道静态参数高精度测量的需求,本文又分别提出了基于高斯曲线拟合的光斑图像亚像素检测方法及基于小波和Zernike矩的标尺靶亚像素检测算法；同时针对像机畸变对高精度测量的影响,提出了基于移动特征靶标的径向畸变标定方法、基于递推最小二乘法像机径向畸变校正方法及基于特征平行直线的几何畸变校正方法；此外在分析了系统误差源的基础上,提出了基于序列图像特征配准的像机旋转误差补偿算法。仿真试验结果显示这些方法能够提高图像测量精度并且减小和消除误差。最后,建立了测量数据处理的部分数学模型,引入了基于三次NURBS拟合的身管空间轴线三维重建算法及基于三次样条插值的药室容积和内表重建算法；此外针对便携式设计提出了基于DSP实时实现的硬件方案。更多还原

【Abstract】 High precision special type tube is widely used in many fields such as weapon, spaceflight and nuclear energy etc. Measurement of the static parameters is a prerequisite and important technique process in designing, studying and use of the special type of tube. With the development of study, manufacture and requirement of application, the classic measurement method based on optics glass and data processing artificially cannot satisfy new requirements of the special type of tube. A kind of high precision, high efficiency, and automatism measurement system need be constructed urgently.Based on thorough study of the measuring principle, measuring method and performance and limitation of the special type of tube, a project to the tube static parameters measurement system with high precision and automatism are designed which integrate the technical of optical, fine mechanism and electron. The critical technical in image obtain, image process, image measurement and data process is studied deeply to the point of accuracy photoelectric imaging measurement in special type tube complex environment during the design. As a result, a series of suitable technologies have been put forward. The feasibility and effectiveness have been verified by practice.Firstly, because an excellent image must be obtain in the special and complex imaging environment, artificial illumination method should be used. In order to get a relatively "good" image, a scene illumination system design idea has been used and a kind of distributed-LED artificial illumination light source which fits the image obtain in special and complex image environment has been designed and practice shows that the designing can ensure obtain high quality image for measure in the special and complex image environment.Secondly, benefit from LED illumination light source, the image quality is improved in some degree but far from the requirement of high precision measurement and include some noise still. So a series algorithms which suit image obtained in the special and complex environment, such as the image denoise means mixed wavelet and self-adaptive median filter, the image enhancement based on human visual property and generalized fuzzy operator have been adopt here,and a novel nonuniform illumination background suppress algorithm based on markov random field model designed. Simulation and test shows that these image preprocess algorithms can improve the SNR greatly and can meet the requirement for the image quality of the later image process. At the same, as the classic image segmentation methods cannot satisfy the need of high accuracy image measurement of special tube, some new methods are put forward to suit the segmentation special image, which include the multi-threshold image segmentation algorithms based on ant optimization and maximum fuzzy entropy, the concentric circles adaptive high accuracy detection algorithms based on local diameter Hough transform. Simulation and test indicate the image process methods can satisfy the need of the subsequent image measure.Thirdly, in order to realize the special type tube static parameters high accuracy measurement, we invent some new algorithms which include the sub-pixel edge detection algorithm for laser spot based on gauss plotfit and sub-pixel detect algorithms for rule target based on wavelet transform and Zernike moments; Simultaneity, in consequence of the lens distortion serious influence the measurement precision, a few lens calibrating and correction methods are construct such as the lens radial distortion calibrating method based on moving feature target, the lens radial distortion correction algorithm based on recursive least squares, the geometrical distortion correction method based on property parallel lines. Moreover, after analyzing the system-error in detail, a kind of camera rotation compensation algorithms based on sequence image frame feature matching. Result shows that the methods can improve the image measurement accuracy and can reduce even eliminate the error.Lastly, the measurement mathematical model are construct partly, the tube space three-dimension axial reconstruct algorithm based on Cubic NURBS plot-fitting and the power chamber volume reconstruction based on three order spline interpolation algorithm are put forward, and these algorithm can get super effect. Meanwhile, aim at the portable design a real-time hard-ware project based on DSP is designed.更多还原