炭素制品X射线图像的缺陷特征分析与选择

针对炭素制品X射线检测图像的特点，对缺陷及其特征提取与选择技术进行了研究。分析了炭素制品生产中易产生的缺陷类型及缺陷的成像特征，在此基础上，从缺陷样本中提取了19个特征值。以特征组合分类能力数学模型为适应度函数，设计了基于遗传算法的特征选择策略，实现了对缺陷原始特征量的优化选择。利用BP神经网络分类器及选择的特征值对缺陷进行了模式分类。研究结果表明，提出的选择方法是比较有效的，可以用于缺陷的识别与分类。     

基于视觉传感的脉冲TIG焊熔池一次图像获取的研究

依据弧光光谱理论，采用辅助光源及复合滤光技术，有效地解决了焊接区熔池成像的干扰和清晰度差等问题，研究了滤光片与辅助光源对成像质量的影响和作用。试验结果表明，同普通图像采集方法相比，该系统获得图像的清晰度得到明显的提高，可以用于焊接过程的实时监控。     

基于数学形态学图像处理的TIG焊接过程正面熔宽控制

焊接过程熔池动态特征信息获取与智能控制一直是焊接界研究的前沿课题.针对TIG焊接过程没有飞溅,熔池形状稳定、清晰的特点,以TIG焊接过程为研究对象,采用机器视觉技术获取焊接动态过程正面熔池图像点,将数学形态学中的二值理论应用于熔池图像的处理,结合模糊逻辑和比例积分微分(PID)闭环控制理论,从弧焊过程熔透质量控制的角度出发,从软件开发到硬件组成,设计了一套焊接过程图像处理及正面熔宽控制系统.结果表明,该系统可以很好地克服焊接过程中的外界干扰,保证焊缝成形的稳定,从而实现了TIG焊质量控制的目标.     

Pro/E自动分模方法与技巧

Pro／E在模具制造业已经普遍应用，基于作者几年来对Pro／E软件应用的一些实践与体会，简单介绍Pro／E2001侧面影像曲线与裙边曲面的巧用。     

Difference in electron- and gamma-irradiation effects on output characteristic of color CMOS digital image sensors

Changes of the average brightness and non-uniformity of dark output images, and quality of pictures captured under natural lighting for the color CMOS digital image sensors irradiated at different electron doses have been studied in comparison to those from the γ-irradiated sensors. For the electron-irradiated sensors, the non-uniformity increases obviously and a small bright region on the dark image appears at the dose of 0.4 kGy. The average brightness increases at 0.4 kGy, increases sharply at 0.5 kGy. The picture is very blurry only at 0.6 kGy, showing the sensor undergoes severe performance degradation. Electron radiation damage is much more severe than γ radiation damage for the CMOS image sensors. A possible explanation is presented in this paper,     

Image segmentation of small defects in precision weldments with complex structure

In this paper, the X-ray nondestructive test method of small defects in precision weldments with complex structure was presented. To resolve the difficulty of defect segmentation in variable grey image, the image processing based on Visual Basic programming method was adopted. The methods of automatic contrast and partial grey stretch were used to enhance the X-ray detection image which has relatively low contrast, then automatic threshold method was carried out to segment the two high intensity zones, and weld zones which contain the small defects was extracted. Smoothing and sharpen processing were proceeded on the extracted weld zones, and small defects in X-ray detection image of weldments with complex structure were segmented by using the method of background subtraction in the end. The effects of raster were eliminated, and because of that the image processing was only proceeded on the extracted weld zones, the calculated speed using the above provided algorithm was improved.     

Gaze direction based vehicle teleoperation method with omnidirectional image stabilization and automatic body rotation control

In robot teleoperation, a robot works as a physical agent at a remote site for a robot operator. There are mainly two tasks in robot teleoperation using camera images: environment recognition using visual information and robot control according to the recognition. In this paper, we propose a gaze direction based vehicle teleoperation method with an omnidirectional image stabilization and an automatic body rotation control. In the proposed method, we manage above two tasks in the same manner that are usually treated separately. This method is an intuitive vehicle teleoperation method where an operator do not need to have concern about vehicle body orientations and can absorb differences of vehicle driving mechanisms. That is, this method frees an operator from being bothered from controlling a vehicle and the operator can concentrate on where he/she intends to go. This method mainly consists of two technologies: an omnidirectional image stabilization technology and automatic body rotation control. The conducted experiments show effectiveness of the proposed method.     

一种融合多传感器信息的移动图像识别方法

多传感器数据融合作为一种特殊的数据处理手段在图像识别领域得到了较大的重视和发展, 本文提出了一种融合多传感器信息的移动图像识别方法. 首先通过在智能手机端提取带传感器信息的图像局部特征,增强局部特征的辨别能力; 其次改进了随机聚类森林的建立算法,减少了样本图像训练时间;最后使用快 速几何一致性校验对匹配结果进行检查, 保证算法的识别精度.实验结果表明,本文提出的方法能够快速 有效地识别移动图像,并具有较好的鲁棒性,同时与传统的Vocabulary tree 方法进行比较,本文方法的识别速度和精度较优,训练代价较低.     

数据场典型相关分析及其在图像分割中的应用

针对数据场环境下多维数据的低维特征提取问题,本文将数据之间的相互作用纳入其相关性求解中,提出一种基于数据场的典型相关分析(Data field based canonical correlation analysis, DFCCA)方法. DFCCA提取的特征具有良好的分布特性,原空间上相隔较远的数据点对的特征聚集在一个较小区域内,而相邻数据点对的特征却有规律地分布在其他点所聚集区域的周围.此特性使得DFCCA具有较好的边界辨识能力,将其应用于图像分割的实验结果表明, DFCCA提取的复杂图像边界具有较好的保真度.     

基于轮廓几何稀疏表示的刚性目标模型及其分级检测算法

刚性目标轮廓具有明显几何特性且不易受光照、纹理和颜色等因素影响.结合上述特性和图像稀疏表示原理,提出一种适用于刚性目标的分级检测算法.在基于部件模型(Part-based model, PBM)的框架下,采用匹配追踪算法将目标轮廓自适应地稀疏表示为几何部件的组合,根据部件与目标轮廓的匹配度,构建描述部件空间关系的有序链式结构.利用该链式结构的有序特性逐级缩小待检测范围,以匹配度为权值对各级部件显著图进行加权融合生成目标显著图. PASCAL图像库上的检测结果表明,该检测方法对具有显著轮廓特征的刚性目标有较好的检测结果,检测时耗较现有算法减少约60%~90%.