A novel lip reading algorithm by using localized ACM and HMM: Tested for digit recognition

Lip contour tracking is an integral part of lip reading application. Fast and accurate lip tracking is an important step in lip reading. This paper uses a novel active contour model for lip tracking and proposes geometrical feature extraction approach for lip reading. Effect of individual features are compared and a joint feature model is obtained by combining weighted decision obtained by a feature vector of difference in inner area, height and width of lip. Ergodic hidden markov model (HMM) is used as a classifier. For each digit Markov Model is tested for 3 states and 5 states. Videos of English digit from 0 to 9 have been recorded for recognition test. Cuave database is used for comparison along with an in-house database. While doing computation of feature vectors, only significant frames are used to reduce the computation complexity. Results of experimentations on digit utterances are given to show that the maximum recognized digit can be used for important programming command of computerized numerical control machines.     

高厚度铝基板的工艺加工

高厚度铝基板是一种特殊印制板,其具有良好的尺寸稳定性和导热性,逐渐被应用于军用设备和高科技民品领域,文章对高厚度铝基板的加工工艺进行了探讨,并对生产过程存在的问题进行分析并找出改进措施。     

模糊分段光滑图像分割模型及其快速算法

灰度分布不均图像是图像分割中一个难点,为此提出一种模糊分段光滑(FPS)图像分割模型.借鉴分段光滑Mumford-Shah(MS)模型与模糊聚类思想,新模型通过两个定义在图像域的光滑函数描述区域特征,并利用模糊隶属度函数代替MS模型中的特征函数.同时,边界检测算子的引入能够有效保护图像中的边界信息.数值求解采用分裂Br...     

运载火箭控制系统的BIT设计与实现

根据拓扑性质知觉原理,视觉原语是从整体到局部,拓扑感知优先。根据此仿生原理,建立感知目标物的形状拓扑信息的数学模型。首先提取物体三维轮廓信息;将空间图形轮廓降维映射到二维投影平面中进行处理;建立提取空间图形拓扑特征的数学模型。将轮廓拓扑空间同胚映射到(Hausdorf或度量空间或)二维投影平面,保持了拓扑不变性,因此便于分析出拓扑特征,可继续应用拓扑映射原理,分析点集拓扑。最终,拓扑特征与知识库模型相匹配,识别出物体。     

基于OPENCV的手势识别系统的设计与实现

人与计算机的交互技术是一种新型的计算机技术,且逐渐演变为一种主流技术和计算机领域的技术热点。为了能够更好的识别手势和跟踪手势的运动轨迹,提出了基于OPENCV的手势识别系统,系统引入了OPENCV计算机视觉库,OPENCV作为优秀的计算机视觉库,为设计的实现提供了便捷的代码,利用OPENCV技术中的图像处理算法,首现通过摄像头采集数据图像,并对采集到的图像进行一系列的缩放,去噪以及锐化等处理,然后对人体手势建立肤色模型,然后经过灰度阈值化来转换成二值图像,得到手轮廓的数据图像后,采用轮廓匹配方法识别出手型。最后通过10种基本的手势模型对比验证了本系统具有一定的实时性,并且识别率可以达到95%以上。     

一种消除LED显示屏动态假轮廓的脉宽调制法

针对LED显示屏播放运动画面时的动态假轮廓问题,提出了集中式门控脉宽调制法.该方法在原有门控脉宽调制技术的基础上,将PWM周期中的小数场固定于周期的中间,然后计算整数权重位对应的子场数并将其紧邻小数场实现.分析和实验表明,集中式脉宽调制法具有与线性PWM相似的点亮时间函数,能够将任意两灰度级间DFC指标的均值从GPWM方式下的0.6降到0.2以下.在有效消除DFC的同时又保持了GPWM的高刷新频率与高灰度级,从而明显改善图像的显示质量.同时算法实现简单,不需要额外的硬件支持.     

基于OpenCV的目标物体颜色及轮廓的识别方法

针对机器人视觉中对目标物体的拾取问题,提出一种基于Open CV函数对从摄像头输入的目标物体的颜色及轮廓进行有效识别的方法。首先利用摄像头读取目标物体及周围环境,经一系列图像处理后进行目标物体的颜色识别,在此基础上进行Canny边缘检测并查找轮廓,最后将轮廓显示。结果表明,该实验程序可以很好地完成对目标物体的颜色及轮廓提取,且有效地避免了目标物体周围相近颜色的干扰。     

基于方向梯度直方图的螺纹连接件端面缺陷检测方法

针对螺纹连接件端面表面的反光性、阈值选取的不确定性, 提出了一套图像处理方法. 使用中值滤波和图像锐化增强等方法处理图片, 解决背景与前景差距较小的问题; 选取利用方向梯度直方图确定图像上的边缘方向, 实现对目标的直线轮廓提取, 与已知目标的轮廓模型结合实现对非直线目标轮廓的提取. 该处理方法可以自动提取目标的轮廓, 而不需要设定阈值, 且对亮度、对比度和噪声变化的适应性较好. 对2500个报警器底盘做的检测实验结果表明, 采用该方法能去除噪声, 与Canny轮廓提取法相比, 能更准确且稳定地提取出螺纹连接件端面的边缘特征; 对螺纹连接件端面磨损和歪斜的检测准确率分别达到了99.4%和98.6%, 满足工业的准确率和实时性的需求; 同时可减少人工成本, 提高检测速度.     

Application of Eigenkets of Bosonic Creation Operator in Deriving Some New Formulas of Associated Laguerre Polynomials

Using the resolution of unity composed of bosonic creation operator＇s eigenkets and annihilation operator＇s un-normalized eigenket, which is a new quantum mechanical representation in contour integration form, we derive new contour integration expression of associated Laguerre polynomials L^ρm （｜z｜^2） and its generalized generating function formula. A series of recursive relations regarding to L^ρm （｜z｜^2） are also deduced in the context of the Fock representation by algebraic method.     

A novel area measurement scheme based on a multi-resolution dynamic contour

A novel scheme for an object's surface area measurement is proposed, which is suitable for area computation of an object with smooth and irregular edges. A multi-resolution dynamic contour is applied and this scheme consists of four steps. Firstly, a photoelectric image collimation system is applied to obtain a target image of a detected object. An image pyramid for the target image is constructed by wavelet decomposition and reconstruction. Secondly, in the image pyramid, a multi-resolution dynamic contour converges to the target's contour edge from coarse to fine scale, via feature search and an iteration algorithm. Thirdly, for the convergent dynamic contour, two formulas for the area and centroid computation of a closed B-spline curve are applied to compute the image target's area and centroid exactly. Finally, a novel centroid self-calibration technology is applied, which measures the pixel's size equivalence with the computed centroid and a dual-frequency laser to measure the true object's surface area exactly. Experiments indicate that this scheme's single-measurement error decreases to ±0.2%, when the number of control points is 20. Compared to conventional measurement approaches, this scheme is robust for a target image with noises or complicated edges.