基于游程与轮廓跟踪的连通区域标记改进算法

为了进一步提高连通区域标记的效率,提出了一种改进的基于游程与轮廓跟踪技术相结合的连通区域标记算法。该算法首先判断当前前景像素点是孤立点还是轮廓点,然后判断是外轮廓还是内轮廓,并用相应的轮廓标记算法对其进行标记,最后用游程标记方法对本行中剩余的前景像素进行标记,整个过程无需等价表,耗时较少。实验证明,改进后算法的标记效率提高大约24.5%。     

边缘检测技术在服饰图案切割中的应用研究

对边缘检测在服饰图案切割中的应用开展研究.在Visual C+ +6.0环境下编制了服饰图案边缘检测图像数字处理软件,实现了灰度均衡、Laplace边缘检测算子和图像腐蚀等多种算法,对各种算子及其组合的适用性做了深入比较.实验成功地提取了复杂服饰图案的边缘轮廓,并驱动激光裁床对服饰图案实物进行边缘切割,实现了从图案的信息采集到切割的即时处理,轮廓的寻边精度高.     

图像LSB隐藏游程检测算法

当前,基于图像LSB的隐藏算法已经成为一种主流的信息隐藏技术。根据LSB隐藏将导致图像空域像素区内LSB长游程长度偏移(减小)的特点,本文提出了能够度量这种偏移的游程检测算法,并给出了算法原理和实现方法。该算法基于盲检测、不需要训练图像数据库、在满足一定假设条件下,检测性能优良而且算法简单易于工程实现,具有一定的实用价值。     

基于Hough变换和无轨迹卡尔曼滤波的眼睛角点跟踪

提出了一种基于改进Hough变换(HT)和无轨迹卡尔曼滤波(UKF)的眼睛外角点跟踪算法。该算法在输入图像中存在虹膜时采用改进Hough变换提取眼睑轮廓并得到眼睛外角点位置,当输入图像中检测不到虹膜时,采用UKF算法对当前帧眼睛角点进行估计。实验证明,本文算法能精确地跟踪眼睛外角点。     

引入Harris角点检测的运动图像边缘轮廓提取

对运动目标图像进行边缘轮廓特征提取,提高运动图像的电子稳像性能,从而准确实现运动目标的检测和识别.提出了一种基于Harris角点检测的运动图像边缘轮廓提取算法,对运动图像进行仿射不变闭合区域增强处理,采用Kalman滤波和运动状态相邻帧补偿的方法进行运动图像Harris角点检测,提高相邻两帧之间边缘轮廓的角点检测能力,实现了运动图像边缘轮廓的提取和稳像处理.仿真结果表明,用该方法进行图像边缘轮廓角点提取的准确度高且稳健性好.     

一种双游程编码的测试数据压缩方案

SOC芯片测试中一个主要的挑战就是处理大量的测试数据.为了减少芯片测试中的测试数据,提出了一种双游程的编码方案,采用变长到变长的编码方式对0游程和1游程进行编码.该算法在编码时同时考虑0游程和1游程,大大减少了测试数据中短游程的数量,同时文中给出了一种基于有限状态机的解压缩算法的实现方案.理论分析和实验结果证明该方案具有高压缩率、硬件实现简单等特点.     

基于统计模型和活动轮廓的运动目标检测与跟踪

提出了一种静止背景视频序列中运动目标的检测与跟踪方法.对连续两帧图像序列作差分计算,对差分图像的灰度分布建立混和高斯模型(GMM),采用期望最大化(EM)算法估计模型参数,并引入基于GMM模型的边界检测算子,进而构造运动边界图像.改进静态图像轮廓提取算法GVF-Snake,利用运动边界图像修改GVF-Snake的能量项,使其能够提取视频序列中运动目标的轮廓.采用一种根据目标区域自动初始化轮廓的方法解决Snake初始轮廓需要手工设定的问题,采用一阶差分预测算法加快轮廓收敛速度.利用改进的GVF-Snake算法对运动目标进行检测与跟踪,结果表明,该算法对刚性和非刚性两类目标都具有较好的检测与跟踪效果.     

基于融合图像特征的角点提取方法

基于图像边缘的角点提取往往对噪音敏感,提取精度较高但运算量大,而基于图像灰度的角点提取易于实现但提取效果不佳.因此提出一种融合图像边缘特征和图像灰度特征的角点检测方法.首先在一较低尺度用Canny算法求出所有边缘点,然后求出每一边缘点的曲率值并求出初始角点集,利用Harris算法通过实验在一较优尺度下对初始角点进行筛选并确定最终的角点集合.所提方法融合图像角点提取的两大特征,可以有效改进在单一特征提取下的不足.通过对比实验,该算法明显地提高了图像角点检测性能.     

基于HSI空间的彩色绣花图样的轮廓提取算法研究

电脑绣花需要彩色图像的轮廓提取与针法填充,而绣花机的针法填充可按区域色块进行.结合绣花机编针的特点,根据彩色图像形态学理论,提出一种基于HSI空间的彩色绣花图像轮廓提取算法.实验结果表明该算法能有效抑制噪声对图像的影响,图像提取的效果满足绣花图案的要求.     

Hough变换在眼睛特征提取中的应用

为了准确地提取虹膜和眼睑轮廓等眼睛特征,提出了一种基于改进Hough变换的算法．该算法采用满足梯度要求的点对及其梯度信息确定输入图像中存在的圆,对于梯度反方向延长线相交且与交点距离相等的两点,定义它们属于同一个圆．在提取眼睑轮廓时,将候选点分为梯度方向向上和向下两部分并构造新的累加器分别确定上下眼睑所对应的圆,避免了上下眼睑的相互干扰．实验结果表明,算法在眼睛非闭合状态和正常光照下能准确地提取眼睛特征．与大多数Hough变换利用3个特征点来确定圆相比,算法通过引入梯度信息的方法,计算复杂度从O(N3)降低为O(N2)．     

Image denoising and boundary extraction based on game theory

Half-quadratic regularization is a classical image denoising method.In removing image noise,the image boundary can be obtained.Since the boundary obtained by the half-quadratic regularization model is too fuzzy and the denoising effect is not ideal,the half-quadratic regularization model is improved by the game method,the image is denoised and the boundary is extracted simultaneously.Two participants are defined,with the classical half-quadratic regularization method used as the target function of denoising,and a relatively novel global sparse gradient model selected as the target function of boundary extraction.The two participants,image denoising and boundary extraction,iterate alternately in a game process,with their convergence points as the Nash equilibrium points.The proposed model is applied to various types of images,and the algorithm proposed can lead to good results in both numerical results and visual effects.Experimental results show that the proposed algorithm can effectively improve the half-quadratic regularization model,thus obtaining better denoising and boundary extraction effects.     

RBF神经网络用于工件边缘特征参数的提取

提出一种基于径向基函数神经网络（RBFNN）的工作特征参数提取的方法,采用了两个径向基函数神经网络,利用第一个RBFNN分别求出工件的边缘点pi邻域内的顺时针边缘与逆时针边缘与x轴的夹角,两边缘夹角小的边缘点pi被认为是具有高曲率的角顶点。根据工件边缘曲线的特征,建立了各种边缘的曲率符号模型,用该模型训练第二个RBFNN,从而识别具有低曲率的切点和拐点及边缘曲线的类型。采用神经网络的方法提取工件特征参数,能准确地定位特征点。     

基于直方图峰值的PCB三维CT图像双直方图均衡算法 

为了改善PCB三维CT图像的对比度和动态范围,针对目前的双直方图均衡算法无法适应PCB图像特性的问题,提出一种基于直方图峰值的双直方图均衡算法。利用PCB三维CT图像的灰度直方图中的最大值对应于PCB基板的概率最大的先验信息,提出以直方图峰值所对应的灰度值为二分点对直方图进行二分,在其两侧分别进行直方图均衡增强,最后再合并增强结果。实验结果表明算法对PCB三维CT图像更具有针对性,增强效果更好。     

Improved Region Merging Algorithm for Remote Sensing Images

To segment high-resolution remote sensing images (RSIs) accurately on an object level and meet the precise boundary dividing requirement, an improved superpixel segmentation and region merging algorithm is proposed. Simple linear iterative clustering (SLIC) is widely used because of its advantages in performance and effect; however, it causes over-segmentation, which is very disadvantageous to information extraction. In this proposed method, SLIC is firstly adopted for initial superpixel partition. The second stage follows the iterative merging procedure, which uses a hierarchical clustering algorithm and introduces a local binary pattern (LBP) texture feature operator during the process of merging. The experimental results indicate that the proposed method achieved a good segmentation and region merging performance, and worked effectively on cloud detection preprocessing in high-resolution RSIs with cloud and snow overlap situations.     

Flotation bubble seed image filling algorithm based on boundary point features

Bubble seed image filling is an important prerequisite for the image segmentation of flotation bubble that can be used to improve flotation automatic control. These common image filling algorithms in dealing with complex bubble image exists under-filling and over-filling problems. A new filling algorithm based on boundary point feature and scan lines (PFSL) is proposed in the paper. The filling algorithm describes these boundary points of image objects by means of chain codes. The features of each boundary point, including convex points, concave points, left points and right points, are defined by the point’s entrancing chain code and leaving chain code. The algorithm firstly finds out all double-matched boundary points based on the features of boundary points, and fill image objects by these double-matched boundary points on scan lines. Experimental results of bubble seed image filling show that under-filling and over-filling problem can be eliminated by the proposed algorithm.     

结合边缘信息优化LSC和改进A*算法的正射影像图镶嵌线提取

为避免正射影像图在镶嵌时镶嵌线穿过视觉显著地物,损害影像中地物的完整性,本文提出一种结合边缘信息优化LSC(线性谱聚类)和改进A^*算法的正射影像图镶嵌线提取方法.首先,将经典LSC的超像素分割理论引入正射影像图镶嵌线提取算法中,并提出边缘强度因子用以优化经典线性谱聚类,有效利用正射影像图中的光谱信息和边缘信息;然后,将优化后的LSC分别应用于两幅正射影像的重叠数据集,得到各类地物的边界特征图,并通过数学形态学法去除边界特征图中的边缘不整齐现象和孤立噪点;最后,改进传统A^*算法,由曼哈顿距离函数替代原有基于欧式距离测度的启发函数,提升A^*算法进行最短路径搜索的效率,在边界特征图中快速搜索出最短路径得到最优镶嵌线.利用真实无人机航拍正射影像图将本文方法与相关方法进行对比分析,实验结果表明本文所提方法可高效、高质量地提取到影像镶嵌线,有效绕过视觉显著地物,减少镶嵌线穿过地物的像元点数,满足实际正射影像图制作的应用需求.     

平方根UKF神经网络及其在预测中的应用

提出一种利用平方根无轨迹卡尔曼滤波(SR-UKF)进行神经网络权值估计的算法,该算法可以克服BP算法存在的学习速率缓慢、计算量大、容易使学习陷入局部极小等缺点.以Mackey-Grass混沌时间序列作为神经网络输入,运用SR-UKF算法、UKF算法、BP算法仿真神经网络.结果表明,SR-UKF算法较之BP算法具有更快的训练速度和更高的预测精度,且可以避免网络学习陷入局部极小,而相对于UKF算法又具有计算量小且能保证状态方差半正定的特点.     

Extraction of line and rounded objects from underwater images

The extraction of geometric primitives from an im-age is one of the basic tasks of computer vision.Straight lines and ellipses are common components inmost underwater images(such as mines and torpedoimages)which are proposed by computer recognitionsystems.The Hough Transform(HT)is the mostpopu-lar technique for extracting geometric shapes.Explana-tions of HT can be found in most pattern recognitionbooks.In brief,the principle concept of the HT is todefine a mapping between an image space a…     

Improved SLIC Segmentation Algorithm for Artificial Structure Images

Simple linear iterative cluster(SLIC) is widely used because controllable superpixel number, accurate edge covering, symmetrical production and fast speed of calculation. The main problem of the SLIC algorithm is its under-segmentation when applied to segment artificial structure images with unobvious boundaries and narrow regions. Therefore, an improved clustering segmentation algorithm to correct the segmentation results of SLIC is presented in this paper. The allocation of pixels is not only related to its own characteristic, but also to those of its surrounding pixels.Hence, it is appropriate to improve the standard SLIC through the pixels by focusing on boundaries. An improved SLIC method adheres better to the boundaries in the image is proposed, by using the first and second order difference operators as magnified factors. Experimental results demonstrate that the proposed method achieves an excellent boundary adherence for artificial structure images. The application of the proposed method is extended to images with an unobvious boundary in the Berkeley Segmentation Dataset BSDS500. In comparison with SLIC, the boundary adherence is increased obviously.     

BRINK:基于局部质变因子的聚类边界检测算法

为有效检测聚类的边界,提出了一种基于局部质变因子的聚类边界检测算法(BRINK).该算法使用加权欧式距离技术解决现有聚类边界检测算法不能适用于高维数据的问题,根据局部质变因子在聚类的边界具有稍大于1的特征来识别聚类的边界.实验结果表明,该算法能有效的检测出聚类的边界.