Shot boundary detection based on mutual information and canny edge detector

Shot boundary detection servers as the preliminary step to video retrieval. Most of error detections in the present algorithms are caused by object and camera. Many researchers make use of mutual information to detect shot boundary, and the effect is good, however, object motion reduce performance of algorithm in the methods. In this paper, the author present a novel method of shot boundary detection based on the knowledge of mutual information and canny edge detector. We extract video frame edge using canny edge detector, then distinguish object motion and shot transform effectively by analyzing frame edge differences, reducing error detections caused by object motion, improving recall and precision. Experiments prove this method is robust to object motion, performance is better.     

高压输电线路巡线机器人障碍物视觉检测识别研究

障碍物的检测识别与定位是高压输电线路自主巡线机器人中的关键技术之一.针对220 kV单分裂输电线路的结构特点,提出了一种基于视觉传感器的障碍物检测识别方法.首先,对采集的图像进行膨胀、腐蚀和高斯平滑等处理,以减少图像噪声;然后,利用Otsu算法对Canny算子进行改良来提取图像边缘,以减少光线变化带来的影响;最后,提取边缘图像中某些图形基元并施加结构约束,实现障碍物的检测识别.大量的试验研究表明,该方法能有效地检测识别高压输电线路的导线、防震锤、悬垂绝缘子和耐张绝缘子等结构物.     

Improvement to the scanning electron microscope image colourization by adaptive tuning

An improvement to the previously proposed Canny optimization technique for scanning electron microscope image colourization is reported. The additional process is adaptive tuning, where colour tuning is performed adaptively, based on comparing the original luminance values with calculated luminance values. The complete adaptive Canny optimization technique gives significantly better mechanical contrast on scanning electron microscope grey-scale images than do existing methods.     

基于图像融合技术的阶梯型边界检测

在Canny算子和模角分离的小波变换(MAS)算法的基础上,针对Canny 边缘检测算法抗噪性差以及MAS算法检测边缘的连续性差的缺点,利用融合技术,提出了一种能够结合Canny算子和MAS算法的方法,保留了各自算法的优点,实现了对阶梯形边界的检测,得到了理想的效果.     

Canny算子在织物疵点检测中的应用

在分析斜纹织物纹理结构规律性的基础上,运用Canny算子对织物疵点图像进行了边缘检测,获得了织物疵点二值图像,再由灰度统计的方法生成投影图,获取织物图像的特征波形,并通过阈值过滤,将疵点定位,从而正确检测织物疵点.在Visual C 6.0环境下编程并调用Matlab6.5图像处理函数实现了本算法.     

基于照片的三维人体模型研究与实现

本文通过人体正面和侧面两张照片,采用基于模板的方式提取人体的特征数据,利用这些特征数据最后建立可产生服装原型衣片的个性化三维人体模型。     

基于全局运动估计的视频对象分割算法

视频全局运动(摄像机运动)所表现的视频序列之间的时间相关性,较其它视频特征更能表达视频序列的高层语义信息.为了能够有效快速的得到视频的全局运动,通过对视频运动估计方法的研究,提出了一种新的基于奇异值分解(SVD)的视频全局运动估计算法.该方法首先通过块匹配法得到局部运动场,利用矩阵的奇异值分解估计全局运动参数,然后运用形态学运动滤波得到前景运动目标的粗略掩摸图像,最后综合利用此掩摸图像和边缘信息分割出运动目标.试验表明,提出的算法能够分割出具有全局运动特征的视频序列中的运动目标.     

基于边缘检测和四点插值细分的SAR图像去噪

针对SAR图像相干斑滤波中存在的降低相干斑与有效保持细节信息这一矛盾,提出了一种基于四点插值细分的SAR图像去噪的新算法,将四点插值细分规则运用到图像去噪中,并与边缘检测相结合。先用canny算子提取图像边缘,进而通过四点插值细分方法分别对边缘图像和原始图像进行去噪,然后再对边缘信息进行边缘信息的重构,得到新的去噪图像。并通过等效视数、边缘保持指数等评价指标对去噪结果进行了评价。实验结果表明,与其他去噪方法相比,该算法在有效地去噪的同时,可有效地保留图像的边缘信息,具有较好的去噪结果。     

基于数学形态学和Canny算子的边缘提取方法

提出了一种基于数学形态学和Canny算子的边缘提取方法。该方法利用数学形态学开运算估计背景,将原始图像与背景进行几何运算,在处理后的图像上运用Canny算子提取边缘。实验结果表明,该方法明显优于传统的经过中值滤波后再进行Canny算子的边缘检测效果,为后续的特征提取、目标识别提供了良好的基础。     

Human and Machine Vision Based Indian Race Classification Using Modified-Convolutional Neural Network

The inter-class face classification problem is more reasonable than the intra-class classification problem. To address this issue, we have carried out empirical research on classifying Indian people to their geographical regions. This work aimed to construct a computational classification model for classifying Indian regional face images acquired from south and east regions of India, referring to human vision. We have created an Automated Human Intelligence System (AHIS) to evaluate human visual capabilities. Analysis of AHIS response showed that face shape is a discriminative feature among the other facial features. We have developed a modified convolutional neural network to characterize the human vision response to improve face classification accuracy. The proposed model achieved mean F1 and Matthew Correlation Coefficient (MCC) of 0.92 and 0.84, respectively, on the validation set, outperforming the traditional Convolutional Neural Network (CNN). The CNN-Contoured Face (CNN-FC) model is developed to train contoured face images to investigate the influence of face shape. Finally, to cross-validate the accuracy of these models, the traditional CNN model is trained on the same dataset. With an accuracy of 92.98%, the Modified-CNN (M-CNN) model has demonstrated that the proposed method could facilitate the tangible impact in intra-classification problems. A novel Indian regional face dataset is created for supporting this supervised classification work, and it will be available to the research community.