智能视觉物联网中视觉特性的提取及视觉标签的建立

发布时间：2019-06-21 02:41

【摘要】：智能视觉物联网(IVIOT),即具有视觉感知功能的物联网。它通常由四个部分组成:视觉传感器、视觉信息传输、视觉信息处理和物联网应用。智能视觉物联网是物联网的升级,它将图像传感器传来的图像,利用图像处理技术或计算机视觉技术等,进行智能化识别、定位、跟踪,便于我们对人、车、物进行智能化管理。智能视觉物联网中最重要的一个核心技术就是视觉标签技术。它可以将视频或图像中的内容进行识别、理解和分类,并为其"贴标签",并将被识别对象所对应的标签信息内容显示出来。本文根据课题需要,设计并实现了一套基于人、车、物的视觉标签系统。系统主要包括人物识别模块,车辆识别模块,物体识别模块。在这个系统中,可以根据用户需求,选择一幅图片,系统会自动识别该图片中的内容,并显示与之相关的其它图片以及其信息标签。在人物识别模块,本文基于人脸识别技术,采用主成分分析(Principal Component Analysis,PCA)算法和支持向量机(Support Vector Machine,SVM)算法相结合的方法进行人脸识别。首先利用PCA算法进行特征提取和降维,再利用SVM算法进行分类和识别。在车辆识别模块,本文采用基于颜色的车牌识别算法,对智能视觉物联网中获得的图像中的车辆,进行车牌定位、车牌校正、字符分割、字符识别等处理,最终识别出车牌号码。在物体识别模块,本文采用基于卷积神经网络的物体识别方法,完成了同类物体的识别(本文以水杯为例)。最后,本文将三个模块整合到一个系统中,使人、车、物一一对应,开发了一套具有视觉标签功能的人、车、物智能识别系统。实验结果表明,该系统具有很好的性能,能够满足用户的基础需求。
[Abstract]:Intelligent visual Internet of things (IVIOT),) is the Internet of things with visual perception function. It usually consists of four parts: visual sensor, visual information transmission, visual information processing and Internet of things application. Intelligent vision Internet of things is an upgrade of the Internet of things. It uses image processing technology or computer vision technology to intelligently identify, locate and track the image from the image sensor, which is convenient for us to manage people, cars and things intelligently. One of the most important core technologies in the intelligent visual Internet of things is visual tagging technology. It can identify, understand and classify the content of video or image, and "label" it, and display the content of label information corresponding to the identified object. According to the needs of the project, this paper designs and implements a set of visual label system based on human, car and object. The system mainly includes character recognition module, vehicle recognition module and object recognition module. In this system, a picture can be selected according to the needs of the user, and the system automatically recognizes the contents of the picture and displays other pictures and its information labels. In the character recognition module, based on face recognition technology, this paper uses principal component analysis (Principal Component Analysis,PCA) algorithm and support vector machine (Support Vector Machine,SVM) algorithm to carry out face recognition. Firstly, PCA algorithm is used for feature extraction and dimension reduction, and then SVM algorithm is used for classification and recognition. In the vehicle recognition module, this paper uses the color-based license plate recognition algorithm to identify the license plate number of the vehicle in the image obtained from the intelligent visual Internet of things, such as license plate location, license plate correction, character segmentation, character recognition and so on. In the object recognition module, the object recognition method based on convolution neural network is used to complete the recognition of the same kind of object (taking the water cup as an example). Finally, this paper integrates three modules into one system, and develops a set of intelligent recognition system of human, car and thing with visual label function. The experimental results show that the system has good performance and can meet the basic needs of users.
【学位授予单位】：内蒙古大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.4;TN929.5

【相似文献】