|
With the computer technology and network technology, the rapid development of modern society gradually to digital and information technology, network forward. At the same time, based on IC card consumption is also increasing. Most IC card consumer spending data just saved (such as the consumption of time, place, amount, etc.), while others impersonate the user of the transaction, only the record consumption data information is difficult to achieve \who impersonate the user to consume. At present in most public places have installed video surveillance equipment, video surveillance equipment can record consumption of the whole process can be done \But in the \For IC card consumption situation, proposed video surveillance technology in IC card consumption, making the IC card consumption process has findability, to a certain extent, can effectively monitor IC card consumer safety. The main tasks: (1) In the analysis of the existing IC card consumer systems and video surveillance technology, based on the proposed video surveillance technology in IC card consumption, and gives the design idea and the overall framework. In the IC card consumption, while consumption data stored in a captured video image, this image as a key frame to establish contact with the captured video, which will be based on consumer video clip keyframes for storage. Users to retrieve, you can browse through the collection of images directly consumer video clips. For non-consumer case, capture video through video surveillance equipment, and video analysis and processing, can achieve content-based video retrieval. Through the analysis of the video, the user submits the image through the system intelligent analysis, the relevant video feedback to the user. (2) the collection of video content-based analysis and processing. First, the use of a sliding window of the video shot detection algorithm, the video is divided into semantic units of the lens and then the lens to extract key frames, keyframes representative of the lens with the contents of the key frame extraction algorithm using the histogram average method. For key frame extraction, using color and texture combination of methods. The resulting keyframe feature-based clustering, clustering algorithm uses artificial immune clustering method. Image retrieval based on the value of the input image feature extraction, with the characteristic value is compared with the cluster centers to determine which category an image to be retrieved in the class, so that you can reduce the image search, thus speeding up the query speed. Through experimental verification, this design method can effectively improve \
|