Research on the Real-Time of the Perception between Objects in Internet of Things Based on Image

Abstract

In this paper, the application of video-based information processing technology in the Internet of things has been researched. Currently, the real-time of most video-based identification methods is not satisfactory. The Hausdorff distance plays an important role in object recognition. However, when comparing the relationship between objects, the traditional Hausdorff distance even some modified Hausdorff distances need to traverse all the points of the image to be matched, which is a non-linear operator. In order to deal with the real-time problem, an improved Hausdorff distance algorithm based on central detection method is proposed. Due to narrowing the search range of space when calculating the Hausdorff distance, the computing speed has been improved compared with the traditional Hausdorff distance for object recognition. An example of vehicles recognition is used to demonstrate the efficiency of the proposed method. Experimental results show that compared with the LTS-HD, the new Hausdorff distance can not only guarantee the accuracy of matching but also enhance the perception between objects in real time.