Target abnormal detecting method and device based on JPEG image

Abstract

The invention relates to a method and a device which are used for detecting target abnormality based on JPEG images, comprising a JPEG image input unit and a JPEG image abnormality detecting unit connected with the JPEG image input unit through a communication interface. The method is characterized in that when the JPEG image is transmitted by the communication interface, the JPEG image abnormality detecting unit performs Huffman decoding to the JPEG image and obtains direct current discrete cosine coefficients DCY, DCU and DCV; whether target abnormality exists in a monitored site is judged according to the comparison of the direct current discrete cosine coefficients; and corresponding switch amount is output. Compared with the prior art, the invention directly processes the JPEG image data, omits the process of decoding the JPEG image into the YUV data, thereby the abnormity target detecting speed is greatly increased By the method, required data storage space, required access times and required calculation magnitude are not great so that the hardware resource of the system is greatly reduced.

Description

technical field The invention relates to a JPEG image-based object abnormality detection method and device. Background technique With the improvement of people's quality of life, the monitoring system develops rapidly. Using image detection technology to achieve monitoring is a common method. At present, the commonly used method for abnormal detection of image objects is the frame difference method. For example, the Chinese patent application number 200510001843.4 discloses a method and system for remote monitoring. The method for judging whether abnormal objects appear on the monitoring site is: collecting images in YUV format Data, remove the UV component in the image, and use the edge extraction algorithm to extract the edge of the gray image. Compare the edge of this grayscale image with the edge of the previous grayscale image, and judge whether the number of edge points whose position difference is outside the set range is greater than a certain threshold, and if so,...

AI Technical Summary

Problems solved by technology

There are two main problems with this method: first, the uncompressed YUV format has a large amount of image data, so the amount of calculation is also large when processing, and it takes a long time from image data input to the end of result judgment, and the required system resources are relatively large. Many, can't be suitable for the real-time requirement in the embedded system; The 2nd, if also adopt existing method, must decode and Obtain image data in YUV format, which is expensive to implement

Images