DCNN CCTV Apparatus for CCTV Video Analytics Based on Object-Image Recognition DCNN

摘要

The present invention relates to an apparatus for analyzing a CCTV image based on an object image recognition DCNN. According to an embodiment of the present invention, the apparatus for analyzing a CCTV image based on an object image recognition DCNN includes: an image conversion unit receiving a plurality of video frame images provided by a video providing apparatus and converting the received video frame images into a format for reducing the load of image processing; and a motion-based image analyzing unit extracting object images of a plurality of tracking objects for motion tracking from the video frame images converted into the format, applying the extracted object images to a DCNN using an object image recognition method to classify the plurality of tracking objects based on a result of the application, and detecting an event based on a designated rule and information on tracking of objects, in which a user is interested, obtained by removing, from the classified tracking objects, wherein the objects, in which the user is not interested, are identified as being deviating from designation criteria among the classified tracking objects. According to the present invention, accuracy for detecting objects is increased by using a DCNN-based image recognition technology to which deep learning is applied.