摘要 |
The invention relates to a method for video analysis, in particular for video monitoring, wherein a digital input video signal is processed in real time, wherein the input video signal comprises at least one first input image and a second input image (11), which follows the first input image at least indirectly, wherein the input images (11) each have a specifiable number of pixels, wherein a background image (12) is determined from the at least one first input image and the second input image (11), and at least one foreground area (18) is determined, the location of which changes relative to the background image (12) in the second input image (11), wherein in order to secure the privacy and/or private sphere of video-monitored persons, the foreground area (18) is made unrecognizable, an output image (14) is formed from the background image (12) or the second input image (11) and the foreground area (18') made unrecognizable, an output video signal is formed from output images (14) that follow each other at least indirectly, and the output images (14) and/or the output video signal are output and/or stored. |
申请人 |
MATUSEK, FLORIAN;SUTOR, STEPHAN;KRAUS, KLEMENS;ROEHR, ROBERT |
发明人 |
MATUSEK, FLORIAN;SUTOR, STEPHAN;KRAUS, KLEMENS;ROEHR, ROBERT |