Apparatus and method for motion detection in video
AssigneeLogitech Europe S.A.
InventorsTodd Webb, Jeremy Noring, Charles Coffing
Life of this patent
Estimating the term…
Abstract
A motion detection method of a video camera includes calculating differences for a sum of accumulated differences (ΔSAD) for corresponding macro blocks in successive frames of video. The method may include identifying motion-generative macroblocks and regions of connected macroblocks. Candidate-motion regions may be defined in a frame and a bounding box may be formed around each of the candidate-motion regions. If corresponding bounding boxes in a plurality of successive frames have substantially the same location, size, and shape in the plurality of successive frames, and if the plurality of successive frames are in a temporal window of successive frames, then the corresponding bounding boxes may indicate motion.
Claims12
A motion detection method of a video camera comprising:
- calculating by the video camera differences for a sum of accumulated differences (ΔSAD) for corresponding macroblocks in successive frames of video;
- for each of the macroblocks in each of the successive frames, identifying by the video camera the macroblock as a motion-generative macroblock if the ΔSAD for the macroblock is greater than a motion-threshold value, or if the ΔSAD for the macroblock is greater than a flooding-threshold value and if the macroblock in a previous frame is in a region of connected macroblocks;
- identifying by the video camera regions of connected macroblocks such that each of the regions of connected macroblocks contains at least one motion generative macroblock and contains macroblocks connected to the motion generative macroblock via a sequence of macroblocks and the macroblocks in the sequence of macroblocks each has a ΔSAD value greater than the flooding threshold, wherein the macroblocks in the sequence of macroblocks touch at at least one point, and wherein the regions of connected macroblocks in a frame define a candidate-motion region in a frame;
- forming by the video camera a bounding box around each of the candidate-motion regions;
- if corresponding bounding boxes in a plurality of successive frames have substantially the same location, size, and shape in the plurality of successive frames, and if the plurality of successive frames are in a temporal window of successive frames, then determining by the video camera that the corresponding bounding boxes indicates motion otherwise determining that the corresponding bounding boxes do not indicate motion.
A motion detection method of a video camera comprising:
- setting a motion-threshold value, and a flooding-threshold value that is less than the motion-threshold value;
- calculating by the video camera differences for a sum of accumulated differences (ΔSAD) for corresponding macroblocks in successive frames of video;
- identifying by the video camera at least a first macroblock as a motion-generative macroblock based on the ΔSAD for the first macroblock being greater than the motion-threshold value;
- identifying by the video camera at least a second macroblock, that is connected to the first macroblock, based on the ΔSAD for the second macroblock being greater than the flooding-threshold value;
- identifying by the video camera a region of connected macroblocks including at least the first macroblock and the second macroblock;
- forming by the video camera a bounding box around the region of connected macroblocks;
- if the bounding box has substantially the same location, size, and shape as corresponding bounding boxes in a plurality of successive frames, and if the plurality of successive frames are in a temporal window of successive frames, then determining by the video camera that the corresponding bounding boxes indicate an area of motion, otherwise determining that the corresponding bounding boxes do not indicate an area of motion.
Dependencies are inferred from the claim text.