| 摘要 |
A video camera having camera-shake detection function divides an image sensed picture into a plurality of blocks and binary codes the luminance signal in each pixel in each block. The presence of a binary coded data of "1" is detected for each horizontal line and vertical line in each block, whereby the presence of at least one causes data "1" to be stored in the bit position in the memory corresponding to the line. Accordingly, data representing the distribution of an object in the vertical direction and the horizontal direction are obtained for each block. By calculating the data correlation between consecutive frames, a motion vector of the object for each block is specified, whereby a motion vector representing the entire picture is selected by subjecting the motion vectors of all blocks to majority decision logic process. In accordance with this selected motion vector, the movement of the object caused by camera-shake is corrected appropriately. <IMAGE> |
