Image capture using client device

摘要

Methods, operable on client devices, for capturing an image of a document. One method comprises: receiving one or more frames; processing each frame by emphasising horizontal and vertical structures to produce an edge image; using an edge detection algorithm on each edge image to detect edges of the document; and automatically capturing a still image of the document once its edges have been located. Another method comprises: receiving a video stream comprising a sequence of frames; generating a focus metric indicating the sharpness of each frame; establishing and varying an adaptive threshold of the focus metric based on the variation of the focus metric over a number of frames; and capturing a still image of the document once it is determined that the focus metric of a current frame is above the adaptive threshold. Another method comprises: receiving a frame containing an image of the document; using an edge detector to detect candidate edges and produce an edge image; detecting corner points at candidate edge intersections; correcting perspective so that the image corner points are at the corners of a rectangle.