摘要 |
In an apparatus for reading-out features such as characters and marks recorded on an object such as labels, boxes and cards, the object is scanned by a linear CCD array to produce a video signal and a background level is detected from the video signal on the basis of an edge signal derived from the video signal. A plurality of slice levels having span and pitch adjusted according to said background level are generated and the video signal is compared with the slice levels to produce an output video signal. Even if the level of the video signal changes abruptly, it is possible to obtain more accurate output video signal and to remove error in pattern recognition. In another preferred embodiment, an amplitude of the video signal is adjusted by a variable attenuator in accordance with the detected background level.
|