| 摘要 |
PURPOSE:To specify the existence area of a recognized object and to execute the recognition processing without fail even when the position of the recognized object is largely dislocated by determining automatically a recognizing area at the time of the first image sampling and executing the recognition at the time of the second sampling. CONSTITUTION:A camera 3 photographs a video in the sight of the camera 3 and outputs an image signal in accordance with the luminance of respective picture elements. The image signal is converted to a digital signal by an A/D converter, binary-coded by a binarization threshold and set to the value to the extent in which the output diameter of a nut can be clearly found. The binary-coded data are stored into an image memory 9 and a processing area is determined. next, the binarization threshold is set to the value where the quality is clearly displayed, and from the image data in the recognizing area, the quality is recognized. By the recognizing result, a non-defective and a defective are seperated by a nut quality dividing part 13. |
