| 摘要 |
PURPOSE:To extract properly an object without causing mis-recognition by converting image data into density data depending on a quantity with respect to a preset threshold level and applying projection processing to the density data when the density data are subject to projection processing so as to emphasize an object in the image data. CONSTITUTION:A television camera 1 is used to sequentially scan a prescribed are including an object to be extracted on a line and to provide image data for each picture element to a gradation discrimination section 21. The gradation discrimination section 21 divides the image data into data to each gradation and fetches density data of each picture element together with their position data. First and second comparators 24, 25 compare the level of the density data having the position data with upper and lower threshold levels from a setting section 23 in which the upper and lower threshold levels are set in advance, and sets the gradation level of the data to the maximum or minimum. Then, the set data are converted into high level data or low level data, a projection section 27 integrates the density data for each coordinate and applies projection scanning to the integrated data, an object extract section 28 extracts an object existing area from part of the high density data to avoid mis-recognition. |
