Abstract
General-purpose imagers have been developed into high resolution and high frame rate, and simultaneously attempted to apply to the measurement and the recognition of objects in various scenes, recently. However, with the rise of image data volume, the transfer-time of the data and the loads for image processing are also increased. In the cases of binary image processing frequently used as pre-processing for the applications of general-purpose imagers, efficient extracting only the information of interest makes the effectiveness higher in the whole image processing. From this point of view, we propose a Pixel Position Extraction (PPE) Method for efficient processing of binary image data using a general-purpose imager. By the proposed method, as the image data captured by imager is binarized, only the x and y coordinates in object domain of the binary image are extracted in real-time on hardware. Thereby, processing volume becomes proportional to the net pixel numbers in object domain regardless of the position of objects in the captured image and the resolution of an imager. By processing only the coordinate data of the pixels of interest in a binary image, the method completely eliminates the redundant processing which is caused by sequential scanning the pixels of background domain in a binary image on software. This paper describes the principle of the proposed method, its algorithm, the validation of data volume change by this method, and the overview of experimental prototype system. Finally, verification experiment by the comparison with the conventional disposal, and its results are discussed.
