Article-processing device having a system for monitoring the quality of images of the articles in real time, a sorting machine, and a method

An article-processing device comprising a conveyor for moving articles in series and on edge, a camera facing transversely relative to the conveyor so as to take a digital image of each article, a data processor unit for performing processing on the pixels of the digital image of the article, and means for inspecting/monitoring the quality of the digital images taken by the camera, wherein:

• said conveyor has a back wall of a dark color and in front of which, relative to the camera, each article passes;
• said camera is arranged to take a digital image within which the article lies against a background corresponding to said back wall of the conveyor, said digital image including both said background and the face of the article;
• and
• said inspection/monitoring means are arranged to scan rows of pixels in each digital image including pixels corresponding to said background in order to extract, for each scanned row, a pixel having a light intensity extremum, said inspection/monitoring means further being arranged to detect, from a plurality of digital images, contrast among said extracted pixels, such contrast being indicative of degradation of the quality of the digital images taken by the camera.

A method of inspecting/monitoring, over time, the quality of the digital images taken by a camera for an article-processing device, said method comprising the following steps:

• taking a digital image of an article using the camera in such manner that the article lies within and against a background of dark color, said digital image including both said background and a face of the article, and recording said digital image in a memory of a data processor unit;
• scanning rows of pixels in the digital image including pixels corresponding to said background in order to extract, for each scanned row, light intensity extrema from the pixels of the digital image;
• and
• analyzing said extrema and detecting, on the basis of a plurality of digital images, contrast among said pixels extracted from said digital images so as to identify any degradation in the quality of the digital images.