| 摘要 |
A solution is proposed for recognising articles to be processed. A corresponding method (600) starts with the step of storing (603-639) an indication of an expected set of repetitive processing of the articles; each article of the expected set is defined by a code compliant with a predefined coding, an offset from a representation of the code on the article, and a frame of predefined size representing a portion of the article at said offset from the representation of the code. The method then provides for repeating (642-696) the following steps for each current article to be processed, corresponding to an expected article of the expected set (being defined by an expected code, an expected offset, and an expected frame). First of all, there is acquired (645) a current image representing at least part of the current article. The representation of a current code being compliant with this coding is searched (648-669) in the current image; a position in the current image of the representation of a corrupted code partially compliant with the coding is saved in response to a negative result of the search of the current code. In this case, a current frame of said size is extracted (681-684) from the current image at the expected offset from the position of the representation of the corrupted code. The current frame is compared (687-689) with the expected frame; a correctness of the article is determined (690) according to the comparison of the current frame with the expected frame. |
