| 摘要 |
The invention relates to a method for determining the length of a queue of objects (8) in a predefined region (1) with at least one entrance (2) and at least one exit (3), in which errors in the detection of objects (6, 7) entering or exiting the region (1) are corrected during the determination of the length of the queue. In a first step, a specific entry signature E of each object (6) entering through the at least one entrance (2) into the predefined region (1) is determined by means of at least one first image sensor (4). Thereafter, the specific entry signature E is stored in an entry list of a calculation unit (9), wherein each entry signature E is provided with an index value i reflecting the temporal sequence of the entries. In addition, a value L reflecting the length of the queue of objects is incremented by one. A specific exit signature A of an object (7) exiting the predefined region (1) through the at least one exit (3) is also determined by at least one second image sensor (5), the specific exit signature A being stored in an exit list of a calculation unit (9) as an exit signature A0 with index value j=0. The index value of each exit signature Aj already present in the exit list is incremented by one and exit signatures Aj having an index value j greater than a predeterminable value M are deleted from the exit list. Then, similarity values Cn for -N<=n<=+N are calculated by the calculation unit (9), wherein N is a predeterminable value that defines a specific number N of entry signatures of objects that have exited the predefined region before, or will exit after, the object with the entry signature EL which, based on the length L of the queue, will next exit the predefined region. In the process, respective similarity measures Umn between an entry signature Am and an exit signature EL+m+n, where m runs from 0 to M, are calculated and summed for Cn. The value L is decremented by the value one. Then, the similarity values Cn are compared to the similarity value C0. If there is a value Cx with index value n = x indicating a higher similarity than the similarity value C0, the value L is corrected by adding the index value x. |
