| 摘要 |
The method comprises determining a specific entry signature E of an object (6) entering through an entrance (2) into a predefined region (1) by a first image sensor, storing the specific entry signature E in an entry list of a calculation unit, determining a specific exit signature A of an object exiting the predefined region through an exit (3) by a second image sensor, storing the specific exit signature A in the exit list of the calculation unit as an exit signature A o with index value j=0, and calculating similarity values by the calculation unit. The method comprises determining a specific entry signature E of an object (6) entering through an entrance (2) into a predefined region (1) by a first image sensor, storing the specific entry signature E in an entry list of a calculation unit, determining a specific exit signature A of an object exiting the predefined region through an exit (3) by a second image sensor, storing the specific exit signature A in the exit list of the calculation unit as an exit signature A o with index value j=0, and calculating similarity values C n for n, where n is -N to +N, by the calculation unit, where N is a predeterminable. Each entry signature is provided with an index value i reflecting the temporal sequence of the entries. A value L reflecting the length of the queue of objects is incremented by one. The index value of each exit signature A j already present in the exit list is incremented by one, and exit signatures A j having an index value j greater than a predeterminable value M are deleted from the exit list. The similarity U m n is measured between the exit signature A m and entry signature E (L+m+n), where m is 0-M, and is added to C n. The method further comprises decrementing the value of L by the value one, correcting the value of L by comparing the similarity values C n with similarity value of C o and in the presence of a C x value with index value n=x of a higher similarity than the similarity value indicating C o, and correcting the value of L by adding the index value x. Each similarity value C n is added to a corresponding histogram value K n in a histogram list, and the new histogram value K n is stored in the histogram list. A correction of the value of L occurs when a histogram value K x with index value n = x indicates a higher similarity histogram value K o by correcting the value of L and adding the index value. The calculation of the similarity values C n and the correction of the value L is not carried out after detemination of exit signature and after a predeterminable number of sensed discharge signatures A, in particular more five, ten or fifteen sensed discharge signatures A. The correction the value L is carried out if a similarity value C x or histogram value K x indicative of a higher similarity, which is higher by a predeterminable value Q as the similarity value C o or the histogram value k o, where Q indicates a 1.5-fold higher similarity. The image sensors detects entrance signatures E, exit signatures A, and color signatures of the objects. The similarity measures U m n are calculated using a distance vector between the exit and the entry signatures in a three dimensional color space. An expected new residence in the predefined area is calculated and displayed by a display device due to the determined length of the queue of the objects at the entrance. Differences between the entrance signatures and the exit signatures due to different lighting conditions at the entrance and the exit are compensated by a correction algorithm. An independent claim is included for a device for determining a length of a queue of objects. |
