Method is for handling purchased goods at automatic cashdesk check-out and involves prior storage in memory of information of a measurable parameter of articles to be handled

Title:

- 1 Automatic cash register system

The present invention relates to a cash register system, usable for use in, for instance, a supermarket, where often multiple customers are present at the same time, and in which usually every customer wants to pay for several groceries. The customer has collected those groceries in a shopping basket or shopping cart and arrives at a checkout. Traditionally, that cash register is staffed, that is to say that staff (a cashier) are present. With conventional cash register systems, that cashier will check the price of each individual item on a cash register. Today, the items are usually bar-coded, and more modern POS systems have a bar-code reader connected to the cash register; In such systems, the cashier will pick up each item and move that item, with the barcode facing the barcode reader, through the barcode reader's field of view. This is already much faster than manually entering the price.

The necessary presence of operating staff (cashier) is a drawback of the known cash register systems. With a cash register system, personnel costs form a significant part of the operating costs. Furthermore, space will have to be kept clear for the operating staff, which implies that successive POS systems must be placed at a relatively great mutual distance; conversely, at a given available space, this means a limitation at a location for the total number of POS systems that can be arranged next to each other. It has been found that there is a growing need among customers to pay for groceries completely independently and with the shortest possible waiting time. This waiting time can be shortened by installing multiple POS systems. Therefore, cash register systems have been developed that can operate without the necessary presence of a cashier. Such systems will hereinafter be referred to by the term automatic checkout system.

Π 9 1 ft 4 H

Known automatic cash register systems are based on the absence of a barcode on each article, and include malicious barcode reading means as well as transport means for transporting the items to be settled along those reading means, such that those reading means can read the barcodes placed on the articles.

The bar code reading means usually comprise at least one scanner, which is adapted to scan an article with a laser beam, and deduce the information stored in the bar code from the intensity variations of the received back light. Since such scanners are well known and useful in the implementation of the present invention, while furthermore the present invention is not concerned with improving a similar scanner, the operation and construction of a nner will not be further discussed here. Suffice it to note that the bar code reading means is a signal output where they provide a read signal representative of the information stored in a recognized bar code. That signal can take the form of a square signal that presents the black and white parts of the barcode, but it can also already take the form of a ral, represented by a digital code. In any case, the output signal is useful for a signal processing unit to look up the desired data of the article in question in an associated database, including the ice. It is even possible that the ice itself is encoded in the output signal. In the following, the output signal k output from the bar code reading means will be designated by the term bar code information signal, and the information contained in that signal will also be indicated by the term bar code information.

It is of great importance that the items are recognized correctly and quickly by the cash register system. If an item is recognized incorrectly, a reverse price will be charged. If an article is not recognized, the intervention of personnel is still necessary to judge the article and enter the correct data in the system ir.

- Ö An important problem in this connection is, for example, that it is possible in practice for several articles to be guided simultaneously along the reading means, at such a small mutual distance, that the reading means regard those articles as one combined article. This problem is also referred to as a cluster problem in the following. Different scenarios are then possible. First, it is possible that the barcode reader detects both barcodes, and therefore may not decide on a single detection. Secondly, only one barcode may be recognized; this can occur if the second barcode is covered by the other item, but also if two identical items are next to each other. In the first example, the system will probably notice that an error situation is occurring; in the second example, it is quite possible that the system is not aware that an error situation is occurring.

It is therefore an important object of the present invention to provide an automatic checkout system which has been improved such that the above recognition problems are avoided as much as possible, so that the automatic checkout system has a high degree of reliability and reliability. More particularly, the present invention aims to provide an automatic checkout system in which error situations of the above-described type are recognized with greater degree of certainty.

In shops such as supermarkets, it is common for there to be items whose price is not known per item but per unit weight. Examples of this type of articles are vegetables, fruit, cold cuts, etc. In order to be able to process these kinds of articles with the known automatic checkout systems, it is now customary that the articles in question are weighed in the store, that a price is calculated, and that a barcode label is printed in which the calculated price is processed, which label to be affixed to the item. Especially with fresh articles, these are actions that the consumer must perform, and that time

<img file="NL1021040C2_D0001.tif" />

<img file="NL1021040C2_D0002.tif" />

ίο:

,; 4.

i

-4and. It is a further object of the present invention to reduce the ok dot problem.

According to an important aspect of the present invention, the scanned articles are also always determined, and the measured weight information is also provided to the signal processing unit of the system. This signal processing unit can use the right information as a check on the bar code information, by comparing the measured weight with a yacht weight derived from the bar code information. In the case of Lkelen priced per unit weight, it is also possible that the barcode information relates to the article type, and possibly to □ row per unit weight, and that the price to be settled is determined by the signal processing unit on based on measured weight.

In a further embodiment, a color histogram is made of the scanned image using one or the color cameras, and a signal representing the obtained hour histogram is provided to the signal processing unit of the POS system, hour histogram information used as a check on the bar code information .

alternative identification possibility offered for the barcode detection does not lead to an unambiguous that the obtained alternative for or

This is the case; identification.

Said and other aspects, features and advantages of the present invention will be further elucidated by the following discussion of an embodiment of a cash register system according to the present invention with reference to the drawings, in which: Figure 1 schematically shows a side view of an article marking station ;

Figure 2 schematically shows an otral processing unit associated with a cash register;

Figure 3 schematically shows a flow chart of steps to be performed by the ztral processing unit;

Figure 4 shows an adaptation of the flow chart of Figure for identification based on a color histogram;

i A 0

Figure 5 schematically shows a flow chart of an operator protocol;

Figure 6 schematically shows a screen presentation of information to be provided to an operator;

Figure 7 schematically illustrates an automatic shopping cart checking system;

figure 8 schematically shows another embodiment of an article recognition station.

Figure 1 schematically shows an article recognition station 1 of a cash register system according to the present invention, hereinafter also referred to as automatic cash register 1. The automatic cash register 1 comprises bar code reading means 20 with a signal output 21, as well as transport means 10 for transporting articles 2 along the bar code reading means 20. The barcode reading means 20 may comprise one or more barcode scanners, but this is not shown in figure 1 for the sake of simplicity. Alternatively, the bar code reading means 20 could comprise, for example, an image pick-up member, such as, for example, a CCD camera, wherein a bar code is then analyzed on the basis of image processing techniques known per se, as will be clear to a person skilled in the art. The only essential thing is that the bar code reading means 20 at the signal output 21 have a bar code information signal S.₂o Offering that is representative of the information stored in a recognized barcode.

The transport means 10 are preferably, although not necessarily, of the type comprising an endless conveyor belt 11, as known per se for POS systems. For the sake of simplicity, only two drive rollers 12 of a mechanism for driving said conveyor belt 11 are shown. The direction of movement of the conveyor belt 11 in figure 1 is from left to right. The articles 2 to be scanned lie on the conveyor belt 11, it is assumed that the articles are placed by the customer on the conveyor belt 11 in such a way that the barcode (not shown in figure 1) is visible. In case of an article

-6 if there is one barcode, that barcode may therefore face downwards. This limitation can be sieved if the articles are provided with multiple barcodes on different parts of their surface, if the barcode reading means are arranged to read barcode also present on the bottom surface.

A detection area of ​​the bar code reading means 20 in Figure 1 indicated by reference numeral 22. The springing of an article is detected by a pass detector associated with the bar code reading means 20 in Figure 1 at the left edge of the detection area 22 sited). The pass detector can, for example, be designed as a light barrier, as will be known to a person skilled in the art.

The automatic cash register 1 is further provided with means 30. Optionally, a separate weighing station could be empty, but this required extra costs and a more complicated construction. Therefore, those weighing means 30 are preferably, and as shown, associated with the sporting means 10. More specifically, the sport belt 11 comprises three successive sections 13, 14, 15, and the middle conveyor section 14 is provided with a weighing sensor 30. The left conveyor belt section 13 is a feed section, and the right conveyor belt section 15 is a feed section. The middle conveyor belt section 14, also referred to as weighing conveyor belt, conveys at a faster speed than the feed section 13; in a suitable liner form, the speeds are approximately cm / sec (13) and 60 cm / sec (14), respectively. As a result of this speed difference, a separation of the successive articles in the section region 22 of the bar code reading means 20 occurs.

In Figure 1, the middle section 14 corresponds to the ection area 22 of the barcode reading means 20, so that item to be scanned can be weighed simultaneously with picking. Although it is possible that the article 2 is kept silent during the weighing, a dynamic weighing is preferred: went during transport); after all, this allows for a more robust throughput, and moreover, it should

-7 stopping and starting again of the conveyor belt 11 a greater wear occurs.

Since conveyor belts with an integrated weighing sensor, suitable for dynamic weighing, are known per se and are commercially available as a standard product, the construction of the weighing sensor 30 will not be further described here. Suffice it to note that the weighing sensor 30 is provided with a signal output 31, and there provides a measuring signal S3Q which is representative of the currently measured weight. This measuring signal will hereinafter also be referred to as weight information signal.

Figure 2 schematically shows a central processing unit 100, which is part of the automatic cash register. In a preferred embodiment, that central processing unit 100 is implemented by a suitably programmed computer, such as a PC. The central processing unit 100 may also, for example, be designed as a separate, suitably programmed processor. This central processing unit 100 is associated with a cash register 110, which may be equipped in the usual manner with an ATM 111 and a cash strip printer 112. The cash register 110 may also be equipped with an ATM, to allow payment with paper money and / or coins. which, however, is not shown in figure 2 for the sake of simplicity. Alternatively, it is also possible for the customer to go to a separate, manned checkout station with the printed receipt to pay.

Furthermore, the automatic cash register 1 is preferably provided with a display 130, for example a monitor or an LCD screen, on which, for example, information about the progress of the checkout process can be shown to the customer, and on which instructions for the customer can be shown. . One or more keys may be present for the possible input of data and / or orders by the customer, as symbolically indicated at 140.

The central processing unit 100 is provided with a first signal input 102 which is coupled to the signal output 21 of the bar code reading means 20, so that the rale processing unit 100 has the bar code information._2C receive. The central processing unit 100 is provided with a second signal input 103 which is coupled to the signal output 31 of the weighing sensor 30, so that the iral processing unit 100 receives the weight information signal.

Furthermore, the central processing unit 100 is equipped or connected to a memory 120 with a file with article information stored therein.

The operation of the cash register system of the present binding will now be explained with reference to aur 3, which schematically shows a flow chart of the steps to be performed by central processing unit 100 when calculating the price of an item offered. It is to be understood that those steps may result from the expiration of a suitable computer program.

A customer arrives at the automatic checkout 1, and enters a checkout cycle, for example by pressing a start button (step) not shown for the sake of simplicity. If customers are required to have a special authorization to use the automatic checkout 1, that oration is checked (step 302), which the customer is asked to do (via display 130) to, for example, introduce a .ntenpas and / or enter a code. tests. As the central processing unit 100, the customer authorization is in order ^rthe conveyor belt 11 is started, and the customer can place n messages on the feed section of the msport belt 11. It is then the intention that the messages i be placed one by one, whereby the relatively fast moving sports belt 14 (approx. 60 cm / cm) s) normally a pass

Physical separation between successive items 2 will be ensured.

An identification cycle is started when an object passes the pass detector 23. The reading means 20 reads the barcode of that article, and generates the barcode information signal S.₂g, which is received by central processing unit 100 (step 311).

-9 First, the central processing unit 100 examines the bar code information signal S.₂o, wherein the central processing unit 100 checks whether there is a complete and / or valid barcode (step 312).

If the bar code information is incomplete or invalid, the central processing unit 100 cannot process the information and the central processing unit 100 detects an error situation. The central processing unit 100 then initiates an alternative identification protocol, or enlists the help of a human operator (step 500). Examples of an alternative identification protocol will be discussed later.

If the central processing unit 100 concludes that the barcode information corresponds to a complete and valid barcode, the central processing unit 100 searches the article information memory 120 for the known data of the relevant article (step 313), and the central processing unit 100 determines the expected weight Gx of that item (step 314). Also, the central processing unit 100 receives the weight information signal S₃o (step 315), and the measured weight Gm derived therefrom is compared with the expected weight Gx (step 316). If the measured weight Gm within predetermined tolerance limits corresponds to the expected weight Gx, i.e. the result of the barcode detection is confirmed by the result of the simultaneous weight measurement, the central processing unit 100 concludes that the identification is successfully successful expired. In that case, the central processing unit 100 passes on the necessary information to the cash register 110, namely the price to be settled and preferably also the description on the receipt (step 317).

If the measured weight deviates from the expected weight by more than the predetermined tolerance, the central processing unit 100 activates the alternative identification protocol (step 500).

- 10A large part of the cluster problems that appear in practice can therefore be avoided with certainty.

The identification cycle always begins when a new boulder 2 passes the pass detector 23 (step 321). If there are more items, the customer indicates this, for example by pressing a stop button not given for the sake of simplicity (step 322). The conveyor 11 is then set. This also happens if no articles are detected during a predetermined timeout.

The customer pays the total amount to be settled, for example by means of the ATM 111 (step 330). the settlement cycle has also ended (399), and the imatic cash register 1 is ready for a subsequent settlement cycle.

Alternatively, the customer can receive a checkout receipt, payment is made at a separate checkout station. In this case, the automatic cash register 1 is already ready for an on-going billing cycle when the actual payment still has to be made at that settlement station.

It should be noted that the present invention is primarily concerned with determining the correct price of the converted items. In a practical implementation of the cash register system, it will of course also be important that the calculated price is paid correctly. It is evident that this is supervised by a supervisor, a person who supervises the proper operation of other POS systems. It is also possible that the scanned Lkelen are held by the cash register system and only released after payment.

It has been stated above that the central recognition unit 100 determines the expected weight Gx of a cell 2. This can be implemented in several ways. It is wise to distinguish between genus information and species information. US information means that the information relates to a certain group of articles with in principle several identical specifications, such as, for example, cartons: all cartons of sugar have the same price, and in

<img file="NL1021040C2_D0003.tif" />

-11 basically the same weight. However, the individual weights of the individual items within that group will show a certain spread. Species information means that the information relates only to one specific individual article.

First, it is possible that in the data file that the central processing unit 100 consults when recognizing the article, the weights of the articles are also stored. In that case, the central processing unit 100 can determine the expected weight Gx of the relevant article simply by reading out the weight information stored in the data file. This implementation is particularly suitable for genus information.

Secondly, it is possible that the barcode itself contains information regarding the weight of the individual item in question. This therefore involves species information. In that case, the central processing unit 100 can determine the expected weight Gx of that article simply by evaluating the barcode read. A limitation of this implementation is that the weight must be measured when applying the barcode, and the barcode to be written must be adapted to the individual item. However, this solution is ideally suited for use with fresh articles such as meat products, vegetables, etc., where the price is known in advance only per unit weight, and where the price of an individual article is precisely determined by weighing that article. The barcode to be written is thus calculated after weighing the article, and since the weight is known at that time, that weight can be incorporated in the barcode to be written. With existing equipment, this only requires a relatively simple adjustment of the software. Furthermore, an important advantage of this implementation is that the weighing of the individual items and the generation of the adjusted barcode takes place within the store where the present cash register system will be installed, so that the administrator of the present cash register system (in this case the relevant retailer) so implementation

- depends on the cooperation of producers of the throats.

In a variant of this implementation it is possible when weighing (and possibly pricing) the relevant .kei only a sequence number is encoded and included in the epj es code, while the combination of sequence number is stored in a weight specified by the irale. processing unit 100 memory to be consulted. The 10th by which the central processing unit 100 determines the mighty weight is very similar to the method described above: by reading the barcode, the central processing unit 100 itifies the article, in particular the serial number of that article, and in memory the central reads processing unit 100 the weight associated with sequence number.

The present invention, however, provides the insight that it is not necessary to weigh such articles (such as articles) in the store, but it is sufficient to use a general barcode including an indication of the type as well as the price per unit weight, or price per kilo). In that case, the identification loop may expire in modified form. Steps 313 and can be skipped, and after step 315, the grating processing unit 100 calculates the item price based on measured weight Gm and the bar-derived js per unit weight, then proceeding to step

It has been stated above that the measured weight dt is derived from the weighting signal S30. If only one single item 2 is always on the weighing belt, the ash signal is a direct measure of the weight of the affected item. In general, however, the following can already reach the weighing belt 14 before the previous article has left weighing belt 14: in that case the signal then refers to the total weight of two licks. Depending on the circumstances, there may even be more than two articles on the weighing belt 14 at the same time. Deriving ”the measuring weight Gm of an ankle

- 13 item then concerns a measurement signal processing, taking into account the magnitude of weight steps in association with the reaching and exiting of the measuring tape 14 by an article; this is also referred to as sequencing.

It has been stated above that the central processing unit takes into account a predetermined tolerance when comparing the measured weight and the expected weight. Information about that predetermined tolerance can be communicated to the central processing unit 100 in various ways.

First, it is possible that the central processing unit always has a fixed tolerance for all articles.

Second, the permissible tolerance may be article dependent and stored in the memory 120 to be consulted or incorporated in the bar code itself, in a similar manner as explained above with respect to the expected weight. Again, a distinction can be made between genus information and species information. If the permitted tolerance concerns genus information, this takes into account both the said article spread and the measuring tolerance of the measuring sensor 30. If the permitted tolerance concerns species information, this takes into account the measuring tolerance of the measuring sensor 30 of the cash register system and also the measuring tolerance of the measuring sensor of the weighing and price station generating the barcode.

Entering the weight and tolerance data into the memory 120 to be consulted by the central processing unit 100 can be performed manually, but the cash register system according to the present invention is arranged in a preferred embodiment to dynamically calculate and store that data in the memory. writing, which is particularly applicable to genus information.

To take advantage of this option to enter a weight and tolerance, for example when new ones

- 14 articles are added to an assortment, an earlier one switches the cash register system to a train mode.

according to the manager, a large number of the same panels are passed through the cash register unit 1, which number can be any but preferably is greater than 10. Incidentally, articles of different types may be mixed together. The central processing unit 100 determines the weight of each individual article, and calculates per type the tistic average of the measured weights as well as the associated spread. At the end of this measuring cycle, the central processing unit 100 writes the calculated average and the spread in said memory, after which the ital processing unit 100 can consult this data in a checkout mode.

In an advanced embodiment, the central processing unit 100 also calculates the spread the measured weights in the checkout mode, and the practical spread is written by the central processing unit 100 the said memory.

It is then possible that the practice that occurs in practice deviates from the tolerance input into the memory. In that case, the central processing unit 100 can, in the billing cycle, first compare a measured view with the permissible tolerance previously stored in the memory, and, if a deviation is found, the measured weight in the second instance equates to the measured weight. spreading occurring in practice which is written or the central processing unit 100 in said memory. If no deviation is observed during the comparison in second instance, the central processing unit 100 can still approve the article in question.

Also, in the event that the spread occurring in practice deviates from the tolerance orally entered into memory, the central processing unit 100 can generate an alert signal so that the situation is investigated by an operator.

In practice, it is possible that two (or more) different articles are so close together that they are simultaneously weighed by the weighing sensor 30, and / or that they are interpreted by the automatic checkout 1 as a single article, while the corresponding two (or more) bar codes are correctly read by the reader 20 and transferred to the bar code input 102 of the central processing unit 100. To cope with this situation, the central processing unit 100 is preferably arranged to process all the received bar codes in the manner described above; to determine the corresponding expected weight Gx of each individual item; to add the relevant expected weights together to an expected total weight Gxt and to combine the corresponding tolerances into a total tolerance; and to compare the weighted weight Gm to said total weight Gmt in the manner previously described, taking into account the total tolerance.

It has been found that it is generally quite possible to identify an article by means of image recognition, such as, for example, a color histogram, pattern recognition, optical reading of characters (OCR), or a combination of said methods. In the following the example of a color histogram will be further elaborated, but alternatively any of the other methods can always be read. In a similar way as described above with regard to weight, it is then possible to measure the color histogram of an offered item and compare it with an expected color histogram, as a check after a successful identification on the basis of bar code. This check can be performed instead of the weight based check, in which case the weigh sensor 30 can be omitted. The color histogram check can also be performed if the weight check does not lead to an approval, in which case a successful color histogram check can still lead to an approval. Also, it is possible that the weight check and the color histogram check are both performed, and the central processing unit

- 16 only decides to approve if both roles are successful: in that case there is a complete check.

However, in a preferred embodiment of the present invention, the cash register system is adapted to use color histogram-based tification as an alternative identification method when unambiguous barcode-based verification has failed. If the color histogram based verification succeeds, the rale processing unit 100 proceeds to the weight control f, s discussed above, and operator&gt; invocation can be omitted; the assistance of an operator is only called upon if identification based on the histogram also fails.

Preferably, and as also illustrated in Figure _s, the automatic cash register 1 according to the present invention provided for this purpose with color histogram generating means, a signal S is output at an output 41₄o offering that is representative of the color histogram, and the central exploration unit 100 is provided with a third input 104 which is coupled to that signal output 41. In the following said signal S₄q are also referred to as color togram signal, and the information contained therein will also be referred to as color histogram information.

Color histogram generators are known per se. For example, they ate a color CCD camera as well as means of counting how often a particular color appears in the engendered image (ie, how many pixels that particular color is). Therefore, the technique of making a color histogram at h need not be further explained here.

According to the present invention, the color chart generating means 40 comprise at least one, and preferably at least two, image pick-up means (such as, for example, two hour CCD cameras) disposed on either side of the weighing sport belt 14 to examine an article from two or more directions where and the color histogram teaching means 40 are arranged to act as the color histogram signal S.₄o generate 1 combination of the two (or more) individual surhistogram signals from the two (or more) individual

17 image pick-up members. This combination is advantageously a simple addition.

In a possible variant, the cameras supply two full images to the central processing unit 100, and the central processing unit 100 itself is programmed to calculate the color histogram from the offered image signals.

Referring now to Figure 4, the identification based on color histogram will be explained in more detail.

If it is decided at step 312 that the bar code information is incomplete or invalid, the central processing unit 100 checks the color histogram. First, (step 401), the central processing unit 100 receives the image signals S₄o of the image capture units (cameras). These image signals may already be an encoded color histogram, but also an encoded original image recording, in which case the central processing unit 100 determines the color histogram (step 402). Then, the central processing unit 100 consults the memory 120, and compares the color histograms stored therein with the measured color histogram of the presented item (step 403). If a color histogram is found in the memory 120 that sufficiently corresponds to the measured color histogram of the article offered, the central processing unit 100 decides that an identification has taken place with a sufficient degree of certainty, and the central processing unit 100 continues with step 313. If not , the assistance of an operator (500) is sought as the second alternative identification method.

Referring now to Figure 5, an example of an operator protocol 500 that is initiated by the central processing unit 100 will be described if an item identification fails or the measured weight does not correspond to the expected weight.

The basic idea behind the second alternative identification method proposed by the present invention is that human help should be sought if automatic identification fails. Of course it is then

-18 to generate an attention signal to draw the attention of a patrolling cashier, who then goes to the cash register in question to share the situation. However, this takes a relatively long time, and has the drawback that bypassing the use of that checkout stent is perceived by bystanders, which can be perceived as indiscreet by the known customer.

Preferably, therefore, the help of a. remote operator by sending a camera image of the article in question to a monitor monitored by that operator. To this end, the automatic cash register set by the present invention is preferably, and illustrated in Figure 1, provided with an image recording device such as a color camera 50, with an output 51 for the purpose of (step 501) image signal S₅₀ that representa: is for an image included in an article. That output -s coupled to the central processing unit 100, which transmits image signal S50 to an operator station 600 ip 502) if the central processing unit 100 states that automatic identification of the relevant Ikel fails. On the basis of the received image, the rator makes an identification (step 503), and the operator sends necessary data to the central processing unit 100 ap 504). To this end, the central processing unit 100 is provided with a fourth input 106 which is coupled to the rator station 600 to receive the information provided by the operator.

The data sent by the operator to the central processing unit can directly be the price information and possibly a description on the receipt. In that all, the central processing unit 100 can continue at p 317. However, it is also possible for the data sent by the operator r to the central processing unit 100 to be from barcodes, and to be processed by the central processing unit in the same manner as the barcodes captured from the barcode scanners; in that case, the ital processing unit 100 may proceed to step 313.

In principle, separate cameras can be present for this. generate a color histogram on the one hand and for it

·. 1 ;

- providing an image recording for the operator station 600, on the other hand. However, there is no need for this. In a preferred embodiment, the automatic cash register 1 has two cameras arranged on either side of the conveyor belt 11, which take image images of the articles offered, which are used both for determining the color histogram by the central processing unit 100 and for effecting a human identification. through the operator station 600, so that these cameras perform both the function of the means 40 and the function of the means 50.

In principle, it is possible for the image pick-up member 50 to record an image only if the identification has failed. In practice it is more convenient that an image recording is always made of each offered article 2 by the image recording member 50 (ie: two recordings by the two cameras arranged on either side of the conveyor belt 11), and that the central processing unit 100 decides whether or not those recordings are used to calculate a color histogram and / or forward it to the operator station 600.

Operator station 600 may be located remotely from automatic checkout 1 within the same premises, but operator station 600 may be located 25 miles away and be part of a multi-store center, and data communication may be via a communication network, for example a computer network such as the Internet, an intranet, etc.

Operator station 600 may be associated with multiple automatic checkouts, so that a single operator can assist in multi-checkout item detection.

There may also be a pool of multiple operator stations 600, and the image signal is passed to any of the operator stations free at that time.

In a simple embodiment, an operator station is only provided with a monitor on which the perceived article image is projected, and a keyboard or other input device with which the operator can display a price and

-20tually can enter a description, and / or an article. The operator then determines the price himself, for example by consulting the article list.

Figure 6 shows an example of an image display for an operator in a more advanced form of operator station. On the screen, for example by a suitably programmed uter, not only the two image recordings 701 and 702 of the boulder are projected, but also some suggestions for similar identifications. Those suggestions can be projected in the form of a written list 703, and / or the form of a series of example images, which are stored in a grade memory associated with the POS system. In Fig. 6, in the bottom right quadrant, the display 700 is a mosaic 710 with nine example messages 711 through 719.

The order of the suggestions in list 703 and iik 710 preferably corresponds to the probability that the suggestion in question is correct. In calculating that probability, the computer can honor the information obtained from the article in question, even if it is incomplete information. For example, it is possible for a barcode fragment to be recognized, in which case the computer will first generate those articles whose barcodes ate that barcode fragment, always taking into account the measured weights.

for example, it is possible that the computer compares the urhistogram with the color histograms of the images stored in one memory, and bases the probability on the degree of similarity between the color histogram and the retrieved urhistogram.

Not only will it be more obvious for an operator to recognize the displayed article in this way, but it will also be easier for the operator to communicate his identification, the central processing unit 100 by, op. as is well known in computer systems, to move an pointer across the monitor screen 700 using

-21 a tool such as a computer mouse, and clicking on the desired text or image.

In practice, it may happen that a customer forgets to place an article from the shopping cart on the conveyor belt 13, so that this article is not detected and is not settled. To counteract this, a further preferred embodiment of the cash register system according to the present invention is provided with a shopping cart checking system 70 which checks whether the shopping trolleys are completely empty. According to a particular aspect of the present invention, that control takes place on the basis of image processing. As illustrated screenatically in Figure 7, the automatic cash register 1 is provided for this purpose with a further camera 71, which is arranged above a path where the shopping trolleys 72 must pass the automatic cash register 1. A suitable trolley detector 73, which includes, for example, an induction loop, detects the presence of a trolley 72, and generates a trigger signal for the camera 71, which takes a picture of the trolley. That recording is passed on to an image processing unit 74, which evaluates the captured image and compares it with an image of an empty shopping cart previously stored in a memory. In case the comparison yields significant differences indicative of a non-empty shopping cart, the image processing unit 74 generates a warning signal, for example to the operator station 600 already mentioned.

In the foregoing it has been described with reference to Figure 1 that barcode reading means known per se can be arranged near a conveyor belt 11, whereby all surfaces of the article can be viewed by the barcode reading means, except the bottom. In a particularly preferred embodiment, a line scanner 80 is arranged under a gap 81 between two successive conveyor belts, such as the conveyor belts 13 and 14, as illustrated in figure 8. The line scanner is associated with suitable, and appropriately positioned, exposure means, as will be apparent to one skilled in the art

-22n. Due to such an arrangement, it is also possible. the bottom of the transported items collect data.

Line scanners are known devices per se, so that a discussion of their construction and operation is not necessary. Suffice it to note that a line scanner is capable of generating a signal that can be compared to a single horizontal line of a video. statue. The signals generated by the line scanner are processed by 1 signal processing device, which may be associated (or may be identical to said processing unit).

The line scanner provides consecutive line images as an article passes. The successive line images thus respond with a line scan of a ticle, similar to a field generated by a 2D camera. The consecutive line images can therefore be edited using special or standard field processing software, similar to the processing of a normal video image. However, it is also the same to use image processing software that is ecologically intended for processing successive individual line images.

It should be noted in this connection that the line frequency n the line camera, ie the number of line images per second, may be equal to the standard line frequency of another video image, for example 25 Hz, but this is not logical. In a possible embodiment, line image capture is triggered by said signal processing device or said processing unit 100, with a repetition rate determined by that device, which may be higher or lower than the standard. frequency.

With the aid of image processing software, various types of information can be extracted from the line images provided. To begin with, the presence or absence of an article can be determined, so that the line scanner 80 can perform the function of the said

<img file="NL1021040C2_D0004.tif" />

-23 passing detector 23. A separate passing detector is then no longer necessary.

Furthermore, it is possible that a barcode is recognized in the obtained image information. This means, among other things, that the customer no longer has any restrictions when placing the items to be settled: the barcode may now also be pointing downwards. The recognized barcode can be processed by standard barcode translation software to retrieve the string represented by the barcode, and based on this, a database can be used to look up the identity and price of the item in question, in the same manner as described previously. In a possible embodiment, the barcode information is obtained only with the aid of the line camera. This offers the advantage that it is no longer necessary to set up a system of barcode scanners in a tunnel construction, so that the costs of the entire installation can be lower, but is associated with the restriction of use that the customer barcodes the articles. facing down on the belt.

Furthermore, it is possible that a color histogram is derived from the obtained image information, which information can be used in the same manner as described above.

Furthermore, it is possible that characters (letters and numbers) written in the obtained image information are recognized, and that this information is used to recognize the articles.

Furthermore, it is possible that in the obtained image information certain characteristic image parts are recognized, such as for instance a logo of a manufacturer, or a product image, and that this information is used to recognize the articles.

Furthermore, it is possible that shape characteristics of the article are derived from the image information obtained, and that this information is used to recognize the articles. For example, it is possible to derive from the number of lines where a part of the article is displayed, the length of the article, in connection with the line frequency used and the speed of the article, which

-24s are exciting with the speed of the conveyor belts. Furthermore, it is just possible to determine the width of the relevant ikel section per line, and to combine the data of all lines into a contour of the article. All these data are useful in recognizing the articles.

In a manner similar to that described above, it is useful to arrange a line camera adjacent to the conveying path of the articles, so as to obtain line images from one side of the articles, which can be processed in the same manner as described above. It is sufficient to have a single line camera, or two line cameras and on top of each other. This makes it possible to determine a side view contour of the panels, which can be combined with the original view contour into a 3Ditour and / or a volume content, which data can be used in recognizing the articles.

In a further elaboration, several line cameras can be set up to transport the articles, see the articles from the front, rear and / or 1 above respectively. If desired, they can thus completely replace the said fashion scanners.

It will be apparent to a person skilled in the art that the present invention is not limited to the examples discussed hereinafter, but that alternatives, modifications, modifications and variations are possible within the scope of the invention as defined in the appended claims.

In the foregoing, with reference to Figure 3, the run of an identification cycle has been described, and it has been noted that that cycle is completed for each item presented. In fact, it is more correct to say that for every item offered, such a cycle is completed. Each individual cycle starts with receiving the dashed information signal S₂o (step 311) and each individual cluster ends with the transmission of the item information ar the cash register (step 317). The individual cycles do not all have to take the same length of time. Some articles are quickly identified, while others take some time

-25 longer, and sometimes even the help of an operator is needed. In principle it is possible that a new identification cycle only starts after the successful completion of the previous identification cycle, but the consequence may be that the conveyor belt 11 has to be stopped regularly, especially when the assistance of an operator is required. According to the present invention, the transport by means of the conveyor belt takes place continuously, and a new identification cycle can start regardless of whether or not the previous identification cycle has ended. In practice, it will generally be possible for several identification cycles to run in parallel, and the order of recognition may deviate from the order of entries.

In addition, it can be noted in this connection that an identification cycle can also be started if the weight residue is greater than a predetermined threshold. Weight residue refers to the difference between, on the one hand, the measured cluster weight and, on the other hand, the total weight of the items already identified within that cluster.

Furthermore, it is possible for operational parameters of an automatic cash register to be passed on to the operator, so that action can be taken at an early stage if phenomena indicate a possible future failure of a part. An example of such operational parameters is, for example, the diode current of the lasers used in the barcode scanners.

In the preferred embodiment discussed above, the articles are moved along one or more fixedly positioned scanners by means of a conveyor belt which is provided with integrated weighing members. Within the scope of the present invention, a simpler version of the cash register system is possible, where the bar codes are read with a customer-operated hand scanner, and where a separate weighing station is provided for weight control.