EP0850112A1 - Device for reducing the digital image data arising on the scanning of loose items on a conveyor - Google Patents

Abstract

The device contains a conveyor with at least one conveyor belt (3, 7, 11, 15) for loose items (25), the surface of which has a brightness which is as complementary as possible to that of the surfaces of fixed regions (1, 5, 9, 13, 17) visible near it and an electronic scanning device (ZK, L1, L2) extending approximately transversely (X) to the conveying direction (Y) over the width of the conveyor belt and the visible surfaces of the fixed regions for the linear detection of the brightnesses (I(x,y)). There is also a program-controlled computer (PC) which stores a sequence of the brightnesses (I(x)) occurring over the width (X) of the conveyor belt (3,...) and visible surfaces of the fixed regions (1,...) in the absence of a loose item (25) in a first store (SP1) and places current brightnesses (I(x,y)*) in a second store (SP2) as long as differences occur on the comparison with the corresponding values or values from the stored sequence (I(x)).

Description

description

Device for reducing the digital image data obtained when scanning piece goods on a conveyor

The surfaces of piece goods of all kinds and in particular of their packaging are generally provided with texts, graphic elements and machine-readable codes. Texts and codes are particularly important here, e.g. contain the contents of a package, the type of goods and, if applicable, a destination for the transport of a piece goods.

Such piece goods can occur locally in very large and inhomogeneous quantities, e.g. for forwarding agents, mail order companies, mail and parcel distribution systems, but also for the acceptance and distribution of parts in large-scale production plants, e.g. in an assembly plant for motor vehicles. As a rule, fast-moving conveyors such as e.g. Treadmills. There is an increasing desire to capture at least parts of the above-mentioned inscriptions mechanically and to trigger actions relating to the respective part fully automatically.

For this purpose, the surfaces of the piece goods transported, for example on a conveyor belt, are scanned line by line with respect to the brightness intensity values, for example with a line camera mounted stationary above the conveyor belt. The resulting binary image data are then evaluated in a computer. The aim is to automatically decipher the meaning of the labels contained in the total amount of image data, ie in particular of texts. For example, texts contained in the image data can be encoded and thus machine-readable using a so-called OCR program (optical character recognition) be transformed. After their evaluation, it is then possible to carry out, particularly automatically, control-technical actions relating to the individual piece goods in a fully automated manner.

For example, in the case of letter sorting systems, the addresses and, in particular, the so-called postcodes (postcode, zip code) are recorded by scanning the tops of the letters in order to achieve the most fully possible sorting of the letter items according to regions, districts or cities and towns.

A problem with such a scanning of the surfaces of piece goods transported via conveyor belts is, however, that extraordinarily large amounts of binary image data are obtained and are fed to a computer for processing. This is due, on the one hand, to the fact that, in particular because of a possible OCR post-processing of the image data, the image data must have a sufficient resolution despite the generally highest possible conveying speed. The line scanning frequency of a line camera scanning the surface of a conveyor belt transporting the piece goods must therefore be quite high.

With e.g. In practice, letter sorting systems can have the following conditions. Assuming that five letters are conveyed under the line scan camera per second, a fully continuous, i.e. uninterrupted scanning of five recording fields for five letters on the top of the conveyor belt, approx. 30 Mbytes (megabytes) of binary image data per second. But for one thing

In hardly any case is the image field completely filled out by a letter in it, and secondly the size of letters can vary greatly, only a part of the pixels from the image data volume corresponds to one segment or pixel on the letter surface per image field, while whereas the other part only represents the surface of the conveyor belt and thus contains no evaluable information.

The invention is therefore based on the object of specifying a device which enables a reduction in the binary image data occurring in the case of optical scanning of the surfaces of piece goods transported via conveyor belts in the manner described above in such a way that, if possible, only those image data are combined Computers are fed for further processing, which are on top of each

Piece goods correspond to lying surface segments. The rest of the image data, which virtually only represents the background, should be able to be discarded beforehand. With the device according to the invention, electronic detection of the edges of the optical object to be scanned should therefore be possible so that a targeted selection can only be made of the image data representing the surface of the piece goods.

The object is achieved with the device specified in claim 1. Advantageous further embodiments of the same are specified in the subclaims.

The invention has the advantage that the amount that can be effectively passed on to a computer for further processing, i.e. binary image data containing evaluable text and image contents can be considerably reduced in comparison to the total amount of image data obtained with optical scans of the type described above. Taking the above-mentioned letter sorting systems as an example, this has the consequence that the original amount of image data can be reduced from approximately 30 Mbytes per second to significantly less than 15 Mbytes per second.

The invention is further explained with the examples shown in the figures briefly listed below. It shows Fig.l. 1 shows a plan view of a section of a conveying device, for example for letters, the brightness intensity values occurring being captured by an electronic line camera,

Fig. 2. the course of the intensity values occurring across the width of the conveying device in the event that no piece goods lie on the surface of the conveying device,

Fig. 3. a block diagram representation of the detection and processing of the resulting brightness intensity values according to the invention, and

Fig. 4. the course of the intensity values occurring across the width of the conveyor in the event that a piece goods lies on the surface of the conveyor.

The invention is explained in more detail using the example of a scanning of a letter surface shown in FIGS. 1 to 4. Of course, the invention can also be used for different types of piece goods of various types.

According to the plan view in FIG. 1, the device according to the invention contains a conveyor device with at least one conveyor belt for piece goods. Its surface has a brightness that is as complementary as possible to the brightness of the adjacent surfaces of the stationary areas of the conveyor.

The conveying device advantageously has a multiplicity of strip-shaped conveyor belts which are spaced next to one another in such a way that the surfaces of the stationary regions of the transport device are visible in between. Such an embodiment is already shown in the figures provides, for example four strip-shaped conveyor belts 3, 7, 11, 15 with an exemplary darker surface being guided along a stationary base of the transport device which has a complementary, exemplary lighter surface. Advantageously, the widths of the strip-shaped, adjacent conveyor belts 3, 7, 11, 15 and the surfaces of the fixed regions 1, 5, 9, 13, 17 visible next to and in between are approximately the same size.

The conveyor belts move in the Y spatial direction shown by an arrow on the left side of the image. On the surfaces of the conveyor belts 3, 7, 11, 15 shown in the example in FIG. 1, two rows 19 and 21 of upright driving wedges are additionally attached. The space in between represents a quasi receiving area for a piece goods. In the example of FIG. 1, a letter 25 lies in the receiving area delimited by the driving wedges and is in Y with a longitudinal edge 31 lying against the quasi rear row 21 of the driving wedges Direction transported.

The device according to the invention serves to reduce the amount of brightness intensity values which are obtained by scanning the surface of a conveying device which is used to transport, in particular, flat piece goods. For this purpose, an electronic scanning device for the line-shaped detection of brightness intensity values I (x, y) extends approximately transversely to the conveying direction Y over the entire width of the four conveyor belts 3, 7, 11, 15 in FIG. 1, ie in the X-spatial direction , and the adjacent and intermediate visible surfaces of the fixed areas 1,5,9,13,17. According to the illustration in FIG. 3, the scanning device preferably contains an electronic line camera ZK with two laterally arranged, preferably rod-shaped lighting units L 1, L 2. The line camera ZK captures those in a so-called scan line 23 current brightness intensity values or intensity gray values I (x, y) lying in the X direction and feeds them to a program-controlled computing device PC for further processing.

In the example shown in FIGS. 1 to 4, the device according to the invention has the concrete task of reducing the digital image data obtained by scanning the area lying between the two rows of driving wedges 19, 21 in such a way that only those between the front and rear longitudinal edge 29.31 and the two side edges 33.35 of the letter 25 lying digital image data, ie the brightness intensity values representing the letter surface are temporarily stored for the purpose of further processing.

For this purpose, according to the invention, the program-controlled arithmetic unit PC stores in a first memory a sequence of the brightness intensity values occurring over the width of the conveyor belt and visible surfaces of the fixed areas in the absence of a piece goods, and stores current brightness intensity values in a second memory, as long as the comparison with the corresponding values or with value groups from the saved sequence deviations occur.

In the example in FIG. 2, such a sequence is the brightness intensity values I (x) that occur over the width X of the conveyor belts 3, 7, 11, 15 and the visible surfaces 1, 5, 9, 13, 17 of the fixed areas in the absence of the piece goods 25. in the form of a coherent curve. The visible, in the example lighter, surfaces of the fixed areas are depicted as intensity hills, and the conveyor belts with the example darker surfaces than those in between are depicted as intensity valleys. According to an advantageous development of the invention, the program-controlled computing unit combines groups of comparable, adjacent brightness intensity values from the sequence of the brightness intensity values that occur in the absence of a piece goods by a kind of averaging to reference intensity values and stores them in the first memory. In the example in FIG. 1, the brightness intensity values located in the black marked areas of the scanning gap 23 are each viewed as a group and are combined to form a reference intensity value by a type of averaging. In this way, the reference intensity values 13, 17, 111, 115 can be assigned to the surfaces of the conveyor belts 3, 7, 11, 15, and the reference intensity values 11, 15, 19, 113, 117 can be assigned to the surfaces of the fixed areas 1, 5, 9, 13, 17 that are visible next to and in between. In the course of the intensity values according to FIG. 2 that occur across the width of the conveyor device, these reference intensity values II... 117 are likewise entered lying approximately in the middle of the respective intensity range.

According to the invention, the program-controlled computing unit stores the current brightness intensity values obtained in a second memory when the surface of the conveyor device used to transport piece goods is scanned only as long as it is compared with the corresponding values or with value groups deviations occur in the sequence stored in the first memory or in the reference intensity values. As a result, the amount of digital image data to be stored for the purpose of further processing can be considerably reduced. In particular, it can be used to separate out image data which only reproduce the pure surface of the conveyor, ie the surfaces of the conveyor belts and the fixed areas. In comparison to the acquisition of the surfaces of general cargo that are of interest, these data represent the background. This is explained in more detail using the example of FIGS. 3 and 4 in particular.

The oval section in the left area of FIG. 3 shows an example of the case in which the piece goods 25, in the example a letter, have been pushed by the conveyor belts and in particular by their row 21 of driving wedges into the effective range of the scanning gap 23 of the line camera ZK . Due to the generally random position of the letter on the conveyor, beginning with the front edge 29 of the letter, the conveyor belts 7, 11 underneath and the visible surfaces 5, 9, 13 of the stationary areas of the conveyor are covered by its surface. Under the exemplary assumption that the top of the letter has an even greater brightness than the surfaces of the visible stationary areas, the scanning device per cell position detects brightness intensity values which represent the course I (x ) exhibit.

According to the invention, the program-controlled arithmetic unit PC now compares these values I (x, y) with the corresponding values or groups of values from the sequence stored in the first memory or the stored reference intensity values II ... 117. According to the invention, only those current brightness intensity values that differ from the values stored in the first memory are stored in a second memory I (x, y) *. In the example in FIG. 4, therefore, only the values lying inside the area designated B deviate from the values or reference intensity values 15... 113 shown in dotted lines and corresponding to the curve in FIG prevails. Only these quasi-inner values are stored, while the value ranges with the reference intensity values 11, 13 and 115, 117 are discarded. This way, a relatively accurate detection of all edges is possible 29,31,33,35 of the piece good 25 possible. Only the digital image data lying within these margins are kept in stock in the second memory. They can, for example, be subjected to further processing with the aid of so-called OCR algorithms (optical character recognition), for example in areas 27 on the general cargo or letter top (so-called AOIs: areas of interest), which usually contain information printed on them , especially to identify text information such as address details.

This detection and processing of the resulting brightness intensity values according to the invention is finally explained in more detail with reference to the block diagram representation of FIG. 3.

The sequences of brightness intensity values I (x) or locally assigned reference intensity values II... 117, which occur in the absence of a piece goods, are inserted into the first memory SPI as a data stream 37 in the manner of an initialization or adjustment process by the PC and are available there as comparative values. Brightness intensity values I (x, y) occurring in normal operation are redirected by the program-controlled computing device PC at the branching point 36 and fed as data stream 38 to a comparison device which is preferably software-based and is shown in FIG. 3 in the form of an oval in block PC. If there are deviations in the comparison with the value sequences or reference intensity values taken from the first memory SPI as data stream 39 and corresponding spatially in the X direction, i.e. applies e.g. I (x) ≠ II ... 115, these digital image data l (x, y) * represent a section of the general cargo surface and are written into the second memory SP2 as data stream 41.

From there, as data stream 43, they can carry out further filtering processing, in particular carried out in software the program-controlled computing unit PC are supplied. In the example in FIG. 3, the areas on the top of the letter which contain text information, that is to say represent so-called AOIs (areas of interest), can be detected, for example with the aid of a software converter. The digital image data are therefore one with the equation I (x, y) * = AOI? undergo writable filtering. If this query turns out to be positive, then this data can continue to be subjected, for example, to a fully automatic character recognition routine, ie OCR {I (x, y) *}. The resulting data stream 45 now contains only machine-readable codes and can, for example, be automatically further processed for a wide variety of control purposes.

Claims

claims 1. Device for reducing the amount of digital images which are obtained by scanning (23) a surface of a conveyor device used for transporting, in particular, flat piece goods (25), wherein a) the conveying device has at least one conveyor belt (3, 7, 11, 15) for piece goods (25), the surface of which has areas (1, 5, 9, 13, 17) that are stationary with respect to the brightness of the adjacent surfaces has complementary brightness, and with b) an electronic scanning device which extends approximately transversely (X) to the conveying direction (Y) over the width of the conveyor belt (3, 7, 11, 15) and the visible surfaces of the fixed areas (1, 5, 9, 13, 17) ZK, L1, L2) for the line-shaped detection of the brightness intensity values (I (x, y)), and with c) a program-controlled computing unit (PC), which cl) in a first memory (SPI) a sequence of the Hel¬ occurring over the width (X) of the conveyor belt (3 ...) and visible surfaces of the fixed areas (1 ...) in the absence of a piece goods (25) intensity intensity values (I (x)), and c2) current brightness intensity values are stored in a second memory (SP2) (I (x, y) *), as long as deviations occur when comparing with the corresponding values or with value groups from the stored sequence (I (x)). 2. Device according to claim 1, wherein the program-controlled computing unit (PC) a) from the sequence of the brightness intensity values (I (x)) which occur in the absence of a piece good (25), groups together comparable, adjacent brightness intensity values by a type of averaging to give a reference intensity value (II ... 117), b) the reference intensity values (II ... 117) are stored in the first memory (SPD, and c) compares current brightness intensity values (I (x, y)) with the stored sequence of reference intensity values (II ... 117). 3. The device according to claim 2, wherein the program-controlled arithmetic unit (PC) from the sequence of the brightness intensity values (I (x)) occurring in the absence of a piece (25), the at least one conveyor belt (3, 7, 11, 15) and Visible surfaces of the fixed areas (1,5,9,13,17) assigned to groups and from this each forms a reference intensity value (II ... 117). 4. Device according to one of the preceding claims, wherein the conveyor device has a multiplicity of strip-shaped conveyor belts (3, 7, 11, 15) that are spaced apart from one another in such a way that the surfaces of the stationary regions (1, 5, 9, 13, 17) of the transport device are visible. 5. The device according to claim 4, wherein the widths of the strip-shaped, adjacent conveyor belts (3, 7, 11, 15) and the surfaces of the fixed areas (1, 5, 9, 13, 17) are approximately the same size.