Remote crane bar code system

[0001] This application claims priority from U.S. patent application Ser. No. 60/601,183 filed on Aug. 13, 2004.

[0002] The present invention relates to a method and apparatus for remotely reading an identifier on an object.

[0003] Items produced in a manufacturing environment will typically be stored in a warehouse for shipping at a later date. A shipping warehouse will typically house a plurality of products, which are made by differing processes or have different characteristics. This collection of warehouse items may be referred to as the ‘field’. The field can be substantially large and therefore may be organized into a set of defined locations resembling a ‘grid’. The warehouse items are placed at appropriate locations within the grid and these locations are recorded, creating a mapping of the items for subsequent rearrangement or retrieval. A shipping order will typically comprise a combination of dissimilar items from this field requiring this combination of items to be located and collected to complete the shipping order. This shipping order is sometimes referred to as a shipping manifest or lift ticket. Further to gathering items for a shipping order, it may be necessary or beneficial to rearrange or move around the items in the field to optimize floor space or to enable a more efficient arrangement of the items.

[0004] When the items manufactured are of substantial dimension and weight, it is typically necessary to retrieve the items from the field using an overhead crane or similar device capable of lifting and transporting items of such dimension and weight. The use of an overhead crane requires the operator of the crane to either be placed at a remote location relative to the field (in the cab of a crane for example) or to operate the crane with a remote control device at field level.

[0005] When the operator is at a remote distance, the operator may be unable to distinguish between items in the field that are required for a given shipping order. This situation is of particular concern where items are of similar shape but different characteristics, such as in the steel industry where coils of stock that are produced with differing specifications appear similar, especially when viewed from a distance. If the operator uses a remote control device to operate the crane, navigating the field while moving the crane, and reading and scanning the items becomes quite cumbersome for one person. Furthermore, the use of one person at the field level to control the crane, identify the items of interest and scan the item becomes cumbersome due the need for multiple devices to both control the crane and scan the item.

[0006] If the crane operator is remotely located relative to the field, a second individual is required to identify the existence and position of the desired items at the field level, to scan the desired items and communicate this information to the operator. The communication between the two individuals is required to identify the item of interest for rearrangement or shipping purposes.

[0007] The use of two individuals to gather items in a shipping order tends to be both inefficient and labour intensive given the task to be completed. In the steel industry where the items in the field are of substantial size and weight, the individual assigned to track the appropriate items at the field level would find the method of scanning to be not only time consuming but also dangerous. The inadvertent movement of large items on the field poses a threat to the safety of the individual at the field level and the large area of the field does not lend itself to an efficient method for identifying the desired items in the shipping order.

[0008] In the steel industry where the items in the field are large coils, typically the individual at the field level manually scans a barcode found on a tag affixed to the coil. This introduces a possibility for human error. The human error can lead to the processing of incorrect coils, which could possibly generate an incorrect shipment to the customer. Further to the time-related inefficiencies and inherent safety risk, the use of a field level individual requires additional floor space for the above-mentioned navigation of the field. By eliminating the use of a floor operator, less floor space would be required. This is due to a reduction in the required size of the lane ways between adjacent coils. Space is then only required to accommodate the jaws of the crane's picker. This requires an apparatus capable of viewing the field from a distance.

[0009] To remotely view labels and barcodes, it has been known to use a camera mounted in a fixed position whereby movement of an item into the field-of-view of the camera allows for remote viewing of a label. This method however requires the position of the labels to be known and the correct item to have been picked by the crane in advance of the camera scan.

[0010] Another method of reading labels and barcodes remotely involves a moveable camera capable of tilting, panning and zooming to focus on a desired label or barcode. This method however, requires additional operations to be manually executed by the operator of the crane to identify not only the item of interest but also to correctly centre and zoom in on the label for reading. These additional operator interactions impose an additional opportunity for human error.

[0011] It is therefore an object of the present invention to provide a method and apparatus to obviate or mitigate the above disadvantages.

[0012] In general terms, one aspect of the present invention provides a method for remotely scanning objects including the steps of using an imaging system to display an image of the objects on an interface, receiving a location input related to an identification tag which is attached to a desired object based on a location in the image, using the location input to orient the imaging system towards the identification tag, magnifying the identification tag, and reading information identifying characteristics of the desired object provided by the identification tag.

[0013] In another aspect, the present invention provides a system for remotely scanning objects comprising an imaging system positioned remotely from the objects and arranged to image the objects. The imaging system has an adjustable lens for magnifying the image. The system also comprises an interface for displaying an image of the objects and is adapted for receiving a location input for an identification tag attached to a desired object based on a location in the image. The system also comprises a processor connected to the imaging system and the interface. The processor uses the location input to orient the imaging system towards the tag, commands the adjustable lens to magnify the tag, and reads information identifying characteristics of the desired object provided by the tag.

[0014] In yet another aspect, the present invention provides a method for aligning a tag in an image, the tag being affixed to an object and having indicia thereon. The method has the steps of obtaining an image of the object having at least a portion of the tag visible in the image; arranging at least one sensor on said image; identifying at least one marking on the tag using the at least one sensor, the at least one marking indicative of the position of the tag in the image, and indicative of a respective region of interest in the image corresponding to at least a portion of indicia on the tag, each of the respective regions of interest intersecting one of the at least one sensor; computing an average position of the at least one marking to determine a deviation of the average position from a preferred position; and aligning the tag in the image according to the deviation.

[0028] Referring therefore to FIG. 1, an overhead crane system 10 is positioned above a field of inventory 20, the inventory in this embodiment being coils 22 of steel varying in specification. The coils 22 are initially placed in the field 20 and the respective positions of the coils 22 in the field 20 recorded using a range finder 13 or other means. Each position may then be correlated to its respective coil 22 using the system 10 or other suitable methods. The correlation of position to coil 22 enables an operator of the system 10 to at a later time target a particular area of the field in order to locate and scan the coil 22 to determine if it remains at its recorded position.

[0029] The overhead crane system 10 includes a trolley 12 mounted upon a bridge 26 and has a communication connection to an operator cab 18, preferably over a festoon cable to accommodate movement of the trolley 12 relative to the cab 18. The cab 18 is situated in a fixed position at one end of the bridge 26. An inventory control system 24 includes coordinates of objects and also has a communication connection with the operator cab 18. The trolley 12 includes a set of motors 28 to facilitate translation of the trolley 12 along the bridge 26. Typically the bridge 26 is mounted on rails 25 transverse to the bridge 26 allowing the bridge 26 to translate fore and aft along the rails 25.

[0030] Translation of the bridge 26 and the trolley 12 in the directions indicated allows the trolley 12 to access objects located anywhere in the field 20. The trolley 12 furthermore includes a picker 16 for vertically hoisting coils 22 from the field 20, a camera system 14, and the range finder system 13 having separate range finders for locating the trolley's position along each axis of the field 20.

[0031] The camera system 14 can be seen in greater detail when referring to FIG. 2. The camera system's components are housed within a casing 36 and this casing 36 is mounted to the underside of the trolley 12. A zoom lens 32 of a camera 34 protrudes beyond the lower surface of the casing 36, which is partially open and covered by a transparent acrylic enclosure 30. The camera 34 is preferably a “smart” camera, which is a camera having a microprocessor capable of processing image data. This functionality enables the camera 34 to process information related to the coils 22, that are acquired in an image.

[0032] The processing may also be done remotely from the camera 34 in a separate processor. The acrylic enclosure 30 allows movement of the zoom lens 32 within its volume and is transparent, allowing the lens 32 to capture images. The camera 34 is controlled by a pan/tilt mechanism 40. The pan/tilt mechanism 40 can orient the camera 34 using various pan and tilt operations in order to point the camera 34 towards a desired area of the field 20. A motor 38 is incorporated within the pan/tilt mechanism 40 and controls its movements. The motor 38 is controlled by an electronic controller 33 which has a communication connection to the smart camera 34 or other system control computer (not shown).

[0033] The interface located within the operator cab 18 is shown in FIG. 3. The cab 18 contains a computer interface 50 which includes a touchscreen 54. A control console 52 allows the operator to control manually, the movements of the trolley 12.

[0034] Making reference now to FIG. 4, the touchscreen 54 displays the images acquired by the camera system 14. These images show objects in the field 20 and in this particular example are coils of steel 22. The coils 22 are of differing specifications, and information pertaining to the coil 22 is stored on a tag 60. The tags 60 are intended to be affixed to the upward facing surfaces of the coils 22 typically in an unspecified manner and therefore do not appear at consistent locations on the upward facing surfaces of the coils 22 or in consistent orientations thereon. The information found on the tag 60 is unreadable from the distance that the operator is located and therefore must be magnified by the camera system 14. A tag 60 is shown in FIG. 5a. The tag 60 includes a barcode 64, a numerical code 66 and a set of alignment markers 62. An alignment marker 62 is located in the proximity of each of the four corners of the barcode 64. One alignment marker 62a is dissimilar to the other alignment markers 62b, 62c, 62d. The dissimilar alignment marker 62a is used by the camera system 14 to determine the orientation of the tag 60 in the image. The orientation of the tag 60 allows the camera system 14 to choose the appropriate direction to perform the barcode scan.

[0035] In FIG. 5a, the dissimilar marker 62a is located in the top-left portion of the image with respect to the other markers 62b, 62c, 62d. The dissimilar marker 62a includes a triangular notch which points towards the centre of the barcode 64. The remaining three markers are triangular in shape and are rotated 90° with respect to each other such that they each point towards the centre of the barcode 64. The alignment markers 62 are located at substantially equal distances from the centre of the barcode 64. These distances are known proportions of the tag's size (for instance a proportion of the width). These proportions and the tag size itself are programmed into the camera system 14. The camera system 14 can use the width of the tag 60 seen in the image to establish scale. Distances can be measured from the alignment markers 62 based on the established scale, the known proportions and the resolution of the camera system 14. The barcode 64 and the numerical code 66 contains identification information pertaining to the coil 22 to which the tag 60 is affixed.

[0036] The communication connections are schematically shown in FIG. 6. The electronic controller 33 includes a zoom controller 82 operating the zoom lens 32 and a pan/tilt controller 84 operating the pan/tilt mechanism 40. The controller 82 commands the motors 38 (not shown) facilitating the movement of the zoom lens 32 (or 32b in a two camera system—explained later). The controller 84 commands the motors 38 facilitating the movement of the pan/tilt mechanism 40. In this particular embodiment, the inventory control system 24 is connected to the operator interface 50 via a wireless Ethernet link 80. It will be appreciated that any of the communication connections described herein may be hard wired or wireless. It will also be appreciated that the touchscreen 50 and operator interface may alternatively be located away from the crane at a remote location, and operated via the communication link 80. In such an arrangement, control of the crane and the picker 16 can be performed from any location.

[0037] Referring to FIG. 7, an automatic scanning process 100 involves a continuous scan of the coil field 102. Referring also to FIG. 1, the camera system 14 is mounted on the underside of the trolley 12 and therefore scans the field 20 below as the operator navigates the trolley 12. Images captured are displayed to the operator 104 as shown in FIG. 4. Coils 22 are observed during this scanning process 100 and the operator must decide whether the coil 22 shown is of interest for reading 106. If the coil 22 is not of interest to the operator, the operator will continue to monitor the image 104 until a coil 22 does appear that is of interest for reading. When a coil 22 appears that is of interest, the operator first indicates whether the coil 22 is situated at a relative far position such as on the floor or at a relative near position such as being mounted in a secured and elevated position on a truck bed. This is done by selecting a “Near” setting or “Far” setting on the touchscreen 54. The settings represent the nominal magnifications required by the camera system 14 to be able to read a tag 60 at the corresponding distance. It will be appreciated that there may be any number of magnification levels that can be chosen and should not be limited to only “Near” and “Far” settings. The operator then selects the coil 108 by touching the image of the particular coil 22 at the position which its tag 60 appears on the touchscreen 54.

[0038] It will be appreciated that the camera system 14 may also use the range finder system 13 to determine where the trolley 12 is in the building and whether it is over a floor area or a loading bay (truck mounted coils) to automatically adjust the magnification and focus to appropriate settings without operator input.

[0039] At this point, the camera system 14 begins an identification process 109. To begin, the camera system 14 is given a set of co-ordinates from the touchscreen 54 representing the position selected by the operator. These co-ordinates are measured relative to a datum wherein the scale of the image is known based on the wide view magnification used by the camera system 14 and the data provided by the range finder system 13. The datum represents the centre of the field-of-view of the camera system 14. The pan/tilt controller 84 then moves the camera system 14 aligning the datum with the given co-ordinates 110 which places the tag 60 substantially within the centre of the field-of-view of the camera system 14. The camera system 14 also uses the data from the range finder system 13 to map the trolley's position within the field 20 to the given co-ordinates. This provides the inventory control system 24 with a floor grid location to be associated with the tag's information.

[0040] This first movement 110 by the pan/tilt mechanism 40 provides a coarse adjustment for centring the tag 60. Following this pan/tilt operation 110, the camera system 14 commands the zoom controller 82 to perform a zoom operation 112, providing an enlarged image of the tag 60. The zoom controller 82 has two predetermined magnifications, one for the “Near” option and one for the “Far” option. Since the tags 60 are presumably affixed to the coils 22 on the upward facing surface, tags 60 with similar designation (specifically “Near” or “Far”) will be at a substantially similar distance from the camera system 14. If the operator had selected “Far”, the zoom controller 82 magnifies the image to its “Far” setting. If the operator had selected “Near”, the zoom controller 82 magnifies the image to its “Near” setting which requires less magnification than the “Far” setting since the coils 22 are positioned closer to the camera system 14. Due to curvature of the upward facing surface of the coils 22, tags 60 of similar designation may be affixed at slightly varying distances. The zoom controller 82 performs minor focusing at this point if necessary to provide adequate sharpness of the image.

[0041] It will be appreciated that the camera system 14 may also use a depth measurement device such as an ultrasonic range finder to determine the distance between the tag 60 and the camera system 14. This would allow the zoom controller 82 to choose specific magnifications for each tag 60. This may be necessary in situations where the dimensions of the objects being selected vary substantially.

[0042] Following the zoom operation 112, the camera system 14 performs an alignment adjustment operation 114. Referring now to FIG. 5b, the camera system 14 analyses the image and identifies the location and orientation of each of the alignment markers 62 on the tag 60 using an object-finding routine built into the software used by the imaging system, e.g. smart camera software, and previously programmed to identify markers 62 having a particular size and shape.

[0043] The camera system 14 determines the position of the dissimilar marker 62a relative to the other markers and this position dictates the relative orientation of the tag 60 and subsequently the barcode scan direction. If the dissimilar marker 62a is the upper-leftmost of the markers 62 (as shown in FIG. 5a and 5b) the camera system 14 determines that a left-right horizontal scan is required. If the dissimilar marker 62a is the upper-rightmost of the markers 62 the camera system 14 determines that a top-bottom vertical scan is required. If the dissimilar marker 62a is the lower-leftmost of the markers 62 the camera system 14 determines that a bottom-top vertical scan is required. If the dissimilar marker 62a is the lower-rightmost of the markers 62 the camera system 14 determines that a right-left horizontal scan is required.

[0044] Using the locations of the markers 62, the camera system 14 then approximates the centre of the barcode 64. Firstly, since the relative orientation of the tag 60 has been determined, the camera system can measure the width of the tag 60 along the appropriate direction in the image 70. Furthermore, since the actual width of the tag 60 and the camera system's resolution is known, the camera system 14 can correlate pixel width in the image to the actual width on the tag 60. Each marker is a particular distance from the centre of the barcode 64 and is a proportion of the tag's width. The distance is measured along a line in the direction that the marker 62b is pointing and is typically perpendicular to the outermost edge of the marker 62b relative to the barcode 64. Based on the proportion of the tag's width, the actual distance on the tag 60 is converted to a number of pixels in the image. This pixel length is then converted to a set of pixel co-ordinates relative to the marker 62b. Using these relative pixel co-ordinates, the centre of the barcode 64 is approximated and a mark 74 is recorded by the camera system 14. This process is repeated for the other three alignment marks 62a,c,d and the average position 72 of the four marks 74 is calculated and its position is recorded by the camera system 14. These markings are shown in FIG. 5b.

[0045] The camera system 14 uses the position of the average centre mark 72 to determine whether the centre mark 72 lies within a window 76 of acceptable positions surrounding the centre of the image 70. If the average centre mark 72 is within the acceptable window 76, the barcode 64 can be read. If the average centre mark 72 is not within this window 76, the pan/tilt controller 84 commands the pan/tilt mechanism 40 to adjust the camera system 14 thereby placing the average centre mark 72 within the acceptable window 76 of the analysed image 70. This alignment of the average centre mark 72 ensures the entire barcode 64 is visible in the image 70 and therefore can be properly scanned.

[0046] With the tag 60 magnified 112, properly aligned (per step 114), and its orientation known, a barcode string is generated by the camera system 14 by scanning the bar code 116. The direction of the scan is based on the determined orientation of the tag 60. This barcode string is sent to the operator interface 50 for comparison with the lift ticket 118. If the information acquired does not match an item on the lift ticket, the coil 22 is rejected and the system 100 returns to the field level image for the operator to make another selection. If the barcode 64 does match an item on the lift ticket, the camera system 14 returns to a wider view to allow the coil 22 to be grabbed and lifted by the operator 119 using the crane's picker 16. The automatic scanning process 100 is reinitialised 120 once a coil has been lifted 119 and resumes scanning the coil field 102 until the next operator selection. The system 10 may then interface with the inventory control system 24 to update the stock of coils 22 and process a shipping ticket for delivery of an order of coils 22.

[0047] Therefore, the system 10 enables the identification, scanning and retrieval of objects in a field of inventory from a remote location requiring only a single input from an operator. The operator may remotely scan a collection of the objects and select an object of interest based on a predetermined location for that object. This can be done through an input such as touching the image on a touchscreen to indicate the location of an identifier on the object. The imaging system 14 may then automatically magnify the identifier based on the input, and automatically perform an alignment procedure to orient the identifier according to a desired orientation. The system 14 then automatically reads the identifier, e.g. by scanning a barcode 64, and uses information provided by the identifier to confirm the location of the object for processing shipping orders, and update an inventory system 24 accordingly. Only a single operator input using a touch or point of a mouse is needed to execute the above procedure. This effectively replaces a manual pan/tilt/focus/zoom operation with a single initial input.

[0048] In a further embodiment of the present invention, the camera system 14 utilises two smart cameras 32a, 32b as shown in FIG. 8. The pair of cameras 32a, 32b are mounted together on the pan/tilt mechanism 40 similar to the apparatus shown in FIG. 2. The first camera 32a is at a fixed magnification and provides a constant overall image of the coils 22 as they are being scanned. The second camera 32b is equipped with a motorised zoom lens similar to the camera lens 32 in the previous embodiment. In this configuration, the second camera 32b maintains a magnification close to the level at which a tag 60 can be read and requires only minor magnification adjustments once the pan/tilt mechanism 40 aligns the second camera 32b with the selected tag 60.

[0049] The use of two smart cameras 32a, 32b eliminates the delay time caused by the long zoom stroke being required to increase the magnification from a wide view of the field 20 to a zoomed view of a barcode 64. While the camera system 14 scans the field 20, the touchscreen 54 displays an image of the field from the fixed camera 32a. When the operator selects a tag 60 on the touchscreen 54, the touchscreen 54 then displays an image from the second camera 32b while it centres the tag 60. Since the tags 60 may be affixed at varying distances, the second camera 32b will make necessary minor adjustments to achieve the desired magnification while centering takes place. Both cameras 32a, 32b are mounted on the pan/tilt mechanism 40, and thus move together to maintain a constant relationship of the location of the tag view within the field of view of the fixed camera 32b.

[0050] During operation, one camera (e.g. 32a) is designated as a field camera, and the other camera (i.e. 32b) is used at the tag camera for reading barcodes. The field camera 32a has a fixed focal length, aperture and focus settings. The image size, and depth of field are set so that all coils 22, no matter what height, are in focus. The overview image is provided to the operator, so that they can select the location (i.e. barcode tag) to enable the tag camera 32b to locate the tag 60 for reading the barcode 64. The field camera 32a monitors the output of the user interface touchscreen 50, looking for tag identification “touches” or other suitable commands to indicate such identification.

[0051] Once the barcode 64 has been identified by the operator, the camera 32a attempts to identify the barcode 64 and locate its center, to thereby increase the accuracy of the pointing instruction to the pan/tilt mechanism 40. If the attempt fails, the pan/tilt command defaults to the exact position that the operator touched. Once the pointing operation is complete, the field camera 32a flags the tag camera 32b to begin the tag reading process.

[0052] The tag camera 32b has a motorized zoom lens, which is capable of adjusting image size, aperture (brightness and depth of field), and focus (object height). Image size is set by the operator, who may specify whether the coil is on the floor or on a truck bed as explained above. The aperture is held constant, and focus may be scanned to optimize image sharpness for the barcode read.

[0053] The tag image may be provided to the operator for manual centering using the touchscreen 50, or to be able to read the tag number in case the barcode is unreadable. The tag camera 32b operates to execute the identification process 109 described above. It will be appreciated that the camera 32b may process the image with an internal processor or may send images to an off-camera processor for processing.

[0054] It will be appreciated that the second embodiment described herein includes all of the features of the previous embodiment with an increased zoom speed imparted by use of a pair of smart cameras 32a, 32b shown in FIG. 8, and described above.

[0055] The identification process 109, particularly the alignment step 114 described above is most accurate when reading tags 60 that are affixed to objects have a substantially planar upwardly facing surface, or when the tags 60 are more or less ensured to be affixed such that their alignment is substantially parallel to the floor 20. When tags 60 are affixed to rolls of steel 22, the inherent curvature of the upward facing surface of the roll often places the tag 60 at a difficult angle for viewing the alignment markers 62 described above, e.g. when the tags are positioned on a sloping surface of the roll 22.

[0056] An alternative procedure for aligning a tag 160 is shown in FIGS. 9a-9d and 10-11, which is most suitable for centering tags 160 that are likely to be affixed to an object having a sloping surface. In this embodiment, like elements are given like numerals with the prefix “1”.

[0057] An image 154 may be obtained according to steps 102-112 shown in FIG. 7, using either the one-camera or two-camera system. The following description is directed towards a two-camera system, but should hold true for a single camera system with different zoom levels, since different zoom levels are inherently at different resolutions. In a two-camera system, when the field camera 32a sends instructions to the pan/tilt mechanism 40 to center the tag camera 32b on a barcode, it is common for the tag 160 to be off-center in the tag camera's field of view. This occurs because the resolution of the tag camera 32b is typically much greater than that of the field camera 32a, and thus, a single pixel shift (horizontal or vertical) command to the pan/tilt mechanism 40 from the field camera 32a, translates to a several pixel shift in the field of view of the tag camera 32b.

[0058] As shown in FIGS. 9a-9c, a portion of the barcode 64 may be cut-off in the image 154, as well as some of the alignment markers 62. The alternative procedure shown in FIGS. 9a-9d enable the tag camera 32b to be repositioned in order to orient the barcode 64 such that it is visible for subsequent scanning (i.e. in a desired orientation). Preferably, the centering operation is executed for each scan, regardless of the accuracy of the coarse adjustment caused by the “touch” of the operator. When a tag 160 is accurately centered after the coarse adjustment, only a minor additional time overhead is required, however, when the tag 160 is substantially off-center, the procedure can save several seconds from the read operation when compared to having the operator initiate a manual re-centering.

[0059] The alternative procedure for aligning tags 160 uses a series of virtual sensors implemented in a software routine to conduct scans along defined paths in the image 154 to identify or “sense” segments. Segments are regions of similar intensity, differentiated from other regions by an intensity gradient, which is preferably user selectable. Each scan effectively causes a “soft” sensor to interrogate the image and mark or identify segments that it intersects. Preferably, three concentric sensors are used. In the embodiment shown in FIG. 9a, three sensors each scan an oval path (inner 202, mid 204, outer 206) to define concentric zones arranged from the center of the image 200 out to the edges of the image field. A marker 208 is placed on the image within each segment identified by a sensor. The number of these points in the image is indicative of distribution of segments in the image.

[0060] A well centered tag 160 should produce an equal distribution of segments, and thus markers 208, about the center of the image 154, such as that shown in FIG. 9d. In such a case, the segment positions would then cancel each other out, to produce an average position of the segments, near center 200. A tag 160 that is towards one side of the image field, e.g. FIGS. 9a-9c, will cause an imbalance in the number of segments on that side, resulting in the average segment position being shifted towards that half of the image 154. In FIGS. 9a-9c, the barcode 164 is located towards the bottom right portion of the image 154, and reports a large number of small segments in that area. Small segments are segments that are of a particular size, measured in pixels, e.g. <10 pixels, and are likely to indicate the presence of a barcode bar (white or black). An average 215 of the position of these small segments, measured from the center 200 computes a vector 214 (see FIG. 9b).

[0061] Referring to FIG. 9c, a horizontal sensor 210 and vertical sensor 212 can also be used to provide greater accuracy. These sensors scan along the image at the average position 215 as shown in FIG. 9c, and are used to adjust the average position 215, to determine a second average position 217, that better represents the centre of the barcode 164. A second vector 216 is then produced that more accurately reflects the offset of the barcode 164. For a horizontal barcode, e.g. FIGS. 9a-9c, the oval segmentation sensors 202-206 would provide the vertical offset, and the horizontal sensor 210, the horizontal offset. Similarly, for a vertical barcode (not shown), the oval sensors would provide the horizontal offset, and the vertical sensor 212, the vertical offset.

[0062] The following describes the alternative procedure for aligning the tag 160, in greater detail, making reference to FIGS. 9a-9c, 10 and 11. In the image 154 shown in FIGS. 9a, the three oval sensors 202-206 are configured to mark segments that are at least 5 pixels in size, which is the typical width of the smallest barcode bar. It will be appreciated that this procedure may be used for aligning other indicia such as an alpha-numeric string, wherein the threshold of 5 pixels may be adjusted to recognize, e.g., the smallest possible character width.

[0063] An edge contrast may be used to identify barcode segments, and is determined through experimentation during an initial calibration. A suitable range is 7-15%, which is high enough to ignore minor noisy segments, but low enough to pick as many valid barcode segments at a relatively poor focus as possible.

[0064] As shown in FIG. 10, when the alignment procedure is executed, a script examines each segmentation sensor 202-206 in turn, and determines the number of segments identified by each sensor. First, the sensor of interest is chosen, e.g. starting with sensor 202, and the number of segments is then determined and compared to a threshold, e.g. 10. If the number of segments is less than 10, chances are that there is no barcode intersecting the sensor 202, just background noise. In FIG. 9a, it can be seen that sensor 202 has only 1 segment, and would therefore be ignored in calculating the offset of the tag 160. However, the next sensor, e.g. 204, clearly has more than 10 segments, and would therefore be used to calculate the average segment position 215 (shown in FIG. 9b).

[0065] Since segments on a barcode 164 should not, ideally, be larger than a certain threshold, e.g. approximately 10 pixels, those that are larger than the threshold are ignored, eliminating stray segments, background segments etc. This ignores the curvature of the path in which the sensors may perform their scan. An oval path may report a larger segment width since the path in which it travels may not traverse the segment along the shortest path. This would result in a measured segment width that is larger than that of the segment's true size. Segments can also be identified as larger than they truly are, if adjacent barcode bars are missed due to poor focus etc. The threshold is chosen to accommodate operational variations.

[0066] Turning to FIGS. 9a and 10 specifically, since sensor 202 has been ignored, sensor 204 is next analysed. There are greater than 10 segments according to the image 154 in FIG. 9a, therefore, the first segment is selected, and its size determined. If the segment selected is smaller than the threshold, i.e., 10 pixels or less, its coordinates are saved to include in the average position. This is repeated until each segment has been analysed. As long as at least one of the segments has not been determined as “bad”, i.e., above threshold, an average horizontal and vertical position are determined based on all coordinates saved during the analysis.

[0067] The above process is repeated for each sensor, which in the example shown in FIG. 9a would involve one more iteration to evaluate sensor 206. If it was determined that all sensors were ignored, the aggregate average position is set to the center 200. If at least one of the segments has not been ignored, an aggregate average position 215 using all included sensors (these are shown in isolation in FIG. 9b), and all included segment positions is found. This calculation produces vector 214 shown in FIG. 9b.

[0068] Once all sensors have been analysed, the horizontal and vertical segmentation sensors may be used, as shown in isolation in FIG. 9c. It will be appreciated that using the horizontal 210 and vertical 212 sensors may be an optional procedure, however, the use thereof does provide a more accurate determination of the center of the barcode 164.

[0069] The steps in using the horizontal 210 and vertical sensors 212 is shown in FIG. 11, making reference to FIG. 9c. The horizontal sensor 210 is placed along the image 154 at the average vertical position (i.e. Y coordinate of 215) determined according to FIG. 10. Similarly, the vertical sensor 212 is placed along the image 154 at the average horizontal position (i.e. X coordinate of 215). A script will determine which line sensor (210 or 212) has a greater number of segments, to decide whether the barcode 164 is oriented vertically or horizontally. It is clear from FIG. 9c that the horizontal sensor 210 has a greater number of segments, and the barcode 164 is clearly oriented in a horizontal fashion.

[0070] In this example, since the horizontal sensor 210 has a greater number of segments, the process continues on the right hand path shown in FIG. 11. Once the proper sensor has been chosen, the number of segments identified by that sensor is determined, and if there are fewer segments than a particular threshold, the process is bypassed. In FIG. 11, that threshold is three (3) segments. If the horizontal sensor 210 has identified three or more segments, which in FIG. 9c is true, a loop commences that measures the size of each segment, and if the segment is smaller than a threshold, e.g., 15 pixels, then the coordinates of that segment are to be included in the second average position 217. Similar to the oval sensors, this process is repeated for each segment until all have been analysed.

[0071] If all segments were bad, the average X position is set to the X coordinate of center 200, and if not, an average X position is computed for all included segments. Differential X and Y measurements are then calculated by subtracting the X coordinate of center 200 from the average X position and the Y coordinate of the center 200 from the average Y position. In this example, the average Y value remains the one calculated by the oval sensors. The differential measurements are then compared to respective thresholds, and if the differential measurements are not above those thresholds then the barcode 164 is within the suitable limits and a move is not required. If however at least one of the X or Y differential measurements are greater than its respective threshold, a second vector 216 extending from center 200 to the position dictated by the X differential and Y differential measurements, i.e. 217, is computed. This vector 216 provides a better estimate of the center of the barcode in the horizontal direction, as shown in FIG. 9c.

[0072] It will be appreciated that the steps taken for measuring a vertical barcode are similar to those that have been described above, and therefore, need not be reiterated.

[0073] As long as at least one of the differential measurements is greater than its respective threshold, a pan/tilt operation will be performed by the pan/tilt mechanism 40, which aligns the tag 160 within the image as shown in FIG. 9d. At this point, the imaging system 14 will analyse the image and determine if further adjustment is needed, or if a particular scan direction is needed. For example, the tag 160 is oriented “upside-down”, and thus the barcode scan operation would need to take this into account. The imaging system 14 may then determine the up-down/left-right orientation and scan accordingly.

[0074] To achieve the most accurate results: a reasonable focus should be used so that the maximum number of barcode segments may be encountered; a reasonably consistent background is preferred, which is difficult to control, however should be considered; and if possible, having no other tags within the field of view of the cameras 32a, 32b is also preferred, to minimize confusion with the background.

[0075] It will be appreciated that the above alternative alignment procedure can be used in place of the procedure shown in FIGS. 5a and 5b, and the choice of which procedure to use, is dependent on the application. For instance, in an application where the objects being scanned are rectangular, e.g., shipping containers, either alignment procedure is suitable. On the other hand, in applications where the objects are curved, e.g., rolls of steel, the alternative alignment procedure is more appropriate.

[0076] Although the invention has been described with reference to certain specific embodiments, various modifications thereof will be apparent to those skilled in the art without departing from the spirit and scope of the invention as outlined in the claims appended hereto. The entire disclosures of all references recited above are incorporated herein by reference.