Classifications

• • G—PHYSICS
  • G01—MEASURING; TESTING
  • G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
  • G01D4/00—Tariff metering apparatus
  • G01D4/008—Modifications to installed utility meters to enable remote reading
• • G—PHYSICS
  • G01—MEASURING; TESTING
  • G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
  • G01D5/00—Mechanical means for transferring the output of a sensing member; Means for converting the output of a sensing member to another variable where the form or nature of the sensing member does not constrain the means for converting; Transducers not specially adapted for a specific variable
  • G01D5/26—Mechanical means for transferring the output of a sensing member; Means for converting the output of a sensing member to another variable where the form or nature of the sensing member does not constrain the means for converting; Transducers not specially adapted for a specific variable characterised by optical transfer means, i.e. using infrared, visible, or ultraviolet light
  • G01D5/39—Scanning a visible indication of the measured value and reproducing this indication at the remote place, e.g. on the screen of a cathode ray tube
• • G—PHYSICS
  • G06—COMPUTING; CALCULATING OR COUNTING
  • G06M—COUNTING MECHANISMS; COUNTING OF OBJECTS NOT OTHERWISE PROVIDED FOR
  • G06M1/00—Design features of general application
  • G06M1/27—Design features of general application for representing the result of count in the form of electric signals, e.g. by sensing markings on the counter drum
  • G06M1/272—Design features of general application for representing the result of count in the form of electric signals, e.g. by sensing markings on the counter drum using photoelectric means
• • G—PHYSICS
  • G06—COMPUTING; CALCULATING OR COUNTING
  • G06M—COUNTING MECHANISMS; COUNTING OF OBJECTS NOT OTHERWISE PROVIDED FOR
  • G06M3/00—Counters with additional facilities
  • G06M3/06—Counters with additional facilities for printing or separately displaying result of count
• • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
  • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
  • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
  • Y04S20/00—Management or operation of end-user stationary applications or the last stages of power distribution; Controlling, monitoring or operating thereof
  • Y04S20/30—Smart metering, e.g. specially adapted for remote reading

Definitions

• the present invention generally relates to automatic meter reading, and more particularly to a method and a device for obtaining a meter reading of a counter
• each symbol position being capable of showing a plurality of symbols in sequential order so as to indicate respective values of the symbol position, by processing a captured image of the counter mechanism.
• a very common example of a meter apparatus with a counter mechanism according to the above is a consumption meter such as an electricity meter, gas meter, fuel meter or water meter.
• a consumption meter such as an electricity meter, gas meter, fuel meter or water meter.
• the remainder of this document will therefore refer to an electricity meter and its counter mechanism, in an exemplifying but non-limiting sense.
• Electricity meters are of course commonplace both in private homes and at public sites such as offices, industries or community facilities and are used to measure the consumption of electrical energy for use by electricity providers when billing the users.
• FIG. 1 A typical electricity meter 1 is shown in Fig 1.
• a rotary disk 4 which is caused to rotate, as indicated by an arrow 5, around an axis 6 at a speed which is proportional to the momentary electric power consumption.
• the rotary disk 4 is mechanically coupled to a counter mechanism 3 in the form of a cyclometer register with a series of rotary reels 12, each reel being provided at its peripheral edge with a plurality of markings which represent respective meter symbols 10 at respective symbol positions IO 1 -IO 5 and which typically are the decimal digits 0-9 in sequential order.
• a fraction of the peripheral edges of the reels is visible through openings in a masking plate 11, each opening being sized to dis- close one symbol in full height.
• the reels are arranged to rotate at different speeds, such that the reel that represents the first symbol position 1Oi rotates ten times faster than the second reel, which in turn rotates then ten times faster than the next reel, and so on.
• the counter mechanism 3 of the electricity meter 1 can represent a five-digit or six-digit decimal integer, often combined with a decimal value (decimal position 1O d in Fig 1) .
• the counter mechanism 3 displays the current meter reading "12345,0".
• the counter mechanism 3 has a cent mark position 10 c .
• An electricity meter of this type is sometimes referred to as a Ferraris meter.
• electricity meters are read by manual visual inspection of the current meter reading as provided by the current positions of the rotary reels.
• the manual inspection is either performed by service personnel from the electricity providers, or by the site owner himself reading the meter reading and reporting it to the electricity providers by phone or over the Inter- net.
• EP-A-O 279 759 discloses a method and a device for optical reading of figures of a counter.
• the device reads figures on the digital display of the counter and consists of a housing which is fixed to the counter and which contains an optical device and a scanning image analyzer.
• the optical device forms an image of the digital display in the image plane of the analyzer.
• the video signal emitted by the analyzer is converted into elementary binary signals by a digitizer.
• These binary signals are sent to a pattern recognition device comprising a microprocessor which is programmed to compare each combination of binary signals corresponding to one decimal figure with ten combinations which are contained in a memory and which correspond to the ten decimal figures.
• decimal position 1O d is also transitional between “9" and "0". If at least one symbol position in the captured image is transitional and contains image components from two adjacent symbols or markings on the same reel, this prior art approach will most likely fail to provide an accurate interpretation of the complete current meter reading.
• each symbol template is a 10x10 pixel block, and a total of 90 such symbol templates are provided, where two adjacent templates have an overlap of 9 pixel rows.
• the 90 symbol templates together represent a development or "flattening out" of the markings on each reel, in the form of a total set of template pixel data consisting of 90 pixel rows of 10 pixels each.
• Another problem which is relevant in optical meter reading originates from the fact that the markings which represent the symbols of the counter mechanism are in fact curved, since they are provided on the peripheral edge of a cyclometer register reel. Therefore, a symbol in a captured image will be somewhat distorted compared to its ideal planar appearance. Moreover, disturbances or defects in the captured image may appear in the form of shadows or reflections, which may lead to truncated or partially non-readable symbols in the captured image.
• a similar problem is skew effects in the captured image because of a misalignment in the mounting of the meter reading device with respect to the front face of the electricity meter and, particularly, its counter mechanism.
• an objective of the invention is to solve or at least reduce one or more of the problems discussed above.
• an embodiment of the invention seeks to provide a method of obtaining a meter reading optically which is efficient in terms of processing power and memory usage, is accurate and robust and deals with the different practical problems set out above, and is adapted for use with different models of electricity meters and handles individual variations between meters of the same model.
• a first aspect of the invention is a method of obtaining a meter reading from a captured image of a counter mechanism, the counter mechanism having at least one symbol position, each symbol position being capable of showing a plurality of symbols in sequential order so as to indicate respective values of said symbol position, the method involving: ⁇
• reference image data is provided as a first set of symbol templates and a second set of symbol templates
• processing of the symbol image data involves a preprocessing stage and a matching stage
• method further involves selecting, based on an outcome of the preprocessing stage, at least one of the first and second sets of symbol templates for use during the matching stage .
• the symbols of the counter mechanism each have a respective associated value, and they will typically be decimal digits, or alternatively other kinds of figures or characters. Often, only one single symbol will be shown at a time at the symbol position and consequently appear in the symbol image data derived from the captured image for the symbol position. In this case, the current value of the symbol position will be the value associated with the symbol, as determined by the matching of the symbol image data with the reference image data. An area in the captured image where such one single symbol appears will be referred to as a whole symbol Image in the following.
• the method according to the first aspect enables optical meter reading with high accuracy using a limited number of matchings.
• each template in said first set unambiguously represents a respective symbol and each template in said second set represents two adjacent symbols in said sequential order.
• Symbol components representing symbols or parts thereof in said symbol image data are identified in the preprocessing stage, and its outcome will be indicative of a number of such symbol components having been identified. If the outcome of the preprocessing stage indicates that only one symbol component has been identified, it is likely that the symbol image data is a whole symbol image, and therefore only said first set of symbol templates is selected for use during the matching stage.
• the second set of symbol templates may be selected for use during the matching stage. More specifically, one embodiment involves:
• an additional step c4) may be performed which involves choosing, as a result of said matching stage, a single best result, in terms of best correlation between symbol image data and template, from said first matching in step cl) and said second matching in step c3) .
• the symbol image data may be binary image data, wherein said symbol components are identified as
• symbol components are also referred to as connected components in the following.
• the matching stage may involve correlation matching of each symbol template in the selected at least one of the first and second sets of symbol templates against symbol components as identified in said symbol image data during said preprocessing stage.
• the correlation matching involves locating a reference position for an identified symbol component in said symbol image data, locating a reference position in a given symbol template, and determining a correlation between individual pixels in said identified symbol component and individual pixels at corresponding
• the first predetermined criteria mentioned in step a) above may stipulate that the connected component has at least a certain size in at least one image dimension. In one embodiment, the first predetermined criteria therefore include that the connected component has a height of at least 2/3 of the nominal height of symbols in the symbol templates.
• the second predetermined criteria mentioned in step c2) above may stipulate that the best correlation between the symbol image data and any template in the first set of templates exceeds a first threshold, such as 75%.
• the second predetermined criteria may also stipulate that the difference between aforesaid best correlation and the second best correlation between the symbol image data and any template in the first set of templates exceeds a second threshold, such as a
• the method further involves performing one matching for said symbol image data starting from the top of a given symbol template and another matching for said symbol image data starting from the bottom of said given symbol template.
• the method is advantageously performed to obtain a meter reading for a counter mechanism having a series of symbol positions, and the same reference image data is advantageously used for at least some of the symbol positions.
• the counter mechanism may be a cyclometer register having a series of rotary reels, a peripheral edge of each reel being provided with a plurality of markings representing said plurality of symbols for a respective symbol position.
• the counter mechanism may belong to a consumption meter such an electricity meter, gas meter, fuel meter or water meter.
• a second aspect of the invention is a meter reading device adapted for use with a counter mechanism according to the above.
• the meter reading device has an image sensor, a memory and a controller coupled to the image sensor and the memory.
• the controller is adapted to perform a method as described above for the first aspect.
• a third aspect of the invention is a method of obtaining a meter reading from a captured image of a counter mechanism, the counter mechanism having at least one symbol position, each symbol position being capable of showing a plurality of symbols in sequential order so as to indicate respective values of said symbol position, the method involving:
• reference image data is provided as a first set of symbol templates and a second set of symbol templates, each template in said first set unambiguously representing a respective symbol and each template in said second set representing two adjacent symbols in said sequential order.
• the processing of the symbol image data may involve a preprocessing stage and a matching stage, wherein the method of the third aspect may further involve selecting, depending on an outcome of the preprocessing stage, at least .one of the first and second set of symbol templates for use during the matching stage.
• a fourth aspect of the invention is a meter reading device where the controller is adapted to perform a method as described above for the third aspect.
• Embodiments of the invention provide optical meter reading with improved accuracy and credibility, since the aforementioned practical problems of flash reflections, symbol transitions, cut-off symbols, and symbol misalignments are handled in a controlled and intelligent manner. Still, the number of symbol templates is kept low, which has benefits both in that little storage space is required and in that the number of required matchings with the captured symbol image data is limited. In turn, this allows a miniaturized implementation of the meter reading device, with a low-power processor and a small memory. Such a miniaturized meter reading device is beneficial in that it allows a slim design and non-obscuring mounting.
• Fig 1 is a schematic front view of a typical
• Fig 2 illustrates the counter mechanism in a
• Fig 3 is a schematic block diagram of a meter reading device according to one embodiment.
• Fig 4 is a schematic sectional view of the meter reading device shown in Fig 3.
• Fig 5 illustrates a typical operating scenario where a plurality of meter reading devices are coupled to a central unit, which in turn communicates with a remote server over a communications network.
• Fig 6 is a flow chart that illustrates a normal or operating mode of the meter reading device. Detailed Description of the Invention
• Fig 3 is a schematic block diagram of a meter reading device 20 according to one embodiment.
• the meter reading device 20 is adapted to be mounted onto a transparent front surface portion 2 ' of the apparatus housing 2 of the electricity meter 1 shown in Fig 1.
• other embodiments of the meter reading device may be adapted for use with other kinds of consumption meters, such as electricity meters, gas meters, fuel meters or water meters, or even with other counter mechanisms than those found in consumption meters .
• the meter reading device 20 has a housing 21 and a plurality of components contained therein, the most important of which being a controller 22 with associated memory 30, and an image sensor 26. Controlling the image sensor 26, the controller is adapted to cause capture of an image 28 of the counter mechanism 3 of the electricity meter 1. To this end, a plurality of illuminating LEDs 25 are provided for illuminating the counter mechanism 3 when the image is to be captured. Optics 24 are provided for guiding light reflected from the front edge surface 12a of the reels 12 of the counter mechanism 3 onto a light sensitive area of the image sensor 26.
• the optics include first and second mirrors 24b, 24d and a lens arrangement 24c.
• the first mirror 24b is partly transparent (dichroic) and is therefore advantageously transmissive in a first light spectrum within the visible wavelength range but reflective in a second light
• the illuminating LEDs 25 and the image sensor 26 will then operate in the second light spectrum, whereas a human user can inspect the counter mechanism 3 through an inspection window 42 in the housing 21 aligned with the masking plate 11 of the electricity meter 1, so as to obtain a manual meter reading by ocular inspection.
• the meter reading device 20 is aligned with the masking plate 11 of the counter
• a rear housing portion 21' of the meter reading device 20 is transparent.
• the image sensor 26 is a CMOS sensor capable of capturing an 8-bit, 640x480 pixel grayscale image 28.
• various other types of image sensors including CCD sensors, having different resolutions and/or color/grayscale depths, are equally possible for implementing the image sensor 26.
• the captured image 28 is provided to the controller 22, which stores the image 28 temporarily in a work memory (RAM) 32 which is part of the controller's associated memory 30.
• the controller 22 may be implemented by any commercially available CPU, DSP or any other electronic programmable logic device such as FPGA, or as an application-specific integrated circuit (ASIC), or as discrete components, or as any combination of the above.
• the meter reading device 20 also has a communication interface 23, through which a resulting output signal 39 can be transmitted as a representation of a current meter reading as ultimately determined by the controller 22. Even if not shown in the drawings, the meter reading device 20 of the disclosed embodiment is coupled to receive electric power for its various components from mains power lines connected to or in the electricity meter. Alternatively, the meter reading device 20 may be provided with its own power supply means, for instance in the form of a long-life battery.
• the components 22, 23, 26 and 30 referred to above are mounted on one surface side of a printed circuit board (PCB) 40, whereas the illuminating LEDs 25 are mounted on the opposite surface side of PCB 40.
• PCB printed circuit board
• the controller's associated memory 30 also includes a non-volatile memory, such as a flash memory or EEPROM, for storing reference image data 34.
• a non-volatile memory such as a flash memory or EEPROM
• the reference image data 34 will be used when interpreting the value of each symbol position in the captured image 28 of the counter mechanism 3.
• the controller 22 will derive symbol image data 29 by extracting a sub area of the captured image 28 in which the symbol position in question is located.
• the symbol position to be determined is symbol position IO 3 , it shows a symbol "3", and the symbol image data 29 derived will thus be based on the sub area of the captured image 28 that contains the image of this symbol.
• a symbol position is transitional between two adjacent symbols when the image 28 is captured.
• the symbol position IO 2 in Fig 3 is transitional between the symbol "4" and the symbol "5". Consequently, for this symbol, the symbol image data 29 derived from the captured image 28 will contain image components which originate both from (the lower part of) the symbol "4" and from (the upper part of) the symbol "5" on the reel 12 for that symbol position.
• the reference image data 34 is provided as a first set of symbol templates 36 (referred to as “whole symbol templates” in the following) and a second set of symbol templates 38 (referred to as “whole symbol templates” in the following)
• a first set of symbol templates 36 referred to as “whole symbol templates” in the following
• a second set of symbol templates 38 referred to as “whole symbol templates”
• Each whole symbol templates 36 contains a reference image of a respective symbol “0”, “1”, “2”, “3”, “4", “5", “6”, “7”, “8”, and “9” that represents the corresponding reel marking when being located fully visible and centered in the opening in the masking plate 11.
• each transitional symbol templates 38 contains a reference image of a respective transitional state between two adjacent symbols, i.e. when two
• the 10 transitional symbol templates 38 represent the following pair of adjacent symbols: "0/1”, "1/2", “2/3", “3/4", "4/5”, “5/6”, “6/7”, “7/8", “8/9", and "9/0".
• selections 1) only against the whole symbol templates 36, 2) only against the transitional symbol templates 38, or 3) against first the whole symbol templates 36 and then, if necessary, against the transitional symbol templates 38.
• other principles may apply as regards the selection of which template set(s) to use for the matching, and/or in what order to use it/them.
• the meter reading device 20 of the disclosed embodiment has three main modes:
• a template creation mode in which the first and second template sets 36 and 38 (whole symbol templates and transitional symbol templates) are generated together with certain related reference data 35 pertaining to a certain type or model of the electricity meter with which the meter reading device is to be used.
• the template creation mode typically takes place at a central location, such as at a manufacturer or distributor of electricity meters, at a manufacturer or
• a distributor of meter reading devices or at an electricity provider.
• an operator connects a meter reading device (or another suitable image capturing equipment having similar optical performance as the type of meter reading device for which symbol templates are to be created) to the face plate of an electricity meter of the relevant type or model.
• a computer such as a personal computer or workstation is connected to receive images of the electricity meter's counter mechanism during forced operation. This means that a large current is looped through the electricity meter, forcing its counter mechanism to rotate much faster than during typical ordinary consumption conditions.
• cent marks 10 c are identified, and their frequency (or mutual distance between adjacent cent marks) is determined through appropriate image processing, involving for instance a Fourier transform.
• the operator may assist in identifying the sub area of the captured image in which the cent marks are located by appropriate user interaction, for instance by marking the relevant area in a magnified reproduction of the captured image on the computer display by means of a computer mouse or joystick.
• the first symbol position (1Oi in Fig 1) is identified in a similar way (the present embodiment does not involve interpretation of the decimal position 1Od)•
• the first reel 12 starts to rotate and its currently visible symbol 10 starts to move vertically within its opening in the masking plate 11.
• Counting the cent marks 10c as they pass in the opening in the masking plate 11 it is possible to determine accurately when the first symbol position changes alternately between subsequent whole symbols "1", “2", ..., and their interleaved transitions "0/1", "1/2", etc.
• Images are captured sequentially as the symbols appear, and the thus captured 10 images of whole symbols and 10 images of transitional symbols are stored in temporary memory.
• a symbol component is a connected component in said symbol image data, identified as a contiguous pixel segment where at least a predetermined number of image pixels of the same binary value are positioned adjacent to each other.
• 8 or more image pixels positioned adjacent to each other in a contiguous pixel segment are regarded to qualify as a connected component.
• Other embodiments may use other criteria for identifying connected components.
• the second largest connected component is kept.
• the respective image is stored as one of the 10 whole symbol templates or 10 transitional symbol templates, as the case may be.
• each symbol will be represented in one of the 10 whole symbol templates (the whole symbol) , and in two of the 10 transitional symbol templates (the upper half of the symbol in one, and the lower half in the other) .
• related reference data is obtained by further image capture and processing, in the form of: • The number of symbol positions IO 1 -IO 5 of the
• the generated symbol templates and related reference data are stored permanently. If the meter reading device 20 is available already at the time of template creation, the generated data may be stored directly as reference image data 34 including the first and second template sets 36, 38 and the related reference data 35 in its memory 30, or otherwise the generated data will be stored in a databank from which it may be downloaded into the meter reading device 20 at a later occasion.
• the calibration mode occurs during installation of the meter reading device 20 by a service person, etc, on the front face of the individual electricity meter 1 for which it is to be used.
• the purpose of the calibration is to determine precisely the image positions of the
• a skew detection step is performed.
• the skew detection serves to detect any angular deviation in the mounting of the meter reading device 20 with respect to the counter mechanism 3 (or more particularly the
• the skew detection will apply edge detection followed by Hough transforms of edge positions to detect naturally appearing edges in the captured image
• skew compensation transformation data will be calculated and stored in the related reference data 35 for use during the rest of the calibration mode, and also for later use in the initial processing of all captured images in normal mode, to compensate for such detected skew .
• the calibration mode locates the cent marks in a captured image, similar to what is done in the template creation mode.
• the approximate image positions of the respective symbol positions in captured images are given.
• the symbol contents are derived from a captured image in a neighborhood of the approximate position and tested against the image reference data 34 so as to determine the exact image position where the symbol contents best matches the image reference data 34.
• the testing is performed by correlation matching
• a human operator performing the calibration can read a resulting meter reading as provided in the output signal 39 and verify its accuracy by comparing with an ocular inspection of the current meter reading through the inspection window 42. After completed calibration, the controller 22 will store the calibrated image
• the operation of the meter reading device 20 can be divided into four main stages: an image capture stage 100, a symbol image data derivation stage 110, a matching stage 120, and a post-processing stage 130.
• the controller 22 controls the illuminating LEDs 25 and the image sensor 26 in a step 102 to produce two (or more) images of the counter mechanism 3 directly after each other (i.e, within a time frame which is short enough so that no parts of the counter mechanism (particularly the cent marks 10 c ) will move in position between the two images) .
• Each image is taken at a different ' lighting condition, and the reason for this approach is to reduce the afore- described problem of disturbances or defects appearing in the captured images due to shadows and reflections.
• the first and third LEDs are used when capturing the first image
• the second and fourth LEDs are correspondingly used to capture the second image.
• the capturing of these two images involves various aspects in terms of illumination times for the LEDs 25, integration times for the image sensor 26, etc. The particulars thereof are left out herein in order not to obscure the invention in unnecessary detail.
• a step 112 will extract a symbol image as a sub area of each of the two captured images at a position therein which can be determined from the related reference data
• a histogram of the intensity of each extracted symbol image is calculated and analyzed, and based on this the best symbol image is selected in step 114 in terms of the least presence of reflections. Exceptionally, it is difficult to determine the best symbol image in step 114, and in such a case the operation may commence with both symbol images in the following steps and leave the ultimate decision to a later stage 128 that involves confidence value determination.
• the symbol image selected in step 114 is binarized using suitable image processing techniques well known per se, typically including actions like background extraction, threshold matrix determination and binarization .
• the symbol image is a binary image matrix, where pixels having a first binary value (e.g. 1) represent foreground image contents and pixels having the other binary value (e.g. 0) represents absence of foreground image contents.
• step 118 symbol components are identified in the resulting symbol image.
• this involves locating all connected components in the symbol image, as has been described above for the
• the two connected components are determined which have the largest pixel masses. If those two components are true symbol representations (i.e. represent first and second partial symbols in transition) , then they should not overlap each other vertically. Therefore, a test to this effect is made, and if it turns out successful, all other connected
• the resulting symbol image at completion of step 118 constitutes the symbol image data 29 which will be used in the following matching stage 130.
• the matching stage 130 performs correlation matching of the symbol image data 29, as derived in the preceding stages, with the reference image data 34.
• a decision is made as to which of the set(s) of whole symbol templates 36 and transitional symbol templates 38 to perform the matching upon.
• a test is first made as to whether any of the connected components has a height which is at least 2/3 of the nominal symbol height, as obtained from the related reference data 35. If so, the correlation matching is first performed based on the whole symbol templates 36. If necessary, the correlation matching is then also per- formed based on the transitional symbol templates 38. The criteria that determines whether or not to involve also the transitional symbol templates 38 will be explained in more detail later.
• the symbol image data 29 contains components, originating from two adjacent reel markings, of roughly the same size, and the best match will probably be found with any of the transitional symbol templates 38. There- fore, in this case, the correlation matching in the remaining steps of the matching stage 120 is performed only upon the transitional symbol templates 36.
• the correlation matching in step 124 is performed against the selected set of templates, as determined above.
• Each template in the selected set is correlated against the one or two connected components in the symbol image data 29, resulting in a correlation percentage from the matching.
• the correlation matching in the disclosed embodiment operates on aforesaid connected components, rather than the entire symbol image data 29 which includes also background pixels, in the following manner.
• the symbol image data 29 is a whole symbol image
• the top of the single connected component i.e. the first (uppermost) pixel where the single connected component starts
• the top of the whole symbol in the whole symbol template is located, and the matching between them occurs by performing a bit-wise logical AND operation between pixels in the connected component and pixels of the symbol in the whole symbol template.
• the result of the logical AND operation will be the number of matching pixel positions between connected component and symbol template.
• the correlation percentage is determined by dividing this number with the mean value of the number of pixels of the connected component plus the number of pixels of the symbol in the symbol template
• the reference position for the matching will be the top of the aforesaid lower connected component in the transitional symbol image (e.g. the top of the symbol "5" )
• the correlation matching in the disclosed embodiment has a sub step 125a which handles the afore-described problem of truncated or partially non-readable symbols in the captured image because of disturbances or defects caused by shadows or reflections.
• the connected component (s) in the symbol image data 29 will be shorter than the nominal symbol height, or differently put, the relevant area of the symbol image data 29 to be used for correlation is shorter in height than the template. Therefore the sub step 125a involves first matching the connected component (s) in the symbol image data 29 with the template starting from the top of the latter. If the resulting correlation percentage is not sufficiently good (i.e. does to reach a percentage threshold), then a second matching is performed, this time starting from the bottom of the template.
• the correlation matching in the disclosed embodiment also has a sub step 125b which is performed if the matching of a particular symbol fails to result in a match that meets a predetermined correlation limit.
• the correlation matching is performed at varying offset positions in vertical and horizontal pixel directions in the symbol image data 29.
• step 127 the best match, i.e. the particular template among the whole symbol templates 36 and/or transitional symbol templates 38 that yields the best correlation percentage with the symbol image data 29, is selected, and the symbol represented by that template will be chosen as the interpreted symbol at the current symbol position IO 1 -IO 5 of the counter mechanism 3.
• step 127 also records a decimal value for the interpreted symbol .
• a confidence value is calculated, the purpose of which is to indicate the relative accuracy of the best match.
• the confidence value is determined as a function of the cor- relation percentages of the best and second best matches.
• a round logic step 132 will compile the resulting interpreted values for the respective symbol positions into a joint meter reading estimation consisting of an integer value having, of course, the same number of symbol positions IO 5 -IO 1 as the counter mechanism 3. If an interpreted symbol at a symbol position 1O n has a decimal value O 0, as determined above, this will be considered when determining an integer value not only for this symbol position but also for at least the following symbol position 10 n+ i.
• the round logic step 132 takes into account that due to the mechanical nature of the counter mechanism, a certain symbol position may be offset by as much as a couple of decimal units compared to other symbol positions, because the corresponding reel either rotates ahead or lags behind other reels.
• the round logic step 132 also takes into account that a symbol position which has not yet switched completely from one integer value to the next must not be rounded to that next integer value, since this could result in excessive billing (if the unit of the electricity meter reading is kWh, an incorrectly rounded symbol at symbol position IO 4 would mean a difference in 1000 kWh) .
• step 134 the result is evaluated in a step 134.
• the confidence values determined in step 128 If the confidence value is low for any of the determined symbols, this may indicate an ambiguity in the correlation matching caused by two different templates yielding similar correlation percentages.
• the controller 22 may wait a predetermined time period and then repeat all the operations starting with a new image capture stage 100. Alternatively, the controller may report the low confidence value in the output signal 39 to a central unit 50 (described below with reference to Fig 5) and receive a command from the latter to repeat the meter reading. Ultimately, the resulting meter reading estimation after the next iteration of step 132 will be found in step 134 to have a better and acceptable confidence value.
• a final step 136 the resulting meter reading estimation is communicated in the output signal 39 together with related information in the form of the con- fidence value and an identity of the meter reading device 20 or electricity meter 1 (preferably established by the operator during aforesaid calibration) . If the meter reading device 20 has a real-time clock, temporal
• information on current date and/or time may also be included in the output signal 39.
• a meter reading device 2Oi at an electricity meter Ii may be connected to the aforementioned central unit 50 together with a number of other meter reading devices 20 2 ⁇ 20s at corresponding electricity meters I 2 -I 6 .
• the communication interface 23 may be compliant with any available local wired communication standard, such as USB, Firewire or RS232, or short-range data communication standard such as Bluetooth or IrDA, or packet-switched data communication such as TCP/IP over Ethernet or GPRS over GSM/3G, or an industry sector-specific communication standard such as M-bus over an electric powerline network, to form a respective communication channel 52i ⁇ 52 6 to the central unit 50.
• the communication channels 52i-52 6 may be used to command reading (i.e., initiate performance of one iteration of the normal mode operations of Fig 6) at the respective electricity meters, and, in response to such a command, the resulting meter reading with its related information may be communicated as output signals 39i-39 ⁇ to the central unit.
• the central unit 50 has a communication channel 54 for connection to a remote server 60 which may be
• the communication channel 54 may involve a communications network 56 - such as a global area network (e.g. the Internet), a telecommunications network or an electric powerline network - and may be compliant with any appropriate communication standard, such as TCP/IP over Ethernet, GPRS over GSM/3G, or M-bus.
• a communications network 56 - such as a global area network (e.g. the Internet), a telecommunications network or an electric powerline network - and may be compliant with any appropriate communication standard, such as TCP/IP over Ethernet, GPRS over GSM/3G, or M-bus.
• meter readings as obtained from the meter reading devices 20i-20 6 may be reported to the remote server 60 and stored in an associated database 62, and, conversely, the remote server may issue commands intended for the meter reading devices 20i-20 6 and relayed by the central unit 50.
• meter reading does not have to be initiated by an external unit (e.g. central unit 50) but may alternatively be initiated by the controller 22 itself.
• the controller 22 may use a real-time clock and predefined calendar settings to perform meter reading at a desired periodicity, such as hourly, daily, weekly or monthly.
• the output signal 39 is not an external output signal but is instead stored locally in the meter reading device 20 (e.g. in the memory 30), for later access e.g. by connecting an external unit to the device 20 through the communication interface 23.
• One example is monitoring of a meter at a fuel tank or dispenser for the purpose of operation surveillance or remote identification of the need for maintenance or fuel refill.

Landscapes

• Physics & Mathematics (AREA)
• General Physics & Mathematics (AREA)
• Engineering & Computer Science (AREA)
• Theoretical Computer Science (AREA)
• Image Analysis (AREA)

Applications Claiming Priority (2)

Publications (1)

Family

ID=38123165

Family Applications (1)

Country Status (2)

Cited By (1)

Families Citing this family (15)

Family Cites Families (2)

• 2006
  • 2006-12-08 EP EP06835824A patent/EP1960741A1/de not_active Withdrawn
  • 2006-12-08 WO PCT/SE2006/001397 patent/WO2007067132A1/en active Application Filing

Non-Patent Citations (1)

Also Published As

Similar Documents

Legal Events