Classifications

• • G—PHYSICS
  • G06—COMPUTING; CALCULATING OR COUNTING
  • G06M—COUNTING MECHANISMS; COUNTING OF OBJECTS NOT OTHERWISE PROVIDED FOR
  • G06M1/00—Design features of general application
  • G06M1/08—Design features of general application for actuating the drive
  • G06M1/10—Design features of general application for actuating the drive by electric or magnetic means
• • G—PHYSICS
  • G06—COMPUTING; CALCULATING OR COUNTING
  • G06M—COUNTING MECHANISMS; COUNTING OF OBJECTS NOT OTHERWISE PROVIDED FOR
  • G06M1/00—Design features of general application
  • G06M1/08—Design features of general application for actuating the drive
  • G06M1/10—Design features of general application for actuating the drive by electric or magnetic means
  • G06M1/101—Design features of general application for actuating the drive by electric or magnetic means by electro-optical means
• • G—PHYSICS
  • G06—COMPUTING; CALCULATING OR COUNTING
  • G06M—COUNTING MECHANISMS; COUNTING OF OBJECTS NOT OTHERWISE PROVIDED FOR
  • G06M9/00—Counting of objects in a stack thereof

Definitions

• the invention relates to the field of counting devices of thin products stacked side by side for small series. More specifically, it is a matter of counting, in an automated manner and with a good rate, the number of thin products contained in a batch of small series.
• the present invention therefore aims to overcome one or more disadvantages of the prior art by creating a device for counting, in an automated manner, the number of thin products produced in small series, with a good rate.
• a lighting means of the stack producing one or more light beams covering at least the entire length of the stack
• detection means comprising at least one detection circuit, comprising a plurality of photosensitive elements, and at least one optical device, associated with the detection circuit, for focusing light rays reflected by the stack,
• processing means receiving signals from the detection circuit or circuits, able to extract from these signals brightness levels correlated with a dimension along the stacking axis expressed in pixels, the processing means generating a determined signal x (n) corresponding to the received signals and including:
• extraction means for extracting, from the determined signal x (n), a pattern representing a thin product; and calculating means for calculating the number of thin products, by intercorrelation of the determined signal with the extracted pattern, to determine an intercorrelation signal corresponding to the number of patterns present and corresponding to the number of thin products of the stack.
• processing means also comprise:
• pretreatment means for carrying out a Fourier transform allowing to provide from the received signals a transformed signal revealing harmonics and then to determine the characteristics of a filtering means for filtering the transformed signal with conservation of at least 1 harmonic ; said determined signal x (n) being a filtered signal resulting from the pretreatment.
• the pretreatment means comprise reconstitution means carrying out a transformation of
• the extraction means are arranged to extract the representative pattern of a thin product in the pre-processed signal.
• the means for extracting a pattern include:
• first calculation means for performing correlation or convolution functions on the preprocessed signal in a first step, then a Fourier transform calculation for estimating in a second time, for each of the frequencies of the Fourier domain, the module and the argument of the Fourier transform of the pattern representing the periodic signal position corresponding to a thin product;
• second calculation means using an inverse Fourier transformation to calculate said first pattern from results obtained by the first calculation means.
• the first computing means elaborate an autocorrelation function c ( ⁇ ) of the filtered signal x (n), defined (for example here in its non-normalized version) by the formula:
• N is the number of pixels of the image of the filtered signal
• Ep is the thickness of a thin product expressed in pixels.
• the first calculation means develop a convolution function conv ( ⁇ ) of the filtered signal on itself, defined by the formula:
• N-1] is the filtered signal; compute the Fourier transform of the n-correlation function b ( ⁇ 1, ⁇ 2) in the Fourier domain, via a two-dimensional Fourier transform, to obtain a matrix set of linear relations expressing the arguments of the function of n correlation according to the arguments of the pattern in the frequency domain of Fourier; and - invert the system (matrix inversion, or reduce to a triangular system) to go back from the n-correlation argument to the argument of the pattern in the Fourier domain.
• an invertible matrix for passing the argument of the transform of the n-correlation function to the argument of the pattern in the Fourier domain can be calculated.
• the resolution of the system can also be done by reducing the linear system to a triangular system. The resolution is then iterative.
• the means for parameterizing the thickness comprise means for estimating the thickness Ep using a first Fast Fourier transformation FFT, the estimation means producing:
• filtering means are provided to provide the extraction means with a filtered and denoised signal, the second calculation means for determining a first periodic pattern representative of a thin product to a possible phase shift.
• the extraction means execute at least one de-signaled signal processing algorithm for determining the signal pattern used for the intercorrelation, the shape of the pattern retained for a series of products being counted being estimated after a comparison between the first periodic pattern detected in the denoised signal and a reference pattern stored in the storage means.
• the parameterization means associated with the processing means are provided for storing the reference pattern during a counting performed by the counting device with a standard batch of thin products.
• the reference pattern can also be chosen from a series of standard geometrical shapes (slot, inverted slot, triangle, parabola portion, etc.).
• the filtering means is a comb filter configured to filter out, in the received signals, noise and frequencies that do not correspond to harmonics, in order to obtain a pre-processed signal in which frequencies far from the harmonics and may correspond to gaps or spaces between thin products are eliminated.
• the signal pattern extraction means comprise circular registration means making it possible to avoid obtaining a pattern offset by phase shift, the circular registration means reproducing from the first pattern patterns with different phase shifts, the phase shift finally applied being determined by using a reference pattern.
• the means for calculating the number of thin products comprise: means for calculating the intercorrelation between the extracted signal pattern and the denoised signal, making it possible to supply the intercorrelation signal; and
• the circular registration means comprise:
• the processing means generate a vector representative of the signals received and perform a FFT fast Fourier transform on this vector, the filtering means receiving the fast Fourier transform of this vector and realizing a frequency Fourier filtering after a determination of harmonics.
• said vector is generated by a program executing a zero-padding zeros filling method so that said vector corresponds to an increased signal size and groups a number N zp of signal samples, N zp being a power of 2, the program being provided with a function of deletion of the added zeros, this deletion function being activated to make it possible to obtain said filtered signal after application of the IFFT inverse fast Fourier transform.
• n is the number of pixels in the image of the denoised signal
• x (k) the denoised signal
• Ep is the thickness of a thin product expressed in pixels.
• a CIS module (equipped with a sensor CIS “contact image sensor”), arranged longitudinally and vis-à-vis the stack constitutes the lighting means and the detection means, the CIS module being of length at least equal to that of the stack, or the CIS module making displacements in the longitudinal direction of the stack vis-à-vis a zone covering at least the entire length of the stack in several steps.
• the device comprises a plurality of CIS modules, arranged longitudinally and opposite the stack, each CIS module comprising detection means and lighting means by a plane beam according to the determined direction , the sum of the lengths of the CIS modules being at least equal to the length of the stack.
• the CIS modules illuminate the stack according to a lighting line, each CIS module being inclined at a given angle so that its light plane beam meets this feature.
• Another purpose is the use of a counting system according to the invention to allow adaptations of certain manufacturing operations depending on the batch and to follow each batch permanently.
• the counting device by which information is transmitted, via means of communication, by the processing means to a processing system, such as a personalization machine, downstream of a processing line.
• the transmitted information including the number of thin products calculated by the device for each series constituting the stack and / or information making it possible to deduce this number and / or an identifier associated with each series.
• the processing system customizes the products of the series, physical or software personalization operations to be applied to each element of a series being associated with the information transmitted by the processing means.
• a further object of the invention is to enable the device to be used for personalization of smart cards or similar portable objects.
• the invention also relates to a use of the counting device, characterized in that a logical customization station, processing a series of thin products comprising an integrated circuit, allows the inscription, in memory of the integrated circuit, personalization information for the use for which the product is intended.
• Another object is to provide a method of processing detection signals that perform well and allow a rapid analysis of the signal to count the number of products of the same thickness in a more or less compact stack.
• a method for processing at least one signal originating from the detection circuit (s) (of the optical type) of a device for counting thin products characterized in that it comprises:
• a step of preprocessing said signal including a filtering of the signal to produce a filtered signal
• an estimation step in the filtered signal of a representative pattern of a product can be thick
• the filtering during the preprocessing step of said signal is performed after a Fourier transformation and by using a comb filter.
• the filtering can also be done by implanting a conventional finite or infinite impulse response filter.
• the method comprises a step of converting the signal, before the filtering, into data representative of brightness levels in correlation with a stack thickness dimension expressed in pixels, the estimation step defining a first representative periodic pattern of a thin product to a possible phase shift, and then using a reference pattern to achieve a circular registration to obtain a second pattern estimated without phase shift.
• the signaling step comprises a display of a number of smart cards to be processed by a chip card personalization machine and / or a transmission of the information representative of this number to the personalization machine.
• An additional object of the invention is to provide a program executable by a computer system for controlling the treatment adequately to obtain a fast and reliable count.
• the invention relates to a computer program directly loadable in the memory of a computer and including computer codes for controlling the steps of the method when said program is executed on a computer, said program thus allowing a count of series thin products of a stack.
• FIG. 1 represents a logic diagram of steps which summarizes the general course of a counting method according to FIG. 'invention.
• FIGS. 2A and 2B show an example of an amplitude graph of a fast Fourier transform associated with the signals coming from the photosensitive elements, and respectively illustrate the decomposition of the signal in harmonics and the pre-filtering of Fourier for the search of patterns.
• Figs. 3A and 3B respectively show a useful signal (denoised) and the corresponding signal including noise.
• Figure 3C illustrates a modeling of the pattern to be searched for in the denoised signal.
• Figures 4A and 4B illustrate the possible presence of a phase shift.
• FIGS. 5A and 5B respectively illustrate a reference pattern used in the circular registration, and an example of a circular registration run.
• FIG. 6 represents a general diagram of a search for the optimal pattern according to one embodiment of the invention.
• Fig. 7A is a perspective view showing an example of a counting device having a CIS module covering the entire stack.
• FIG. 7B shows a perspective view showing an example of a counting device comprising a CIS module covering the entire stack by longitudinal displacements.
• Figure 8 illustrates the cross-correlation between the pretreated signal and the estimated pattern.
• FIG. 9 represents an example of a counting device comprising a transverse CIS module performing a longitudinal displacement.
• FIG. 10 represents an example of a counting device comprising a CCD matrix camera performing longitudinal analyzes along several longitudinal lines.
• FIG. 11 represents an example of a counting device comprising a CCD matrix camera performing one or more longitudinal analyzes by displacement in the longitudinal direction.
• Fig. 12 is a perspective view showing an example of a counting device comprising a CCD camera.
• Figure 13A illustrates a signal obtained with a better contrast than that of Figure 2A.
• Figure 13B illustrates a similar signal obtained with poor contrast from that obtained in Figure 2A.
• FIGS. 7A and 7B show a counting device comprising a CIS module (3).
• One or more CIS modules (3, 3d) may be arranged longitudinally.
• a module (3, 3d) CIS comprises illumination means, a photosensitive cell and an integrated optical focusing device.
• FIG. 12 represents a counting device comprising lighting means (7), mirrors (9a, 9b) and a camera (8) CCD.
• Other cameras of the same type, comprising an optical device and a photosensitive circuit and producing an electrical signal depending on the light received are usable.
• Focusing the light rays reflected by the stack (5) allows recovery of one or more signals via at least one detection circuit. These signals are extracted to allow processing, in which it is sought to analyze the variations in brightness levels in correlation with a stack thickness dimension expressed in pixels.
• the device allows counting of series of products (2) thin, stacked side by side, by determining the repetition of a representative pattern of a product (2) in a filtered signal and denoised resulting from a transformation of received signals.
• a first Fourier transformation is used before comb filtering to obtain the denoised signal thereafter.
• a system based on Fourier transform and statistics with an order greater than 2 is used to allow to precisely define a periodic pattern representative of a thin product to a possible phase shift, via a calculation of the argument and the module of the transformed signal pattern in the Fourier domain.
• the device can comprise a rectangular container (4) containing the products (2) thin, only the products (2) at the ends of the stack (5) being shown in Figures 7 to 11.
• the products (2) thin can be maintained, so non-limiting, by a shrinkable transparent film or shims resting on the tray (4).
• the tray (4) serves in a non-limiting manner means for holding thin products (2).
• a magazine for processing thin products (2) is directly used.
• the stack (5) is illuminated along its entire length, by a plane beam of light rays (6, 6d) produced by the lighting means of a module (3, 3d) CIS or by a lighting means. diodes whose rays are focused in a plane by an optical device.
• the beam (6, 6d) plane projected against the stack (5) produces a bright line (T).
• the line (T) is then analyzed by means (3, 3d, 9a, 9b, 8) for detecting the reflected light intensity, associated with processing means (10).
• the lighting means comprise a fluorescent tube (7) which illuminates, by means of multidirectional spokes (7a), the entire upper part of the stack (5), including the area of the line ( T) light cited above, analyzed by the detection means associated with the processing means.
• the analysis of a longitudinal light line (T) by the detection means (3, 3d, 9a, 9b, 8) associated with the processing means is called longitudinal analysis of the stack (5).
• the analysis according to several segments of the stack (5), over its entire length, by the processing means (10) associated with the detection means is also understood as a longitudinal analysis.
• the light rays (6) emitted by the light source (s) allow a longitudinal analysis of the batch of products, that is to say parallel to the long side of the tray (4).
• the relative displacement of the tray with respect to the CIS module (s) is transverse, that is to say parallel to the small side of the tray (4), and involves longitudinal analyzes on different longitudinal zones.
• the line (T) longitudinal light is indeed moved to different levels depending on the width of the stack (5).
• 100 longitudinal analyzes are performed in a transverse movement (M4a, M3a) of back and forth, back and forth.
• different longitudinal analyzes are performed by transverse displacements not perpendicular to the longitudinal direction, the line (T) on the stack (5).
• a fluorescent tube (7) more powerful than diodes illuminates the entire upper part of the stack (5).
• a matrix photosensitive cell for example a CCD matrix 1
• a CIS module (3, 3d) or the camera (8) CCD are connected to a processing circuit for transmitting electrical signals from the transformation of light energy into electrical energy by the photosensitive cells.
• the electrical signals produced contain information for each pixel of the CIS or CCD photosensitive cell.
• the electrical information is generally translated into levels, digitized and stored by the storage means.
• the storage and storage phases, already contained in patent FR 2 854 476 entitled "device for counting stacked products", will not be described here.
• Each photosensitive cell CIS or CCD comprises, for example 10,000 photosensitive elements, to analyze the entire length of the stack (5) and allow the counting of a batch of products of, for example, a maximum of 1000 products.
• Each photosensitive element makes it possible to detect a light signal and to express this signal in the form of an electrical signal representative of at least 256 brightness levels. This signal for 256 levels of brightness is translated into 8-bit words, each word is stored in the device memory. Thus for the example given, the memory consists of 10,000 one-byte words.
• the photosensitive elements of the CIS or CCD photosensitive cells may be sensitive to rays of different colors and to their constitution by a combination of red, green and blue.
• the photosensitive cell is a matrix comprising for example 2000 photosensitive elements, for the analysis of the length, by 2000 photosensitive elements, for the analysis of the width.
• Simultaneous longitudinal analyzes are therefore possible along several longitudinal lines (T) of the stack (5), at different distances from a long edge of the stack (5).
• the analysis of the light rays reflected by the stack (5) is performed in two dimensions, unlike other embodiments in one dimension.
• the analysis performed in two dimensions allows several different longitudinal analyzes of the stack (5), the counting device being fixed, while the analysis in one dimension requires a displacement, for example of the stack (5), to perform several different longitudinal analyzes.
• the representative information for example of the brightness level, stored in memory in digital form is translated in the form of a graph, as illustrated by the curve (C1) of FIG. 8, and show variations in brightness.
• the graph has peaks representing maximums and troughs representing minimums of the signal from the electronic circuits associated with the photosensitive cells.
• the processing means (10) make it possible to analyze these variations by treating, for example, all the values taken in order according to their position. For example the rightmost pixel is processed, then the next one to the left and so on.
• a processing algorithm relies, for example, on the comparison of at least two successive values in order to determine the direction of variation of the curve.
• the s (n) signal (s) collected by a longitudinal analysis of the stack (5) of thin products (2) are recovered by the processing means (10) which then determine the repetition of a pattern (M) representing a product using a processing algorithm of a Noise signal.
• Fourier filtering is carried out beforehand to eliminate the hollows in the recovered signal, the noise elimination being able to be carried out just after for the counting signal reconstituted by an inverse Fourier transformation.
• the flow of the count is illustrated in FIG. 1.
• the counting method thus comprises:
• a step (51) of pretreatment of the recovered signal including a filtering of the signal to produce a filtered signal, the filtering preferably being a filtering performed on the Fourier transform (fast or otherwise) FFT of the signal with a comb filter;
• the method first comprises a step (50) of converting the signal, prior to filtering, into data representative of brightness levels in correlation with a stack thickness dimension expressed in pixels.
• the signaling step (54) comprises a display of a number of smart cards to be processed by a chip card personalization machine and / or a transmission of the representative information. from this number to the personalization machine.
• the aforementioned steps (50, 51, 52, 53, 54, 55) can be performed automatically on a computer connected to the detection means (8). All signal processing and calculations can be i performed by a program loaded directly into the computer memory and specifically used to count the number (N) of thin products (2).
• the shape of the pattern retained for a series of products (2) being counted can be estimated after a comparison between the first periodic pattern (M1) detected in the denoised signal and a reference pattern ( Mref) stored in the storage means.
• the estimation step (52) may make it possible to define the first periodic pattern (M1) representative of a thin product (2) with a possible phase shift, as illustrated in FIGS. 3C, 4A and 4B.
• the reference pattern (Mref) is used to perform the circular registration illustrated in Figure 5B, to obtain a second pattern (M2) estimated without phase shift.
• parameterization means associated with the processing means may be provided in the device to obtain the reference pattern (Mref) during a counting performed by the counting device with a standard batch of products ( 2) slightly thick.
• Other configuration modes for the reference pattern (Mref) can of course be used.
• the processing means (10) provide, for example, an intercorrelation signal (C2), as illustrated in FIG. 8, and pattern counting means (M2). ) in the denoised signal, by detecting the local maxima (S) of the intercorrelation signal (C2).
• the pretreatment step (51) can be performed as follows.
• the pretreatment step (51) may consist of denoising this signal s (n) to the maximum by filtering the frequencies that do not correspond to harmonics (comb filter).
• the steps of the filtering are for example the following ones: i) "Zero padding" method (filling of zeros) An addition of zeros is made at the end of the counting signal s (n) so that the number of samples N zp is a power of 2 (This is necessary to calculate the FFT transform).
• the recovered vector corresponds to an increased signal size and groups an even number N zp of signal samples.
• S 2p (") - a succession of decreasing peaks of height is observed
• the graph shown is produced for a thickness Ep of product (2) parameterized at 18 pixels and shows the modulus of the FFT transform as a function of normalized frequencies.
• the peaks represent the periodic character of the signal.
• the first peak (hO) is called the fundamental (or first harmonic) and the other peaks (hi, h 2 , . ) are called the harmonics iv) Frequency filtering
• the filtering is done by truncation of the FFT transform, as illustrated in Figure 2B.There are only the frequencies around the harmonics.
• the frequency bandwidth (p) of each bandwidth is This frequency width is denoted by 2 * b P.
• the filter thus obtained forms a comb filter
• the processing means (10) advantageously use this type of comb to eliminate by filtration. rage of noise and especially the frequencies do not not corresponding to harmonics. Frequencies distant from the harmonics and which may correspond to differences between the products (2) thin are eliminated.
• the means (10) for processing the device are provided with at least one program that makes it possible to memorize all the intermediate results obtained successively during the processing, for example using storage tables.
• the different calculation algorithms are respectively used by calculation modules arranged to retrieve the appropriate information (portions of the signal being processed, results of the previous operations, etc.).
• Comb filtering retaining at least the first three harmonics is necessary and sufficient.
• the signal x (n) is the sum of a noise w (n) and a useful signal y (n) composed of a repetition of motives word (n) representing the slice of a map.
• the signals y (n) and x (n) have the appearance represented by the respective traces (Sd, Sf) of the Figures 3A and 3B.
• the denoised signal plot (Sd) has a geometric character easily recognizable in the example of Figure 3A.
• the thickness of the card expressed in pixels may be denoted Ep. Its value is fixed arbitrarily at the beginning of the treatment. The estimation of the thickness can be made at the beginning of treatment using a first FFT: - Calculation of the FFT and its module.
• the processing means (10) perform an estimation of the Fourier transform (FT) of the pattern:
• Word (f) R m (f) e ' ⁇ n .
• the search for Word (f) takes place in two phases:
• the word pattern (n) will be easily calculable by an inverse Fourier transform.
• the processing means (10) then make it possible to estimate respectively the module and the argument of Mot (f).
• the processing can simply consist in carrying out the autocorrelation c ( ⁇ ) of the observed signal.
• c (r) ⁇ x ()) x (r +))
• r [0, Ep-I] (R3)
• the Fourier transform of the relation (R3) gives the module of Word (f):
• Theta B A.
• Theta M (R9) The value of matrix A depends only on Ep.
• the last row of matrix A of the system varies according to the parity of Ep.
• Theta M A 1 .
• Matrix A links the arguments of the bicorrelation (correlation to a higher order) to the arguments ( ⁇ m ) of the pattern.
• the means (10) for processing the counting device make it possible to perform a circular registration to eliminate any phase shifts. Indeed, in many cases, the estimation of the pattern by the calculations described above is not yet satisfactory.
• the algorithm used to search for the pattern will give the estimate of the pattern (M1) as shown in Figure 4B. A phase shift is apparent.
• the estimate is a correct estimate of the pattern (M2) to a pure phase shift.
• a reference pattern (mref) is used for the estimated pattern to be correct.
• the reference pattern (Mref) may for example have the shape shown in Figure 5A, inverted U-shaped (here three segments).
• FIG. 5B An example of additional processing applied to the pattern obtained in FIG. 4B is illustrated in FIG. 5B.
• the registration may consist of applying different phase shifts to the reconstituted periodic pattern (M1), until the pattern (M2) resembles the reference pattern (Mref).
• M1 the reconstituted periodic pattern
• M2 the pattern
• Mref the reference pattern
• FIG. 5B shows that the scalar products found (from top to bottom) are respectively:
• Product_Scalary (Motif, MotifRef) 0.7
• the pattern (M2) obtained after registration then corresponds to a correct estimate of the pattern of a card or similar portable object, thin.
• FIG. 6 summarizes the processing method implemented to make it possible to estimate a trace in the signal representative of a thin product (2).
• Signal x (n) repeatedly contains the word pattern (n) whose Fourier transform can be expressed as r (n) e l ⁇ (n) .
• f the module and the argument of Word
• Mref the reference pattern
• the counting is done by calculating the cross correlation / ( «) between the estimated motive word (k) (of size Ep) and the denoised signal x (k) (of size N).
• This step also called adapted filter, is carried out as follows:
• the counting is done by detecting the local maxima (S) or vertices of the intercorrelation signal (C2), as shown in FIG. 8. Having a preprocessed signal x (k) makes it possible to establish an exact count, without risk of error due to a small spacing between two consecutive products (2).
• the device is composed of a CIS module (3) projecting a beam of light rays (6).
• the light rays (6) are projected on the stack (5) of thin elements (2), contained in the tray (4), in a longitudinal direction, forming a line (T) light on the stack (5).
• the device may comprise three CIS modules combined so that the light rays and the modules 3a, 3b, 3c cover the entire length of the stack (5).
• the CIS modules are for example placed so that a part of the treated areas overlap.
• the modules can be inclined so that the illuminated areas are aligned. Two of the modules can be inclined at a given acute angle to the vertical and the other module can be inclined at an acute angle to the vertical. In this case, the modules are inclined so that the intersection of the plane light beams with the stack (5) forms a line (T) light.
• the CIS modules are not inclined, the longitudinal analysis being performed according to several segments whose sum of lengths is at least equal to that of the stack (5).
• An initialization phase makes it possible to determine the relative positions of the CIS modules.
• the device comprises only one CIS module (3d) which moves relative to the stack (5) in several positions (PO1, PO2, PO3) in a longitudinal direction.
• This module (3d) traverses the entire length of the stack (5) after several displacements and several stops at given positions (PO1, PO2, PO3) in order to treat, each time, an area (ZO1, ZO2 , ZO3) of the stack (5).
• the different positions (PO1, PO2, PO3) are chosen so that each zone partially overlaps the adjacent zone.
• the processing means identifies the signals corresponding to the overlap and eliminates the duplicate signal portion.
• a calibration step concerning overlapping areas is also described in patent FR 2 854 476 in order to efficiently process duplicate data.
• the relative displacement of the module (s) (3) CIS with respect to the tray (4) is realized, according to one embodiment, by a transverse displacement ( M4a) of the tray, relative to the longitudinal direction of the lighting, the module or modules (3) being fixed. In another embodiment, this same relative displacement is achieved by a transverse displacement (M3a) of the CIS module (s) (3), the tray (4) being fixed. In the embodiment of FIG. 7B, the relative displacements are in a transverse or longitudinal direction.
• a relative longitudinal displacement is carried out parallel to the longitudinal illumination in order to position the module (3d) CIS above the various zones of the tray (4), this displacement (M4b, respectively M3b) being achieved either by moving the tray ( 4), the module (3d) CIS being fixed, either by moving the module (3d) CIS, the tray (4) being fixed.
• a possible displacement (M3a, respectively M4a) relative transverse module (3d) CIS relative to the tray (4) is performed, for example, perpendicular to the longitudinal illumination.
• the transverse displacements (M3a, respectively M4a) relative to the module (s) relative to the tray (4) involves several longitudinal analyzes according to different longitudinal zones of the stack (5).
• Figures 9, 10 and 12 illustrate the use of a camera (8), for example matrix or linear CCD type.
• the camera (8) CCD is associated, but not limited to two mirrors (9a, 9b) and means (7) lighting.
• This type of device is detailed in patent FR 2 718 550.
• the sensitive photo sensor is, for example, linear and allows the longitudinal analysis according to a feature (T).
• the associated lighting means are, for example, a fluorescent tube or diodes whose illumination rays are focused or not.
• Several longitudinal analyzes are, for example, made according to the same line (T) with different illumination intensities.
• several longitudinal analyzes are, for example, performed according to different features (T1, T2, T3), by a relative displacement of the stack (5) relative to the camera (8) CCD and the device lighting.
• the means (7) of illumination for example, made by diodes whose rays are, according to a nonlimiting example, focused by an optical device, and requires transverse relative displacements, in order to perform several different longitudinal analyzes.
• the lighting means is formed by a fluorescent tube (7)
• the entire upper surface of the stack (5) is illuminated, but with different intensities.
• the area closest to the tube is illuminated at a higher light intensity than the more distant areas.
• This type of illumination of variable intensities is combined or not to relative transverse displacements to achieve different longitudinal analyzes according to different features (T1, T2, T3) longitudinal, with different light intensities.
• One variant comprises the variation of the luminous intensity obtained by controlling the lighting means, according to a variable power.
• the means (8, 9a, 9b) of detection are fixed and the tray (4) is movable (M4a), the tray (4) is fixed and the means (9a, 9b , 8) are movable at least in part, the mirrors (9a, 9b) and / or the camera (8) CCD being movable.
• the photosensitive sensor of the camera (8) CCD is matrix. This type of photosensitive sensor allows analysis in two dimensions, depending on the length and width of the stack (5). In the case of a matrix photosensitive sensor, transverse displacements are not necessary to perform several longitudinal analyzes.
• the camera (8) CCD analyzes, for example, the entire length of the stack (5), as shown in Figure 9, where the stack (5) is analyzed over its entire length with a longitudinal displacement (M8) of the camera (8) CCD.
• Several lines, covering the entire length of the stack (5) are analyzed, the lines being very close, or even glued, at a distance for example of 5/100 of a centimeter or more distant at a distance, for example from one or several millimeters.
• the lines (T, T1, T2, T3) analyzed are also illuminated according to different light intensities.
• the elements or products (2) thin are stacked in a tray (4) and are set so as to have the long edge to the top of the tray (4).
• the products (2) to be counted are arranged side by side, without limitation a front face of a product against a reverse side of another product.
• Figures 7 to 11 show a view of products (2) thin stacked side by side, the tray (4) being shown under the stack (5).
• the products (2) thin are therefore placed on their edge, oriented transversely in the tray (4), that is to say parallel to the short sides of the tray (4) rectangular.
• a stack contains up to 500 cards.
• the counting device detects the slice of each product (2) and thus determines the number (N) of products.
• An example of data processing is the detection of the variation of brightness.
• the data translated in the form of a graph represent the brightness as a function of the position.
• a maximum will be the value of an electrical signal corresponding to a received high intensity light signal, with respect to the adjacent signals.
• a minimum will be the value of an electrical signal corresponding to a light signal received of low intensity, relative to adjacent signals.
• a maximum can be interpreted by the treatment program as the medium of a product (2) to be counted and a minimum is interpreted as the joining of two products (2) to count. The junction between two products (2) thin is indeed darker and the middle of a thin element is lighter.
• the counting device can indicate the number of thin products (2) in a series. Thanks to the memorization of information provided by the operator, concerning the nature of the products, the device associates with each series the nature of the products.
• another processing system downstream of the processing chain receives data specifying the nature of each product (2) and can therefore determine the customization or verifications to be performed.
• the downstream processing system communicates with the processing means of the counting device by means of communication, in a known manner.
• the communication means comprise, for example, a wired link or infrared or radio wave and communication interfaces adapted to the type of connection.
• the communication means are mediums, such as diskettes or disks, associated with readers of these mediums.
• This treatment is therefore done automatically, directly by inserting the tray or the magazine containing the stack (5) into the processing system, or by transferring the stack (5) to another support.
• a check can be made by comparing the number (N) found by the device for the products of the complete stack (5), with a number of products provided by a product series management device (2).
• the number of products (2) in each series is deduced from these results.
• the operator knows the nature of each small series comprising the stack and thus determines the nature of each product (2) at a given position.
• the entire stack can be processed directly, information additional nature of the series that can advantageously be provided to the personalization machine.
• the personalization machine has processed a total of N elements, the processing performed being a function of their position in the stack (5).
• An alternative embodiment, as shown in FIG. 11, comprises at least one transverse CIS module (3t) producing a transverse illumination, for example perpendicular to the longitudinal direction of the stack (5).
• the module (3t) transverse CIS comprises detection means and lighting means in a transverse plane beam which illuminates transversely the stack (5).
• the module (3t) transverse CIS placed vis-à-vis the stack (5) performs the analysis of the illuminated transverse linear zone.
• the analysis of the entire length of the stack (5) is performed by a displacement (M3t) of the transverse module, in the longitudinal direction of the stack (5).
• the longitudinal displacement (M3t) of the transverse CIS module (3t) is performed at a determined speed.
• the photosensitive cells of the transverse module transform the light energy of the rays reflected by the stack (5) and focused on photosensitive cells of the detection means, into electrical signals which are the image of the light intensity.
• the processing means of the counting device sample these signals and convert the analog values of the electrical signals into image computer codes of these analog values, placed in the storage means.
• the transverse CIS module has covered an area comprising the entire length of the stack (5) by its lighting means associated with its detection means, the stack (5) has been analyzed over its entire length and over a zone. of determined width.
• the two-dimensional analysis thus makes it possible to perform several longitudinal analyzes on the stack (5).
• the longitudinal analyzes are performed along lines (T1, T2) near or far (T1, T3) of several millimeters.
• FFT Fast Fourier Transform

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=38743572

Family Applications (1)

Country Status (7)

Families Citing this family (11)

Family Cites Families (5)

• 2007
  • 2007-04-26 FR FR0703031A patent/FR2915601B1/fr not_active Expired - Fee Related
• 2008
  • 2008-04-23 EP EP08805504A patent/EP2145293A2/fr not_active Withdrawn
  • 2008-04-23 WO PCT/FR2008/000585 patent/WO2008145859A2/fr active Application Filing
  • 2008-04-23 KR KR1020097024666A patent/KR20100040700A/ko not_active Application Discontinuation
  • 2008-04-23 AU AU2008257340A patent/AU2008257340A1/en not_active Abandoned
  • 2008-04-23 CN CN200880020797A patent/CN101816012A/zh active Pending
  • 2008-04-23 US US12/597,678 patent/US20100226576A1/en not_active Abandoned

Non-Patent Citations (1)

Also Published As

Similar Documents

Legal Events