WO2016082055A1 - Smart label with static and dynamic information fields - Google Patents

Abstract

The invention relates to a smart label device, to the method for writing information thereto, to the method for reading the information contained in the label and to a system associated therewith. The label is a visual representation of a matrix of two-dimensional data, containing information units (smallest unit that makes up the matrix of two-dimensional data) exhibiting different states.

Description

 SMART LABEL WITH STATIC INFORMATION FIELDS AND

 DYNAMICS

FIELD OF APPLICATION The present invention corresponds to a smart tag type device, the method of writing information therein, the method of reading the information contained in the tag and a system associated therewith.

The label corresponds to the visual representation of a two-dimensional data matrix, containing information units (smaller unit that makes up the two-dimensional data matrix) that have different states. Such information units may have both static and dynamic behaviors. A dynamic behavior refers, in the present invention, to the ability of an information unit to vary its state or mutate with respect to another variable, while a static behavior refers to an immutable state in time. The information units that have dynamic behavior (dynamic information units) can be interpreted as sensors according to what is indicated in the method of reading information of the present invention, in order to gather information of environmental variables and / or of interest : temperature, UV radiation, pressure, etc., presenting dynamic information. All or part of the dynamic information units can be used in conjunction with the information units that have static behavior (static information units) to encode static information on the label. The dynamic information, as well as the static information, can be read from the label by a recognition system (SR) according to the reading method of the present invention. Part or all of the dynamic information can be read visually by the user without the need for an SR. BACKGROUND

DESCRIPTION OF THE STATE OF THE TECHNIQUE

In the state of the art there are multiple types of labels with information encoded using points, bar codes, QR codes, figures, etc. This information can be static or dynamic depending on the application and use of the label.

The static information is printed on the label and commonly corresponds to information associated with the manufacture, price, origin or destination of the product in which it will be incorporated.

The dynamic information varies over time as some sensor element changes state according to the condition to be measured. As a result of this change in the sensor, the label modifies its appearance by changing the color of an icon, figure, bar, etc., and this change may be reversible or irreversible.

There are currently labels that meet the condition of containing static and dynamic information. Document US2013 / 0048736 describes a label consisting of two components. The first component corresponds to a surface that contains information that is invariable over time (static information) and the second component corresponds to a portion of the label that contains information that It varies over time. For example, it can be seen in Figure 1 (a), a first component of the label consisting of a barcode and a second component that is a tape that changes color as time passes, which can be associated with the expiration time of the product in which the label is inserted. Similarly, Figure 1 (b) shows a label in which the barcode contains static information and the bottom of the label is variable in time.

The document label US2013 / 0048736 is composed of two independent labels, located on the same base, because the upper part containing the bar code and the variable lower part can be separated, maintaining each one full functionality.

JP200181356 describes a price tag that is provided with a first barcode that shows data of the non-variable type in which no alteration occurs after it is issued, and a second barcode that shows a set of data of variable type in which an alteration factor occurs after its emission, as seen in Figure 1 (c). Also in this case the label consists of two or more labels incorporated in the same base, under the same argument that the dynamic part and the static part separately maintain full functionality.

Messrs. Shimizu, Isami and others [Reference 2], investigated the use of colors other than black and white in the construction of two-dimensional codes. They proposed a method to improve the efficiency in the classification by the recognition system, with respect to the colors used in the coding. its The method consists of including a set of colors (seed colors) in the two-dimensional code, which are not necessarily equal to the colors used in the coding. As input of the classification system (features) they used the RGB values of the color to be classified, together with the RGB values corresponding to the seed colors. Specifically, they carried out experiments considering as coding colors the 64 colors generated with the combinations of RGB values [0.85, 170,255], and as seed colors they used the 8 colors generated with the combinations of RGB values [0,255]. They did not consider the incorporation of dynamic elements in the label used nor the use of different color states. The state of the art does not describe a label that incorporates static and dynamic information in a single unit, where part or all of the dynamic information can be read with the naked eye. The state of the art does not describe a label that contains static as well as dynamic elements, where dynamic elements can be interpreted as sensors, and whose method of reading / writing information considers elements that exist in more than two states.

SUMMARY

The present invention corresponds to a smart tag type device, the method of writing information therein, the method of reading the information contained in the tag and a system associated therewith.

The label corresponds to the visual representation of a two-dimensional data matrix, containing units of information (unit smaller than it forms the matrix of two-dimensional data) that present different states. Such information units may have both static and dynamic behaviors. A dynamic behavior refers, in the present invention, to the ability of an information unit to vary its state or mutate with respect to another variable, while a static behavior refers to an immutable state in time. The information units that have dynamic behavior (dynamic information units) can be interpreted as sensors according to what is indicated in the method of reading information of the present invention, in order to gather information of environmental variables and / or of interest : temperature, UV radiation, pressure, etc., presenting dynamic information. All or part of the dynamic information units can be used in conjunction with the information units that have static behavior (static information units) to encode static information on the label. The dynamic information, as well as the static information, can be read from the label by a recognition system (SR) according to the reading method of the present invention. Part or all of the dynamic information can be read visually by the user without the need for an SR.

Both the units of information that exhibit dynamic behavior and those that exhibit static behavior can have more than 2 usable states by the present invention, allowing to store more information per unit of space (for example in comparison to binary codes).

Static information corresponds to invariable information. Part of its uses are: container of an identification number or part of a mechanism that allows implement different levels of access to information (both static and dynamic) for the different users of the label. The coding of the static information may include redundancy information as an error correction method, according to the method of writing the information of the present invention. The dynamic information units of the label, used to gather dynamic information, are added to the label under a predefined structure for each configuration (for example: product, product type, family of products or variables of interest to be measured).

The combination of static and dynamic information allows information about the origin, history and current state of the product to be stored.

In addition to the information contained in the static and dynamic information units, it is optionally possible to include an electronic system for storing and transmitting information, which acts as a backup for the analog part of the label. The electronic system can use wireless (RFID, NFC, WIFI, BLUETOOTH, etc.) or contact technology for sending / receiving information, which would allow the device to be integrated into automatic tracking and / or logistics systems that they consider electronic and automation components.

The analog information (both static and dynamic) contained in the label can be read from an image of it obtained from a capture device (scanner, digital camera, etc.), using a recognition system (SR) of according to the reading method of the present invention. The information read from the label can be used to update an external database, and can be incorporated into a logistic and / or quality assurance system, from its manufacture to the end user.

BRIEF DESCRIPTION OF THE FIGURES

Figure 1: Examples of prior art.

 Figure 2: Label design according to the present invention.

 Figure 3: Label barcode.

 Figure 4: Coding used in the elaboration of the bar code.

Figure 5: Example of distorted quadrilateral.

 Figure 6: Example 1: Coding using two possible states for the information units.

 Figure 7: Example 1: Coding using four possible states for the information units.

Figure 8: Example 2: Base case.

 Figure 9: Example 2: Adaptation of the shape of the label in order to accommodate additional sensor.

 Figure 10: Example 3: Coding of compressed information.

 Figure 1 1: Example 4: Coding of encrypted and compressed information Figure 12: Example 5: Label with a visually interpretable sensor.

 Figure 13: Example 5: Sensor detail.

 Figure 14: Example 5: Sensor states.

Figure 15: Example 6: Label with visually interpretable sensors. Figure 16: Example 6: Possible states and sensor arrangements.

 Figure 17: Example 6: Example instructions for visual label reading.

 Figure 18: Example 7: Label including colorimetric temperature sensor. Figure 19: Example 7: Photograph of the label to be analyzed by the image processing system.

 Figure 20: Example 7: Subsectors of the image to be analyzed by means of the corner detector algorithm.

Figure 21: Example 7: Results of corner detector algorithm.

Figure 22: Example 7: Corner classification.

 Figure 23: Example 7: Image after spatial transformation and transformation to black and white.

 Figure 24: Example 7: Examples of regions of adjacent pixels and their respective bounding box.

Figure 25: Example 7: Zone of adjacent pixels with the highest height among those detected.

 Figure 26: Example 7: Barcode.

 Figure 27: Example 7: Regions with lower value in the size pattern detection function.

Figure 28: Example 8: Integration of the label into the logistics and product quality assurance system.

Figure 29: Types of information contained in the label.

Figure 30: Coding scheme. Figure 31: Decoding scheme.

 Figure 32: System scheme

 Figure 33: Color space color diagram xyY.

 Figure 34: Density of indicator data distribution a.

 Figure 35: Density of indicator data distribution b.

 Figure 36: Cluster projections in the CIELAB color space.

 Figure 37: Groups in the xyY color space.

 Figure 38: Groups 1 to 8 xyY color space.

 Figure 39: Groups 9 to 16 xyY color space.

 Figure 40: Evolution of the xyY color space solutions.

 Figure 41: Colorimetric transformation for the red, green and blue channels

Figure 42: Measured temperature versus estimated temperature.

 Figure 43: Cumulative distribution of the absolute error.

LIST OF ACRONYMS AND ACRONYMS

Listed below are the acronyms and / or acronyms used in this document: UV: Ultraviolet electromagnetic spectrum.

 RFID: Radio frequency identification, of the English Radio Frecuency Identification. NFC: Near Field Communication, from Near Field Communication. QR: Quick Response Code Two-Dimensional Code. ASCII: American Standard Code for Information Exchange, from the American Standard Code for Information Interchange.

UPC: Universal Product Code, English Universal Product Code. GF: Galois Field.

 RS: Reed-Solomon.

 LZW: Lempel-Ziv-Welch.

 TLC: Thermochromic Liquid Crystal, from English Thermochromic Liquid Crystal.

SR: Recognition system.

REFERENCES

[1] Reed, Irving S .; Solomon, Gustave (1960), "Polynomial Codes over Certain Finite Fields", Journal of the Society for Industrial and Applied Mathematics (SIAM) 8 (2): 300-304. [2] Color Recognition by Extended Color Space Method for 64-color 2-D Barcode. Takuma Shimizu, Mariko Isami and others, MVA2011 IAPR Conference on Machine Vision Applications, June 13-15, 201 1 Nara, Japan.

[3] ISO / IEC 15420: 2009 Information technology - Automatic Identification and data capture techniques ~ EAN / UPC bar code symbology specification. [4] High speed data compression and decompression apparatus and method, Welch, T.A., US4558302, 1985/12/10 /.

[5] A method for the construction of minimum-redundancy codes. David A. Huffmann. Preceedings of the IRE DETAILED DESCRIPTION OF THE INVENTION

The present invention corresponds to a smart tag type device, the method of writing information therein, the method of reading the information contained in the tag and a system associated therewith. The tag has information units (elements) that have static behavior and information units (elements) that have dynamic behavior.

The static information is that included at the beginning of the product tracking, according to the coding (writing) method of the present invention. Other uses are as a container of an identification number, or part of a mechanism that would allow different levels of access to information (both static and dynamic) to be implemented for the different users of the label.

The dynamic information corresponds to that presented by the dynamic information units included in the label, which are interpreted as sensors by the SR, in order to obtain information on variables such as temperature, UV radiation, time (useful life), etc.

Additionally, the tag can integrate wireless technology for the transmission and reception of data and / or together with an electronic information storage unit that is updated based on the information retrieved from the analog part (in particular the dynamic part) of the tag . The information units present in the device do not require energy from the same label for their state changes. The label allows different levels and alternatives of access to information: simple reading by the user based on a manual or indications attached to the product, reading by an SR, according to the method of reading of the present invention, from the image obtained by means of an image capture device (scanner, digital camera, etc.) or through wireless communication means (RFID, NFC, WIFI, BLUETOOTH or others). Another way to implement levels of access to information is by encrypting all or part of the information, to which only a fraction of the users would have access.

The label considers the incorporation, if necessary for the proper functioning of the SR, of information units intended for state calibration (which in its preferred execution corresponds to a metric colorimetric calibration) and spatial, which allow the recovery of information in a multitude of situations or reading conditions.

All of the above allows the device to be incorporated into a management and logistics system, managing to track the entire production process and distribution of a product. The monitoring does not require electrical energy for its operation (since it does not require the energy information units coming from the same label), being able to recover the information stored in analogue way using the SR at any point of the path or process and record the information of the moment at the time that said information was read by the system, either in an internal memory or in an external database for later access.

The invention has the following characteristics: Device tag: a. The tag has static and dynamic information, corresponding to the static and dynamic information units of a two-dimensional data matrix. b. Part or all of the dynamic information can be read by the user with the naked eye without the need for an SR. C. The analog information contained in the tag (both static and dynamic), can be read using a recognition system, according to the reading method of the present invention. d. The information units of the label can present more than two states, according to the amount of information to be included in the label and to the states that the SR is able to recognize in practice. and. The shape of the label is configurable, its preferred execution being square or rectangular. F. It allows to include an electronic means of storage / transmission of information. g. The means of transmitting the information stored electronically on the label may be of the wireless type. h. The dynamic information units correspond to one or a set of materials that in the case of experiencing changes recognizable by the SR or Visually, depending on one or more environmental or interest variables, they can be interpreted as sensors.

Regarding the coding method (writing): i. The coding method can use units of static, dynamic or joint information (depending on the configuration). j. Part or all of the dynamic information can be configured to be interpreted visually. k. Some or all of the information can be configured to be interpreted only by the recognition system. I. Static label information may include redundancy information as an error correction method, according to the writing method and information reading method of the present invention, allowing the correct information to be retrieved, even though a fraction of She is read incorrectly. m. The static information may be subjected to a compression process before being added to the label, according to the methods of the present invention. n. The static information can be subjected to an encryption process before being added to the label, according to the methods of the present invention.

Regarding decoding method (reading): or. It can be automatic, using a recognition system that delivers information as a result of the interpretation of a label image. p. It can be manual (visual) using an instruction attached to the label or intuitively to interpret part or all of the dynamic information units configured for such purposes. q. You can recognize and interpret the levels of information present on the label. r. Can recognize and interpret the existence of information compression. s. You can recognize and interpret the existence of information encryption. t. Some or all of the information read by the recognition system or by visual interpretation can be sent to a remote or local storage system manually or automatically.

Regarding the system: u. The logistics and / or product quality assurance system, as a result of the use of the label, may have different configurations.

1. LABEL DEVICE

The technologies and methods that support the characteristics of the label described above are described below. A. Static device information

The static information of the label is included in it by a particular arrangement of the static information units therein, in some cases also using dynamic information units. The interpretation of these elements is done through the SR.

A particular type of information corresponds to the recognition and calibration patterns, described below.

Recognition and calibration patterns

At the time of decoding the label, it is necessary that the SR be able to determine its size, orientation and configuration, which is why a set of recognition patterns is included in it.

Figure 2 shows pattern 201, which allows, during the analysis of an image of the label by the SR, to determine the total size of the data matrix in units of pixels, since it covers the entire first column of the label. In addition, pattern 201 contains additional information in the form of a barcode, constituting a double use thereof. The barcode used corresponds to an adaptation of the UPC code, where the digits are coded as a set of four bars of variable width, with alternating color between black and white (Figure 4). In the present invention, said pattern can be encoded with two or more states. Pattern 201, of which an example is shown in Figure 3, includes a starting pattern corresponding to three bars (black-white-black) (301), a two-bar term pattern (white-black) (302) and a check digit dependent on The coded digits. Unlike the UPC code, here the code "L" is used in all the digits.

The pattern 205 is of a known size in units of information, allowing to determine the size of the image of the data matrix in units of information by comparison with the size of the pattern 201 in pixels. The size of pattern 205 may vary, depending on aspects such as the size of the data matrix in units of information or the expected resolution in the image of the label to be analyzed by the SR. The size of the pattern 205 is such that errors and normal distortions in the determination of its size in pixels from the image analyzed by the SR do not result in errors in the determination of the size of the data matrix in units of information.

Pattern 204 is used in the spatial transformation stage (3. DECODING METHOD Section A Decoding by means of the recognition system), as it is indicative of the upper right corner of the data matrix.

Another type of pattern corresponds to the state calibration patterns, 202 and 203, respectively, according to the nature of the dynamic information units of the device and the configuration of the states of the static information units (indicated in sections B Dynamic device information and C Configuration of the device information units). Depending on the situation, these zones can be merged into a single zone or eliminated, depending on the method of reading to be used. Static information contained in the recognition patterns

The adaptation of the UPC coding used in the generation of pattern 201 allows an arbitrary number of digits to be included. In an implementation of the invention, pattern 201 is used to store information relative to the size of the data matrix, together with a unique identification number for each product. For this, 4 digits corresponding to the identification number of the label are concatenated, together with two digits corresponding to the size of the same in information units (in the rectangular case 4 digits are included), in addition a verifying digit is calculated according to:

where:

This set of 7 digits is transformed into an image according to the coding shown in Figure 4, also incorporating the start and end pattern, together with an outline that facilitates recognition according to the reading method of the present invention. . It should be noted that any set of digits can be encoded in this way, being able to incorporate different and / or additional information to the size of the matrix and / or the identification number of the label.

Pattern 205 is used in a complementary method of determining the size of the data matrix, which is based on the fact that it is of a size known in information units, its preferred execution being the size of 3x3 information units. Because the pattern 205 is required to be isolated from the rest of the figure, a white safety edge is incorporated, which constitutes a use of space of 4x4 information units. As indicated above, as the size of the data matrix increases, it may be necessary to change the size of this pattern.

B. Dynamic device information

The dynamic information units of the label, once interpreted visually or through the SR, allow measuring and / or quantifying environmental and / or interest variables. These information units are generated using materials that have thermodynamic, optical, chemical, pressure dependent properties, etc., variables with the magnitude to be quantified, and that do not require energy from the label itself as a power source for its operation.

The type of material to use depends on the variable you want to measure. The material is required to undergo a detectable change due to changes in the value, presence or absence of the measured variable. For example, if the variable is temperature, thermochromic or thermosensitive materials can be used. On the other hand, if the variable of interest is UV radiation, the required compound should be photosensitive or sensitive to light, so it must undergo a detectable change by varying the lighting to which it is subjected. Within the variety of materials with potential for use, those of colorimetric type stand out, which undergo a visible change to the human eye due to changes in some specific variable. These changes may occur as the variation of a range of colors, for example in the TLC, or materials that undergo discoloration or coloration before certain threshold values (photosensitive materials), due to phase changes in the reversible (hysteresis) or irreversible material. Among the materials considered are those that have colorimetric changes, which will be used in the preferred embodiment of the present invention, and can be classified as monochromatic or polychromatic.

Monochromatic materials

These materials have the particularity of experiencing a reversible or irreversible color change, as the case may be, in the presence and / or absence of a certain variable that causes a phase change. For example, the reference of sensitive materials in the range of the ultraviolet spectrum could be taken, where if a certain threshold value is reached the material will show a change that can be from the decrease in hue to becoming translucent, or from a translucent phase until you reach a certain color in steady state. Therefore, a colorimetric calibration, necessary for the interpretation of the state of the material as a sensor, in this case focuses on the amount of color rather than the differences in color.

Polychromatic materials

Polychromatic materials can have a range of colors or phase changes associated with a certain level of excitation of some variable on which their phase change depends. The complexity of the analysis for this case is greater than for materials Monochromatic, this is due to the complexity of analyzing n phases and (n-1) transitions.

General methodology that relates the behavior of a material and a variable of interest.

To transform the behavior of the materials into recognizable states by the SR of the information units, the following method is proposed. This method relates the behavior of the material with a variable of interest allowing its use as a sensor.

Application of methodology for the use of calibration colors in conjunction with monochromatic and polychromatic materials

1. Analysis of the behavior of the material

The analysis of the behavior of a material that presents reversible or irreversible phase changes will depend on whether it belongs to the monochromatic or polychromatic type and the variables to which the phase changes will be associated.

The first step is to identify the behavior of the material within the range of use. This is achieved by acquiring images or frames of the material associated with different values of the variable to be measured.

If the variable were temperature and the temperature utilization range between T ₀ and Τ ₁ , where Τ ₁ > T ₀ , adequate sampling should be established to obtain greater precision in the analysis of its behavior. For example the acquisition of an image every (Τ ₁ - T ₀ ) / N; N being the quantity of samples, or it could also be a video of the material with an equal step a frame {framé). It is important that there is a coherence between the information obtained from the measured variable and the equivalent image. It is worth mentioning that to improve the data acquisition process, reference colors are incorporated for subsequent color correction. 2. Color correction

Color correction is intended to normalize colors / image in different situations of light and shadow. For this, the colorimetric transformation methodology is used, which is detailed in section 3. A.4. Figure 41 shows the curves of a TLC material with respect to temperature, in addition the curve obtained after applying a second order colorimetric transformation is included.

3. Choice of the method of estimation of the variable of interest

Once the experimental curves of the behavior of the material with respect to the variable of interest are known, the method of estimation of the variable of interest is chosen based on the information related to the material. There is a wide variety of methods for estimating variables, such as: least squares, soft computing methods such as neural networks, fuzzy logic, among others. Since there is no single solution to relate response or change of phases for all materials with respect to a variable, several solutions can be evaluated and determine which one is best suited for each material.

4. Selection and extraction of patterns Pattern selection is a process that helps decrease the amount of data to be processed and increase the response speed of the method of estimating the variable of interest. Pattern is understood as the characteristics or properties of a data set. The number of patterns is not only limited to the values obtained from the RGB model, but to all possible color transformations to models and color spaces, statistical values, texture, invariant moments, etc. The objective is to choose from all the possible patterns are those that contribute the most to the estimated model according to the material and variable selected. There are several methods of pattern selection: based on feature filtering regardless of the estimator (filtering); based on estimator performance (wrapping); and based on estimator training (embedding), there may be some mixed and other families of methods not mentioned. The decrease in the dimensionality of the characteristics eliminates the redundancy data and increases the performance of the system.

5. Training, error and evaluation of the method.

Training, error and evaluation of the method will depend on the estimation method to be implemented. An implementation of the method is set forth in Example 9.

C. Configuration of the device information units Each device information unit may have more than one state, with the requirement that each state be distinguishable by the SR at the time of reading. In the case of configuring information units for visual interpretation, the states of these units must also be distinguishable by the user. The number of available states is directly linked to the amount of information stored within a unit of information. For the binary case, the representation of two states is called bit: one and zero; These states can be represented (in print) as the existence or absence of a color, usually black is usually used on a white background, and can be any other.

To increase the density of information, a greater number of possible states is incorporated for each unit of information. By increasing the number of states, the binary bit can no longer be referred to, since the base of the system would no longer be two. If the encoding were only the ASCI 1 table (256), it is necessary to store the 256 characters 8 bits (2 ⁸ = 256), on the other hand if you wanted to store the ASCII table (128) the requirement would be 7 bits (2 ⁷ = 128). However, for the storage of the ASCII table (256) for a base greater than two, when using bases 4, 16 and 256, the amount of information units required to represent a character has to be reduced while increasing the density of information on the label. In order not to confuse the concept of bit (two states), or qdits (D-dimensions in Hilbert space) in quantum computing, the vEbit unit is defined in this invention. The vEbit unit aims to explicitly define that the number of states can be variable and the letter E represents the number of states used, with a v2bit being the equivalent of one bit.

Once the unit of information greater than two states is defined, it is necessary to encode the ASCII table (256) of 8 v2bits (2 ⁸ = 256), 4 v4bits (4 ⁴ = 256), 2 v16bits (16 ² = 256 ), or 1 v256bit (256 ¹ = 256); for 2, 4, 16 and 256 states, respectively. This methodology, which consists in the use of a greater number of states in order to increase the density of information, can be used not only for the ASCII code (256), but for any symbol table. Being the amount of states dependent on the technological feasibility and sensitivity of the SR. Representation of States:

Each unit of information on the device can be configured in more than one state, with the requirement that each state be distinguishable by the SR at the time of reading. In the case of configuring information units for visual interpretation, the states of these units must also be visible. In this way any type of measurable property can be used, with different levels representing different states, for example: colors, textures, magnetic properties, etc. It should be noted that the combination of states of different origins, such as color / texture or other selection of states, may also be included, being a mixture of two or more, as the case may be. In the state of the art, color has been used in the case of coding information in two-dimensional codes, using two states, with the black indicating a state and white the complementary state.

The use of more than two states per unit of information is aimed at increasing the density of information within the same physical space. On the other hand, the selection of states is obtained by calculating perceived distances for a given tolerance. Use of colors

The use of solid colors, as a representation of two or more states, is one of the possible options in increasing the density of information stored in the data matrix. The increase in information density is generally associated with a decrease in the identification capacity of each state.

The amount of colors to use depends on the vEbits in which the coding is performed, but the choice of colors (states) will directly influence the ability to identify each state, associated with an error by state identification. Due to physical and technological limitations, the colors chosen may not necessarily be achieved or reproduced by printing, optical and display devices, such as printers, cameras, scanners, and monitors. In addition to this, the error produced from propagation from one device to another should be considered, for example: a given color is reproduced by means of a monitor, then this digital color is reproduced analogically with a printer, then scanned by a scanner returning again analog to digital information, to finally display it again with the same monitor. Comparing each result shows the existence of an error in the gamut or color gamut.

On the other hand, in addition to the errors related to color reproduction, it should be noted that the colors chosen must satisfy the condition of maximum color difference between them. In addition to meeting the boundary conditions according to be the case, as indicated by the segmented line of the selected region Ω in Figure 33.

The preferred execution begins by finding the common area between the color reproduction devices in the xyY color space, for this purpose the chromatic diagram of the xyY color space is used as indicated in Figure 33. The selected region can be found by restriction equations that form the boundary of the resulting polygon, or by evaluating membership in the region.

To evaluate the membership of the colors to the selected region, an array is generated that contains all the colors belonging to a finite color model. If the RGB color model is used, the maximum number of colors is 256 ³ = 16,777,216, then a transformation is made to the XYZ color space, to then obtain the coordinates in the xyY color space, the XYZ to xyY color space transformation equations are shown below:

Once defined the membership of the points to the selected region, a clustering algorithm is used, in this case K-means, using a distance relative to color as a distance metric, implementing the color distance CIEDE2000 (ΔΕ ₀₀ ). It should be mentioned that the color distance ΔΕ ₀₀ , has a better response in the differentiation of colors than the typical distances used in the K-means algorithm, for example Euclidean distance.

The objective function is to maximize all distances ΔΕ ₀₀ between the K colors chosen. The clustering algorithm manages to determine the set of data closest to a given initial solution μ ⁽⁰⁾ , updating the value of the iteration to iteration solution until its convergence (if it exists) or up to a predefined number of iterations. The distances are calculated in the CIELAB color model, the set of data belonging to the selected region Ω (see Figure 33) in CIELAB coordinates is referred to as x, and C¡ ^(z) corresponds to the elements belonging to ax which they are closer to the element μ¡ ^(z) , of the solution in iteration z, with respect to the distance ΔΕ ₀₀ . N being the number of points belonging to the selected region Ω.

To improve the result of the data grouping, the values a (j, ¡ _{, m)} ^(z) , which correspond to the distance ΔΕ ₀₀ between the jth element between the and _c ¡ ^(z) belonging, are first calculated to the group C¡ ^(z) , and (). The sum of the amount of Data pertaining to each cluster n _c ¡remains constant for each iteration.

Then for each grouping i, the value j * that maximizes b¡ ^(z) (j) is determined, where b¡ ^(z) (j) corresponds to the geometric mean of the values for grouping i divided by the standard deviation of the values for the grouping i.

where,

Once the value of j¡ ^(z) * is found, the value of the solutions that maximize the color distances between the K solutions is obtained.

Finally, a learning rate w is included, ranging from zero to one.

D. Shape and size of the device

The size of the data matrix in information units depends on the volume of static information to be included, which could be compressed, on the dynamic elements and on the number of states in which each information unit can be configured. The choice of the proportion of the smart tag (high width ratio) can be arbitrary, adjusting to the number of units of information required, and the space in the place where the smart tag adheres. Preferred shapes are square and / or rectangular. The size of the data matrix is calculated, both in the square and rectangular format, as the smallest size that is capable of containing both static and dynamic information units. In the case of the square format this corresponds to determining the minimum such that:

Where it corresponds to the information units used by the static and dynamic elements T The information units required to encode the information depend on the coding scheme selected, as explained in section C Configuration of the information units of the device. In the case of using a rectangular format on the label, the proportion between the sides may also be subject to an aspect ratio for practical and / or aesthetic reasons.

E. Optional: Electronic device elements The label can optionally include an electronic system for storing and transmitting information. The information stored in the electronic system can be used as a backup for the dynamic information collected by the analog part of the label. A wireless or contact type data transmission system may also be included. 2. CODING METHOD

The technologies / methods that support the characteristics of the coding method described above are described below.

A. Information settings inside the device

As can be seen in the diagram in Figure 29, the information units of the label may have static, dynamic behavior or be information units that cannot be interpreted by the SR. The static information is encoded on the label by assigning a defined status to each static element (see 1. LABEL DEVICE: C Configuration of the device information units), or using dynamic dual type elements (according to Figure 29) . The dynamic information units used in this way must meet the condition of being recognized as static information units by the SR.

Before assigning the values that correspond to static information to the data matrix, a set of positions in it can be indicated as reserved spaces, which means that during the assignment of values these positions will be omitted. The reserved spaces can be used to include in the label dynamic elements that due to their shape, operation, size, etc. Do not use to encode static information (simple dynamic elements), recognition / calibration patterns or others. Additional elements such as logos, perforations, fasteners, dynamic elements not interpretable by the SR, etc. They can also be located in a reserved space of the label.

The static information is written in the data matrix (in its preferred execution) vertically, from top to bottom, covering the columns from left to right, ignoring, as explained above, the reserved spaces.

A possible situation in which part of the dynamic information units cannot be used to encode static information is when they are required to have a particular organization, for example, in order to generate a pattern that facilitates their visual interpretation. , form a logo, etc. B. Error correction method

An error correction mechanism is included in the information coding, whose objective is to detect and correct the errors that occur in the reading of the information on the smart tag, due to damage to it, or due to the performance of the SR .

An implementation of the error correction method uses the Reed-Solomon algorithm, this algorithm has the peculiarity that it uses elements belonging to a Galois Field of order p (GF (p)) and also an extension of the Galois Field of order p ^ m (GFip ^ m)), sets that are characterized by being closed for addition and multiplication, thus avoiding the problem of overflow (that some operation generates a value that the system is not able to represent) . In GF (p) the operations of addition and multiplication are carried out module p, which means that the result is corrected by replacing it with the module of division by p, in practice only prime numbers are used as a base, because of otherwise, not all elements have a multiplicative inverse, and in some cases the neutral element is not unique limiting the possible sets of values to use.

On the other hand the elements of GF (p ^ m) can be interpreted as polynomials of order (m -1), whose coefficients can be added using addition module p. The construction of the elements requires the definition of a generator polynomial of order m that allows to avoid overflow. As an example, in the following table 1, the construction of the elements of the GF set (2 ^ 2) is shown, using the generating polynomial (a ² + a + 1 = 0), in addition to the comparison with the values greater than (p ^A m) - 1 transforming the equivalent element into both GF (2 ^ 2) and GF (4). It should be noted that there are p ^m different elements in each case.

Table 1. Elements of the GF set (2 ^ 2), using the generator polynomial (a ² + a + 1 = 0)

Within the framework of the Reed Solomon algorithm, the information to be transmitted is interpreted as the coefficients of a polynomial, to which the coefficients of the product of the operation are added with a generating polynomial as follows:

In this construction, n corresponds to the total length of the data chain to be transmitted, while k corresponds to the length of the message itself, the rest (nk) corresponds to the redundant information added as a method of error detection and correction.

The generating polynomial g (x) corresponds to a polynomial of order (nk), whose roots are the elements a, a ² , .... cr ^{(n "k)} in GF (p ^ m). You can check the (nk) roots in the polynomial message received, the evaluation in each of these roots should give a zero result.

The zeros information of the received polynomial allows to detect and correct errors up to a maximum of (nk) / 2, due to the fact that the information of an evaluation is required to detect the error and an additional one to correct it. In the typical implementation, n = (p ^ m) -1 is used, where the generating polynomial corresponds to:

C. Generation of label in printed format

A physical representation is generated, where each position in the data matrix is transformed into a square of defined size. The status of each selected from a set according to the configuration of the information units defined for the device in relation to what is explained in section 1. DEVICE: C Configuration of the information units of the device. In addition, the corresponding states are assigned to the recognition and state calibration patterns. Pattern 201 is rescaled according to the size of the first column of the label image. Surrounding the image of the label includes a security border of white color 206, so that the label is minimally isolated from the environment. The extent of pattern 206 depends on the performance of the SR and the characteristics of the product where the label will be included. For example, in the case of a white product, it might not be necessary to include a border.

Then, the image of the label goes to the printing stage. Depending on the nature of the materials used in the dynamic information units of the label, these can be incorporated in the same printing equipment or at a later stage. D. Optional: Compression of information

Data compression is an optional feature that aims to reduce the size of the data, prior to coding on the smart tag. An important premise is the use of lossless compression algorithms, among which are those based on static models and those based on dictionaries.

Any algorithm that meets the requirement that there is no loss of information can be used, for example the Lempel-Ziv-Welch algorithm.

The data compression process is divided into two parts, coding and decoding. In the case of the Lempel-Ziv-Welch algorithm, the compression algorithm encoder fulfills the function of developing a dynamic index dictionary, which will later be used by the compression algorithm decoder. In the decoding stage the original information is reconstructed, previously coded, through an inverse process with the information present in the dictionary.

E. Optional: Information encryption

Data encryption is an optional feature that aims to implement levels of information security in the code in order to protect information by user types. The applicable encryption methods can be any present in the state of the art provided they do not present data loss. An example of such methods is the AES encryption method, whose implementation is detailed in Example 4. A correct implementation of encryption methods considers the configuration of the label and the requirements defined for the product or element to which it will adhere, as well as also for the different user levels that are considered to be implemented.

It is important to mention that it is possible to implement different methods of encryption of information on the same label in case you want to define, for each user, a different level of security.

3. DECODING METHOD

The following describes the technologies / methods that support the characteristics of the decoding method described above. A. Decoding through the recognition system

The reading of the information from the device by the recognition system involves a series of specific stages.

3.A.1 Acquisition of tag image Information can be obtained from the tag image from a variety of image capture devices (eg scanner, digital camera or other). It is possible to include previous stages of processing, such as the application of a medium filter in cases where the image presents noise, or the reduction of the image size, in order to reduce the processing time, among others, depending on the application of the label and the surface on which it will adhere.

3. A.2 Spatial transformation

In the stage of spatial transformation, the projective distortion of the image is corrected. In a first step, the original image is subdivided into a 4x4 grid in its preferred execution, of which only the 4 sectors containing the corners of the original image are analyzed.

Each sector is transformed to grayscale. Intensity gradients (I) in x and x direction are calculated by applying filters [-1 0 1] and [-1 0 1] '. The objective function for corner detection is calculated, given by:

Where sf is a sensitivity factor in detection. Regions containing local maximums of the objective function, corresponding to the groups of pixels having the same value, while surrounded by pixels of lower intensity, are isolated. The least significant local maximums are eliminated, as a percentage of total detections. Regions containing remaining local maximums contract to a single point. Next, for each of the 4 sectors analyzed, the point recognized as the corner closest to the corner of the original image is selected, according to the position of the sector under analysis. This set of 4 points is considered to delimit the grid that contains the image of the label. The transformation matrix between the coordinates of these detected corners and a rectangle of similar size is calculated, according to:

With X the reference points in coordinates of the original image and Y the reference points in the coordinates of the corrected image. This transformation can be calculated in two stages. Referring to Figure 5, the matrix that maps the unit square to the coordinates of the parallelogram is given by:

If A _x is the matrix that maps the unit square to quadrilateral X and A _and is the matrix that maps the unit square to quadrilateral Y, then you have:

 This transformation is subsequently applied to the entire image.

3.A.3 Pattern recognition

Barcode pattern This is done by transforming the image to a black and white representation, then all areas are isolated with adjacent black pixels. For each zone the rectangle in pixels that contains it (bounding box) is determined, the major edge of all the bounding boxes corresponds to the length of the image of the information matrix in units of pixels. The subsector of the image that corresponds to the barcode is analyzed by extracting a line of information. Points along that line where there is a color change between black and white are detected, thereby calculating a value for the widths of the columns that make up the barcode in pixels. The column widths obtained in this way are normalized by the average of the widths corresponding to the starting pattern, which corresponds to a unit in the UPC coding. Once the normalized width is known, groups of four columns can be translated into whole numbers. Finally, it is verified that the information has been correctly recovered by means of the verification digit.

Pattern of size Pattern 205 is distinguished in similar ways based on the value of the objective function that considers the proximity of the form to the lower side of the data matrix, its eccentricity, and the size ratio between the shape and the total height of the data matrix, according to:

 With the values of the constants adjusted based on a set of label images:

 The above values can be readjusted depending on the specific implementation. The comparison between the size of the pattern 205 in pixels with the size of the pattern 201, allows to determine the size of the data matrix in units of information. The size determined in this way can be verified with the value obtained from reading the barcode. In the event that the barcode information cannot be retrieved reliably, the value of the data matrix size determined from the analysis of pattern 205 is used. 3.A.4 Colorimetric transformation

The colorimetric transformation is part of the recognition of the states and of obtaining information from the information units. Within the possible colorimetric transformations, the digital values of the color to be corrected (previously known design values), and the values are taken as reference representative of the measured colors (mean, median, geometric mean, effective value or RMS, or other of the sample) and then find the polynomial that adjusts the measured colors as close as possible to the theoretical values.

One way to find the adjustment polynomials is through the least squares adjustment approximation, thus minimizing the quadratic error. The polynomial adjustment in matrix notation is presented below:

where y corresponds to colors without distortion, X corresponds to the representative values of the experimental colors, and corresponds to the transformation matrix. The above equation can be solved by premultiplying by the transpose:

The numerical solution of the matrix equation can be found by clearing as follows:

The polynomial order affects the stability of the transformation in the areas related to extrapolation of data. So usually the first and second order settings get more robust results at the time of mapping. However, a higher polynomial order also increases the accuracy of the approach in areas related to interpolation. The colorimetric transformation can be carried out directly with the values of the RGB color model, or in the XYZ, CIELAB, CIELUV or other color space. For practical and processing speed effects, it is that the application of the least squares method can be applied to models and not to color spaces.

Some polynomials used for colorimetric transformation are described below, and polynomials of greater order than those detailed below can be used.

First order transformation polynomial

Let r ₀ , g ₀ and b _{0 be} the values of the reference colors and r ,, and bi the representative values of the measured colors. It has to be that the colorimetric transformation can be expressed in matrix form as follows:

After the colorimetric transformation is obtained, the colors of the entire image are corrected, where r ₂ , g ₂ and b ₂ represent the column vectors of the image. So:

Then we proceed to take the information contained in the arrangement to another dimension (O x P x 3), where (M = 0 x P)

Second and third order transformation polynomial

Like the first order colorimetric transformation, the arrays of the second and third polynomial colorimetric transformations are generated. To simplify the notation, the reference colors are:

second order arrangement:

third order arrangement:

second order colorimetric transformation matrix:

third order colorimetric transformation matrix:

arrangement with the data to correct:

Then you have that the corrected images for the second and third order transformations are:

Then we proceed to take the information contained in the array to another dimension (O x P x 3), where (M = 0 x P).

Finally, the data is standardized as in the first order transformation

3. A.5 Reading information

With the size of the known data matrix, as well as the size of a unit of information in pixels, the data matrix is sequentially traversed, in the same writing order, that is, from top to bottom, traveling the columns from left to right, omitting the zones that correspond to reserved spaces. A subimage is extracted for each unit of information, whose status is compared with the list of states used in the coding. The information of the states used in the coding is uploaded to the SR, for example, by means of a configuration file for a particular execution of the tag. Based on this comparison, an integer is assigned as a read value, the set of these integers corresponds to the decoded message, which may or may not contain errors.

Alternatively, a reading improvement method is included that corresponds to, in addition to analyzing the aforementioned sub-image, analyzing sub-images that correspond to slight displacements of the sample window, considering the one with the lowest variance in the comparison of states to assign a reading value . 3. A.6 Verification and correction of errors

Under the Reed Solomon algorithm, the message received corresponds to the

Message transmitted plus an error component according to:

Assuming the errors are in the positions

and they are of magnitude

, then the polynomial error can be written as:

The syndrome value is defined as the error polynomial evaluated at the roots of the generating polynomial, in the event that no errors occurred during the transmission of the message, all syndrome values should be equal to zero. Because the generator polynomial has (n-k) roots, a system of (n-k) equations and (n-k) unknowns is generated:

The notation can be simplified a little by replacing

The difficulty that the previous system is non-linear is attacked by resolving the error positions separately and then the magnitudes of them. The error locator polynomial is defined as:

The roots of the polynomial locator correspond to the reciprocals of Pi, thus, finding the roots is equivalent to determining the positions of the errors in the received message. It has:

The previous equation is still valid when multiplied by

(In the following equation the notation is used

Then add on p from 1 to (n-k) / 2:

Remembering the definition of syndrome:

Repeating this for different values of m, a system of equations is obtained that allows to obtain the coefficients of the error locator polynomial, and from its roots find the positions with errors in the message:

Once the values have been determined

From the reciprocals of the roots of the error locator polynomial, the magnitudes of the same can be determined, and with this correct the message. In this algorithm the number of errors must be resolved iteratively.

B. Visual interpretation

The visual interpretation of all or part of the dynamic information by the user may require that you have previously received instructions. In general, these consist of the location of the interpretable elements in the label, together with a list of the states in which the element (s) can be found, together with the value / condition of the associated variable of interest. Visual interpretation can also be generated based on intuitive concepts (such as representative figures of concepts known to the collective unconscious) or by logical sequences in certain industries or applications where interpretation does not necessarily require additional instructions. As an example of the previous explanation, the use of the figure X (X) or a ticket (approval or approval symbol) can be considered using colors of the information units to generate the figures.

C. Decompression

If the optional data compression component indicated in the writing method was used, the reading method will use a method according to the compression algorithm used to process the information, before or after decryption depending on the order in which the data were used. optional components.

Once the appropriate decompression method has been used, the corresponding reading sequence is continued. D. Decryption

As an optional module for reading compressed information is used, an optional module of encrypted information is also used that can be used for the implementation of information levels (as indicated in point Έ. Information levels "below) or for the implementation of information security levels.

The decryption module of the SR identifies segments that indicate whether there is encryption or not in a given fragment and proceeds to use the method adequate decryption. Just as it is possible to use various compression methods it is also possible to use various data decryption methods according to the specifications and requirements of the implementation, provided that it is appropriate for the encryption method used. E. Information levels

At the time of decoding, the system is able to recognize the existence of information levels, according to the type of implementation that was used during the coding. This can be done in a variety of ways, for example an unencrypted segment of the information can be used as an indication that the next block of information is encrypted and therefore a password is required to access the information, it is also possible to create different versions of the reading software (producer version, consumer version, etc.).

4. System

Figure 28 shows the optional integration of the smart tag into a traceability system and the integration of the analog, digital, and recognition system information. Such a system involves the interaction of the different means that compose it:

1. Means for configuring the device (A): At this stage the appropriate configuration for the label is chosen based on the variables to be monitored and the user's requirements. Among the aspects to be configured are: the dynamic and static part of the two-dimensional code, encryption algorithm, protocols coding and decoding, device verification number, physical arrangement of spatial and state calibration, shape, size, etc.

2. Means of interaction with the system database (B): The system database consists of several fields for each product. Within the database, two major types of information stand out: the pre-tracking information (C), and the information generated during the product tracking.

3. Encryption and / or compression media (D): At this optional stage, the data is encrypted and / or compressed. Then this, the two-dimensional code (E) adheres to the product, collecting information through the active elements on the label.

4. Means of acquisition and feedback of data. Label information is retrieved from the image obtained by means of a capture device (E). Both spatial and state calibrations are used to retrieve the static and dynamic information contained in the smart tag (F). After reading the two-dimensional code, it is decoded (G), stored information in the database in the fields corresponding to the object tracking (H, I, J), and feedback of the device to through the wireless connection (optional) between the device and the data acquisition portals (K, L, M), where the information fed back to the device is stored inside the internal memory (optional), as a backup of the data processed in Data acquisition portals. At any point in the follow-up, the user can visually access the information contained in the label (Ε ').

Examples

Example 1

An example of coding information regarding a product, corresponding to the character string, is shown in Figure 6 and Figure 7: "MANU Soprale PROD whole milk VOL 11 CONSTEMP 8 C ELAB 12-09-14 VENC 12-11-14 ", identification number" 0019 ", considering 36% of redundancy symbols (the message can be reconstructed even if 18% of the symbols are read incorrectly), in addition in each case a set of information units that have dynamic behavior is included before changes in a variable (eg temperature). In the case shown in Figure 6, the result of the coding is shown considering two possible states of the information units. Representing each character using 7 units of information, a data matrix of 31x31 units of information is required. In the case shown in Figure 7, considering 4 possible states of the information units, each character is represented using 4 information units, requiring a matrix of 24x24 information units, this illustrates the increase in information density when using more than two possible states for each unit of information in the coding.

Example 2

The coding of the information corresponding to the text string "MANU Marfruit PROD salmonahu PESO 500g CONSTEMP 0C ELAB 20-06 VENC 05-07", identification number "0120", using 36% is shown in Figure 8 and Figure 9 of redundancy symbols. Figure 8 shows the coding using four possible states of the information units, considering a set of information units that exhibit dynamic behavior in the face of changes in a variable (eg temperature) requiring a matrix of 24x24 information units. Figure 9 shows how the appearance of the label can vary to rectangular to accommodate a set of information units that exhibit dynamic behavior in the face of changes in a variable (eg pressure), in this case requiring a matrix of 21x30 information units .

Example 3

Figure 10 shows the coding of two types of information; a segment of the genetic code of the species Atlantic salmon (150 characters), followed by fields with product information constituting the message "GEN AGTGCTAGAGGAGTTCTGTATCTATAATTCTTAAAATAATTGTTATGCTTTTTGTGA ACGTTTATCGTGAGTAATTTAGTAAATTCTTAGTGAATTCGGCGGAAGTTTGCGGG GGTATGCTAGTTCTGAACGTCACATGCTAATGTAAAA \ n PRODUCT INFO \ n MANU Marfruit \ n PROD salmonahu \ n WEIGHT 500g \ n CONSTEMP 0C \ n ELAB 20-06 \ n VENC 05-07 \ n PORTION 55g \ n CALORIASTOT 107 \ n GRASASTOT 6g9% \ n CHOLESTEROL 37mg12% \ n SODIUM 32mg1% \ n CARBOHIDTOT 0g0% \ n PROTEIN 13g26% \ n ", identification number" 01 15 ". In this case the information was compressed using Huffman's algorithm. The algorithm is based on encoding the characters that are repeated a greater number of times in the message using the least possible number of bits. Generating an equivalence table as follows:

Table 2. Binary equivalence

The uncoded message consists of 378 characters, equivalent to a requirement of 3024 bits in case each character is encoded using 8 bits. Using the Huffman algorithm, the entire message is reduced to a 1731 bit string, which represents a compression of the information to 57% of the original size. Subsequently, adjacent 8-bit segments are used as integers in the framework of the Reed Solomon algorithm, allowing information to be encoded on a 47x46 information unit label, considering 14% redundancy symbols. Example 4

The result of the coding of the following message is shown in Figure 1 1; "Route 25 efficiency distribution monitoring. Departure from dispatch center X, 1 1: 48 22-07, register arrival at warehouse Y. \ n PRODUCT INFO \ n MANU Marfruit \ n PROD salmonahu \ n WEIGHT 500g \ n CONSTEMP 0C \ n ELAB 20-06 \ n VENC 05-07 \ n ", identification number:" 0318 ". In this case it is presumed that it is not convenient for distributors to know that this particular unit of the product is under special monitoring, in order to avoid preferential treatment, so that prior to coding the information is subjected to an encryption process , in this case with the AES128Bit algorithm, using the "admin_key" key, generating the modified message:

"YTTBDYgpqvQ + XmnCjrqojsKR7tsBkW6 / l7U2xMh / zgqvV + ov0N5ldgAPH2fqrhGtRVmX lagXbhOpr24YmQU6k84tel4j95mqgYm91322pCSanad + / YxTOzN / Ocg8mxhMYYBi / Go esqy9 + PB3Gj4Fm8L7ZIAuXtBDBn6smbbnX4c = \ n PRODUCT INFO \ n MANU Marfruit \ n PROD salmonahu \ n WEIGHT 500g \ n CONSTEMP 0C \ n ELAB 20-06 \ n VENC 05-07 \ n ", where the sensitive information has been replaced, after this the information is compressed using the Huffman algorithm and inserted into the label, in this case using 4 states of the information units and 18% of Redundancy symbols, on a 34x33 information unit label. The original message consists of 219 characters, the encrypted message of 274 characters (2192 bits), and the compressed message of 1642 bits (compression at 75% of the size of the encrypted information).

Example 5 Figure 12 shows the label generated by encoding the following information: "MANU Soprale \ n PROD whole milk \ n VOL 1 l \ n CONSTEMP 8C \ n ELAB 12-09-14 \ n VENC 12-1 1- 14 \ n CALORIES TOT 158 \ n FAT TOT 5g8% \ n CHOLESTEROL 22mg7% \ n SODIUM 150mg6% \ n CARBOHID TOT 18g6% \ n PROTEIN 10g20% \ n ", identification number:" 01 10 ", in a matrix of 47x47 units of information, considering 20% redundancy symbols. The active zone "SENSOR", changes its appearance according to the environmental conditions to be monitored (eg temperature), near the active zone are calibration colors, indicative of specific conditions of the information units used (Figure 13). Figure 14 shows possible states of the active zone, the calibration colors can be used to determine the environmental condition by direct comparison by the user (according to the instruction manual), or by using the SR. Example 6

Figure 15 shows an example of a label that includes dynamic units of information, which can be read both with the naked eye and using the SR. The triangular area corresponds to a protected space and contains sensors that change irreversibly when a certain combination of time and temperature occurs. While the circular-shaped sensor that also corresponds to a reserved space contains sensors that reversibly change to certain temperature values (Figure 16).

It is assumed that the user previously had access to an instruction of the type shown in Figure 17. Visually the user may decide to discard the product, based on the observation of the triangular area. Likewise, observing the circular area gives an idea of the temperature of the product in relation to its recommended consumption temperature (proportion of the circular area). On the other hand, using the SR, you can, based on the specific sensor (s) (if it is information units compatible with the SR) that was activated within the triangular area, determine the cause of the deterioration of the product. Also, based on the configuration of the colored areas in the circular area, you can get a specific recommendation on how to handle the product to reach its optimum temperature. An example of reading the label shown in Figure 15 is:

Example 7

Figure 18 shows the image of the label obtained when encoding the information: "MANU Soprale \ nPROD whole milk \ nVOL 1 l \ nCONSTEMP 8C \ nELAB 12-09-14 \ nVENC 12-11-14 \ n", identification number "0515", including a set of dynamic colorimetric information units that change state with respect to the temperature variable (S), in a matrix of 31x31 information units, considering 30% redundancy symbols. Figure 19 shows a photograph of the same label. The photograph is 1796x1843 pixels in size. Below is the decoding algorithm, which allows you to retrieve the tag information, through the main stages.

Spatial Transformation Stage

The photograph associated with the tag image is broken down into a 4x4 grid. Sub-images containing the corners of the original image are analyzed (Figure 20). Figure 21 shows the detail of the pixels detected as corners in the sub-image corresponding to the upper left corner of the photograph. The corners detected near the edge of the subimage are then discarded (indicated by a square in Figure 22). Of the corners remaining, in this case, the one closest to the upper left corner of the subimage is selected (indicated by a triangle in Figure 22). An analogous procedure is performed with the remaining three sub-images. Once the coordinates of the corners of the data matrix are obtained, the transformation matrix is obtained according to what is described in section 3. DECODING METHOD: A Decoding by means of the recognition system.

Pattern Recognition Stage

After the application of the spatial transformation, the image is transformed to black and white (Figure 23), and the properties that characterize each region of contiguous pixels (Bounding box, eccentricity and centroid) are obtained.

Bounding box analysis

For each of the regions, the rectangle that contains it is determined, characterized by the coordinates of the upper left vertex, and the width and height values. An example of the values obtained, with respect to the regions indicated in Figure 24, is shown in the following table 3:

Table 3.

In this particular case, 32 different regions are detected, the height of the data matrix corresponds to a greater height between the detected regions (Figure 25), corresponding to 1660 pixels.

Next, the size pattern 205 is identified by evaluation of the detection function (see 3. DECODING METHOD: A Decoding by means of the recognition system), the values corresponding to the regions with lower scores, and in reference to the Figure 27, are shown in the following table 4:

Table 4 It is observed that the pattern is detected correctly. This determines a provisional value for the size of the data matrix in information units (31), and the size of each information unit in pixels (54).

Bar code analysis The part of the image that corresponds to the bar code is analyzed below (Figure 26). From this image, the value of the length of each of the bars is determined, values that are subsequently normalized by the average of the detected length for the 3 bars corresponding to the coding start pattern. Both sets of values are shown in table 5, where the 31 values shown correspond to the three bars of the start pattern, along with the 28 bars corresponding to the 7 coded digits:

Table 5.

Normalized lengths allow each digit to be identified according to table 6:

Table 6.

This verifies the check digit:

Since the check digit matches, it is assumed that the information retrieved from the barcode is correct, in this case the size of the matrix determined by both methods (31) matches.

Reading the information

The matrix is then sequentially traversed, assigning each unit of information a value, depending on its status (in this example: color). The median of the color values is used in the case of coding using two states. Table 7 shows the value assignment to some units of information in the data matrix as an example:

Table 7.

It should be noted that in some cases in Table 7 this allocation is more reliable than in others. After this, the read values are concatenated and we proceed to verify that no errors occurred during the reading. In this particular case the message is recovered without errors.

Example 8

In Figure 33, the selected region Ω is shown in the xyY color space. This region was obtained by determining the intersection between the color ranges corresponding to ProPhoto RGB, Adobe RGB 1998, Colormatch RGB, sRGB and SWOP CMYK.

16 clusters were determined (K = 16), using 200 iterations (iter = 200) and a learning rate of 0.8 (w = 0.8). The initial colors μ ° were chosen randomly respecting the boundary conditions of the Ω region, the transformation to the CIELAB color space of the initial colors is shown in the Table 8

Table 8

The functions of empirical distribution of the indicators a ^(z) and b ^(z) (for z = 200), respectively, are shown in Figure 34 and Figure 35. Table 9 shows the colors obtained at the end of the calculation.

 Table 9

 Figure 36 shows the result obtained from the groupings in the selected region Ω belonging to the CIELAB color space for iteration number 200.

Figure 37 shows the result obtained from the groupings in the selected region Ω belonging to the color space xyY for iteration number 200. Figure 38 and Figure 39 show the results obtained from the groupings in the selected region Ω belonging to the xyY color space, for the 200th iteration, individually. Figure 40 shows the variation of the color with respect to the number of iterations, in the color space xyY. The initial and final solution are indicated as μ _ι ⁽⁰⁾ and μι ⁽²⁰⁰⁾ , respectively.

Example 9 Characterization of the colorimetric behavior of a polychromatic material (TLC), with respect to temperature:

 The range of use of the material is between 15 and 29 [° C].

 A second order colorimetric transformation was used for color correction, the comparison of the values for each color channel in RGB, corrected and uncorrected, is shown in Figure 41.

 As a regression method, a Feed Forward Neural Network was used, with 20 neurons in the first layer, 20 neurons in the second layer, and 1 neuron in the output layer.

45 characteristics corresponding to each of the channels of the following color transformations were used: YCbCr (3 channels), LCH (3 channels), RGB (3 channels), YPbPr (3 channels), YUV (3 channels), YIQ ( 3 channels), HSV (3 channels), HSL (3 channels), HSI (3 channels), XYZ (3 channels), Lab (3 channels), Luv (3 channels), 'JPEG- YCbCr (3 channels), CAT02 LMS (3 channels), YDbDr (3 channels). The method used for the selection of characteristics was one of the wrapping type, in each iteration the characteristic with a better performance is added to the regression model, while the performance increases. As an indicator of performance, the mean square error (MSE) was used, the results are shown in Table 10. From the results, 4 characteristics were selected since the fifth iteration worsens the performance.

 Table 10

Figure 42 shows the relationship between the adjusted and experimental temperature for the data set used in the calculation, quadratic correlation coefficient R ² = 0.9969.

 The cumulative distribution of the adjustment error is shown in Figure 43.

Claims

1. Tag-like device to be incorporated into an item or product or used independently, comprising: information units with one or multiple states, which store coded information; a basis for arranging said information units; wireless means for data transmission and reception; memory unit for data storage; and data processing means, CHARACTERIZED because the information units contain the information encoded in a static and dynamic way under a single code, where said information is read visually, directly by the user, by means of a recognition system and / or through a wireless access system.

2. The tag type device according to claim 1, CHARACTERIZED in that the information units have one or more states.

3. The tag type device according to claim 1, CHARACTERIZED because the information units represent static information that does not vary over time.

4. The tag type device according to claim 1, CHARACTERIZED because the information units represent dynamic information varies over time.

5. The tag type device according to claim 1, CHARACTERIZED in that the information units are interpreted as sensors.

6. The tag type device according to claim 5, CHARACTERIZED in that the information units that are interpreted as sensors are used to encode dynamic and static information.

7. The label type device according to claim 5 and 6, CHARACTERIZED because the information units that are interpreted as sensors that measure multiple variables such as temperature, ultraviolet rays, pressure, pH, humidity and time.

8. The tag type device according to claim 1, CHARACTERIZED because it further comprises calibration fields.

9. The label type device according to claim 8, CHARACTERIZED in that the spatial calibration fields consist of squares of known size in information units (205), and recognition patterns (201 and 204).

10. The tag type device according to claim 1, CHARACTERIZED in that it comprises the shape of a square or rectangle.

1 1. Information coding method of the tag type device of claim 1, CHARACTERIZED in that it comprises: defining application requirements of the device according to: amount of information to be encoded; quantity and nature of the variables to be monitored; and redundancy of information; establish type and quantity of information units, where the quantity varies according to the number of states that each unit of information takes and the amount of information to be stored; establish the size and shape of the label, based on the amount of information units to be stored; reserve spaces for recognition, calibration and sensor patterns; apply an error correction method to generate redundancy data for error detection and correction; write, in the label, the information and redundancy data as values of a data matrix, without using the reserved spaces, where each unit of information takes a state according to the type of coding selected, where the values of the matrix correspond to the states that take the static and dynamic elements of dual type; incorporate, in the reserved spaces, patterns of recognition and calibration of states; incorporate additional information associated with the tag size; incorporate a security border, without information, around the image of the label; and incorporate the dynamic information units of the label.

12. The coding method according to claim 11, CHARACTERIZED because the tag information is encrypted.

13. The method of coding the label according to claim 1, CHARACTERIZED because the information on the label is compressed.

14. The coding method according to claim 13, CHARACTERIZED because compressing the information comprises decreasing the size of the stored data and the label.

15. The coding method according to claim 12, CHARACTERIZED in that the encryption of the information comprises defining levels of access to the tag information for the user.

16. The coding method according to claim 11, CHARACTERIZED because the state calibration is colorimetric.

17. The coding method according to claim 11, CHARACTERIZED in that the amount of information to be stored in each unit of information is given by the type of coding selected.

18. The coding method according to claim 11, CHARACTERIZED in that the error correction method uses the Reed Solomon algorithm.

19. The coding method according to claim 11, CHARACTERIZED in that the process of writing information as values of a data matrix comprises writing the information from top to bottom, covering columns from left to right.

20. The coding method according to claim 11, CHARACTERIZED in that the step of incorporating additional information associated with the size of the label is carried out in the form of a one-dimensional barcode.

21. The coding method according to claim 11, CHARACTERIZED in that the safety edge is a white border.

22. The coding method according to claim 1, CHARACTERIZED because the safety edge is a line enclosing the label.

23. Method of decoding the information contained in the tag type device of claim 1, CHARACTERIZED in that it comprises: capturing a digital image of the tag; rescale the tag image; eliminate noise in the tag image; perform a projective transformation of the tag image; identify and read the information contained in the bar code; identify size patterns (201, 205); determine and validate the size of the data matrix; perform state correction; read dynamic information associated with the information units that are interpreted as sensors; read the encoded information, associated with the static information; detect and correct errors in the encoded information; and report the information stored on the label to the user.

24. Decoding method, according to claim 23, CHARACTERIZED because it further comprises the visual reading of all or part of the dynamic information contained in the label, where said reading is carried out without the need for a recognition system, according to instructions included in the label, provided separately, by intuitive interpretation according to the provision of dynamic elements (simple or dual ).

25. Decoding method according to claim 23 and 24, CHARACTERIZED because the step of reading the dynamic information associated with the information units that are interpreted as sensors comprises visually interpreting the label by the user, according to the states that take the information units.

26. The decoding method according to claim 23, CHARACTERIZED in that step of eliminating the noise in the captured image comprises the application of filters.

27. The decoding method according to claim 23, CHARACTERIZED in that the projective transformation comprises the steps of: subdividing the image into a 4x4 grid; determine the sectors that contain the corners of each of the squares generated by dividing the original image into a 4x4 grid, which includes: transforming each sector to grayscale; calculate color intensity gradients in the x and y direction; apply an objective function on intensity gradient values for equine detection; Y in each sector, select the point recognized as the corner of the original image based on the value of the objective function and criteria of proximity to the edge of the image.

28 The method according to claim 23, CHARACTERIZED in that the stage of state correction is colorimetric correction.

29. The decoding method according to claim 23, CHARACTERIZED in that the reading of static information comprises: traversing the data matrix in the same order in which the information was written during the encoding of the information, where the reserved spaces they are omitted; extract an image from each unit of information; and compare the status of each unit of information with the corresponding state of the coding process, to assign it an integer value.

30. The decoding method according to claim 23, CHARACTERIZED in that the determination of the size of the data matrix comprises comparing the size patterns (201 and 205) with the value obtained from the bar code reading.

31. The decoding method according to claim 23, CHARACTERIZED in that the report of the information to the user is carried out by presenting the information in a user interface according to the recognition system used.

32. The decoding method according to claim 23, CHARACTERIZED because if the data is compressed, decompression is performed after the step of detecting and correcting errors in the encoded information.

33. The decoding method according to claim 23, CHARACTERIZED because if the data is encrypted, decryption is performed after the step of detecting and correcting errors in the encoded information.

34. The decoding method according to claim 33, CHARACTERIZED because decrypting data comprises establishing levels of access to information for the user.

35. The decoding method according to claim 23,

CHARACTERIZED because it also includes storing the information in a data storage medium on the label.

36. Logistics system for the label of claim 1, because it comprises: a database; configuration and coding means, to configure and encode the information on the label according to static and dynamic information units; means of label recognition; means to interact with the database; encryption and compression means; means of data acquisition and storage; data feedback means; wireless media; data storage media; CHARACTERIZED because static and dynamic information encoded under a single code is read directly by the user, by an optical system, an image capture and processing system or through a wireless access system.

37. The system according to claim 36, CHARACTERIZED because the data storage medium is an internal memory of the tag.

38. The system according to claim 36, CHARACTERIZED in that the optical system is a digital camera, a scanner or other device that allows the capture of an image.

39. The system according to claim 36, CHARACTERIZED in that the wireless communication is carried out through RFID, WIFI, NFC, BLUETOOTH or other wireless communication means.

40. The system according to claim 36, CHARACTERIZED in that the encryption means allow generating levels of access to the information encoded for the user.

41. The system according to claim 36, CHARACTERIZED in that it comprises the implementation of information levels due to differences in the methodology of disposition of the information units in the data matrix and according to the type of reward medium used.