Method and system for reading an identification code in a scene

Abstract

A computer-implemented method for reading an identification code (ID) in a scene, the method comprising the steps of acquiring (E1) a first digital image (IMG1), determining (E2) a motion image (IMGΔ), determining (E4) a plurality of primary cells in the motion image (IMGΔ), determining a probability of the presence of an identification code (ID) in each primary cell of the first digital image (IMG1) in order to deduce secondary cells and a relevant focus area, acquiring a second image with the relevant focus area, and performing an identification code (ID) reading in the second image. Abstract figure: Figure 3

Description

Title of the invention: Method and system for reading an identification code in a scene technical field

[0001] The present invention relates to the field of reading an identification code in a scene, for example, to identify products in a warehouse.

[0002] In the field of logistics and commerce, it is known to place on a product a label containing an identification code in order to allow the product to be identified by computer in order in particular to know its nature, its price and other.

[0003] The identification code is generally in the form of a barcode and may include parallel black and white bars (one-dimensional or 1D category barcode) or a matrix of black and white squares (two-dimensional or 2D category barcode).

[0004] In a warehouse containing a large number of products, it is known to carry out an inventory of the products using a mobile robot equipped with an acquisition device, for example a camera, to acquire an image of a scene likely to include one or more barcodes. The image undergoes a processing procedure comprising steps consisting, on the one hand, of detecting one or more barcodes and, on the other hand, of reading the detected barcode(s).

[0005] In practice, barcode reading is optimal when the barcode is sufficiently sharp in the image. Thus, if the focus area of ​​the acquisition device is not correctly set, the image is blurry and reading is not possible. Such a problem is common when the acquisition device is moved rapidly within the scene, the relative movement resulting in an unsuitable focus area.

[0006] To eliminate this drawback, it has been proposed to acquire a large number of images with different focus areas and then evaluate the sharpness of each image. The sharpest image is selected and processed to detect and read a barcode. Such a method has the disadvantage of requiring significant computing resources and increasing the barcode reading time. Furthermore, an image may be considered generally sharp even though the barcode is not readable.

[0007] To better define the focus area, it has been proposed to evaluate the distance between the acquisition device and a product using equipment such as a laser sensor (LIDAR) or an ultrasonic sensor. Such equipment is expensive and increases the size, which reduces its potential for integration on a mobile robot or other device.

[0008] To overcome this drawback, one solution would be to reduce the speed of the mobile robot in order to limit focus errors that prevent optimal reading. However, such a solution is not suitable since it would significantly increase the time required to perform a product inventory in a warehouse.

[0009] The invention thus aims to eliminate at least some of these drawbacks. PRESENTATION OF THE INVENTION

[0010] The invention also relates to a computer-implemented method for reading an identification code in a scene, the method comprising steps consisting of: • Acquire an initial digital image of the scene with a default focus area, • Determine a motion image by comparing the first digital image to a previous digital image, each image having a plurality of pixels with determined positions and a value, the value of each pixel of the motion image being determined by subtracting the value of a pixel of the previous digital image from the value of a pixel of the first digital image at corresponding positions, • Determine, within the motion image, the pixels with a value greater than a predetermined primary threshold, hereinafter referred to as primary pixels, • Determine a plurality of primary cells from the primary pixels in the motion image, • Determine the probability of finding an identification code in each primary cell of the first digital image, • Determine the primary cells having a probability of presence greater than a predetermined secondary threshold, hereinafter referred to as secondary cells, • Determine a relevant focus area from these secondary cells, • Acquire a second image with the relevant focus area and • Read an identification code from the second image.

[0011] Thanks to the invention, a first image is advantageously used sacrificially to determine a relevant focus area located at the probable position of an identification code. This makes it possible to acquire a second image with a very sharp identification code that can be read conveniently and quickly with limited computing resources. Given the high sharpness, the computing resources required for reading are advantageously reduced. The use of primary cells allows for the convenient and rapid processing of pixel groups. Only a small amount of information is processed. The determination of the relevant focus area is carried out quickly so that the second image It has a small amount of motion compared to the first image. Thanks to the invention, high-speed identification code reading is possible, enabling implementation in mobile robots or on transport carts. Product inventory can then be carried out quickly. This avoids the need for numerous focus area adjustment operations.

[0012] According to one aspect, the step of determining a plurality of primary cells comprises substeps consisting of: • Identify a plurality of cells within the motion image, each cell encompassing the same number of pixels. • Determine, for each cell, the number of primary pixels contained within that cell, • Consider as a primary cell only a cell having a number of primary pixels greater than a predetermined movement threshold.

[0013] Advantageously, a primary cell allows a reading window to be formed on the most relevant area of ​​the image. This reduces the number of operations and speeds up processing.

[0014] According to one aspect, the process includes a step consisting of: • Determine the number of primary cells relative to the total number of cells and • Acquire a new first image if the number of primary cells relative to the total number of cells is greater than a predetermined filtering threshold.

[0015] Advantageously, if the amount of motion is too great, the image is rejected. Only images with a low amount of motion are used to determine a relevant focus area. This avoids processing images that cannot be read.

[0016] According to one aspect, the relevant focus area is determined by giving priority to secondary cells with the highest probability of presence. This ensures that the relevant focus area is centered on the identification code.

[0017] According to one aspect, the probability of the presence of an identification code in a primary cell is determined by means of at least one mathematical correlation function applied to the pixels of said primary cell. Such a mathematical correlation function makes it possible to detect the presence, nature, and orientation of the identification code. This advantageously allows an identification code to be distinguished from text and for several different identification codes to be distinguished in an image.

[0018] According to one aspect, the mathematical correlation function includes a measure The intensity of a gradient is measured along different angular orientations, and the intensity measurement is compared to at least one characteristic intensity profile of a category within an identification code. This allows for the measurement of a similarity criterion, i.e., a probability of presence.

[0019] According to one aspect, an identification code may belong to at least two categories, the method comprising a step consisting of: • Determine a probability of presence for each category of identification code in each primary cell of the first digital image, • For each identification code category, determine the primary cells with a probability of presence greater than a predetermined secondary threshold, hereinafter referred to as secondary cells, • For each identification code category, determine a relevant focus area from the aforementioned secondary cells, • For each identification code category, acquire a second image with the relevant focus area and • For each category of identification code, perform a reading of an identification code in the second image.

[0020] The invention also relates to a computer program type product, comprising at least one sequence of instructions stored and readable by a processor and which, once read by this processor, causes the execution of the steps of the process as previously presented.

[0021] The invention also relates to a computer-readable medium comprising the computer program-type product as previously described.

[0022] The invention also relates to a system for reading an identification code in a scene, the reading system comprising an acquisition device configured to acquire several digital images of the scene and a computer configured to • Command the acquisition device to acquire an initial digital image of the scene with a default focus area, • Determine a motion image by comparing the first digital image to a previous digital image, each image having a plurality of pixels with determined positions and a value, the value of each pixel of the motion image being determined by subtracting the value of a pixel of the previous digital image from the value of a pixel of the first digital image at corresponding positions, • Determine, within the motion image, the pixels with a value greater than a predetermined primary threshold, hereinafter referred to as primary pixels, • Determine a plurality of primary cells from the primary pixels in the moving image, • Determine the probability of finding an identification code in each primary cell of the first digital image, • Determine the primary cells having a probability of presence greater than a predetermined secondary threshold, hereinafter referred to as secondary cells, • Determine a relevant focus area from said secondary cells, • Command the acquisition device to acquire a second image with the relevant focus area and • Read an identification code from the second image. PRESENTATION OF THE FIGURES

[0023] The invention will be better understood upon reading the following description, given by way of example, and referring to the following figures, given by way of non-limiting examples, in which identical references are given to similar objects.

[0024] Fig. 1 is a schematic representation of a scene comprising a plurality of products, arranged in different positions, each of which is associated with an identification code.

[0025] Fig. 2 is a schematic representation of a system for reading an identification code according to one embodiment of the invention.

[0026] Fig. 3 is a schematic representation of the steps in implementing a reading process.

[0027] Fig. 4 is a schematic representation of primary cell determination in the motion image.

[0028] Fig. 5 is a schematic representation of the probability of presence in the primary cells of the first image.

[0029] Fig. 6 is a schematic representation of histograms produced for a "one-dimensional" category barcode, for a "two-dimensional" category barcode and for text.

[0030] Fig. 7 is a schematic representation of the determination of primary cells having the highest probability of presence (secondary cells).

[0031] Fig. 8 is a schematic representation of the determination of a relevant focus area determined from secondary cells.

[0032] Fig. 9 is a schematic representation of the acquisition of a second image with the relevant focus area.

[0033] Fig. 10 is a schematic representation of a step in reading the second image.

[0034] It should be noted that the figures explain the invention in detail to put it into works the invention, the said figures being of course able to better define the invention where appropriate. DETAILED DESCRIPTION OF THE INVENTION

[0035] The invention relates to a computer-implemented method for reading an identification code in a scene. For example, the scene could represent shelves in a warehouse with numerous products in different positions. Each product has an identification code.

[0036] The identification code is generally in the form of a barcode and may consist of parallel black and white bars (one-dimensional or 1D barcode) or a matrix of black and white squares (two-dimensional or 2D barcode), also called a matrix code. Examples of matrix codes include the QR Code, the Da-tamatrix, the Aztec code, and the Maxicode. Such an identification code is easy to read in an image if the ID code is clear in the image.

[0037] By way of example, with reference to [Fig. 1], a scene of a warehouse SC is shown containing five products P1-P5, each of which is associated with an identification code ID1-ID5. The products P1-P5 are in different positions in the warehouse.

[0038] In a known manner, in order to carry out an inventory, to check prices or to ensure logistical management, it is necessary to control the products P1-P5 present in the warehouse by reading their identification codes ID1-ID5 using a reading system 10 which will now be presented.

[0039] With reference to [Fig. 2], a reading system 10 according to one embodiment of the invention is schematically represented. The reading system 10 comprises an acquisition device 11 configured to acquire at least one digital image of the scene SC, each digital image having a plurality of pixels with definite positions and values. In a known manner, the pixels are arranged in rows and columns. Each pixel has a position defined by its row and column. Preferably, the value of a pixel is defined in grayscale. Preferably, each image has the same number of pixels.

[0040] The reading system 10 further includes a processing unit, also called a computer, for performing, in particular, image processing steps. Some of the processing steps can be performed locally by a processor 12 or remotely, for example, by a processing server 2 accessible via a communication network 3, for example, via the internet. Thus, the processing unit can be local, remote, or local / remote.

[0041] In this example, with reference to [Fig. 2], the reading system 10 is in the form of a mobile phone 1 comprising an acquisition device 11 and a electronic board equipped with a processor 12. The electronic board also includes a network module to allow connection to the processing server 2 via the communication network 3. Thus, part of the computer's processing can be carried out by the processor 12 while another part of the processing can be carried out by the processing server 2.

[0042] The reading system 10 can be in the form of a mobile robot, preferably autonomous, to acquire images during the robot's movements. To increase integration, it has also been proposed to integrate the reading system 10 with a forklift so as to read ID1-ID5 identification codes when products are moved by the forklift or when the forklift is circulating within the warehouse.

[0043] With reference to [Fig. 3], the acquisition device 11 is configured to acquire a plurality of successive IMG0, IMG1 images with a time interval of, for example, between 20 ms and 100 ms. As previously stated, the acquisition device 11 can be moved, either manually or robotically, which induces movement.

[0044] The acquisition device 11 is configured to acquire an image based on an adjustable focus area. In this example, with no adjustment, the acquisition device 11 is configured to acquire an image based on a default focus area Fd. In practice, the default focus area Fd is determined at the center of the image ([Fig. 3]).

[0045] An example of the implementation of a reading process will be presented with reference to [Fig.3]

[0046] The process includes a step of acquiring a first image El The first image IMG1 is a digital image of the scene with a default focus area Fd. The first image IMG1 is acquired by the acquisition device 11 with a default focus area Fd located at the center of the first image IMG1. In this example, the first image IMG1 has a unique ID code located away from the default focus area Fd. In this example, the ID code is a one-dimensional barcode.

[0047] With reference to [Fig. 3], the method includes a step of determining E2 a motion image IMGA by comparing the first digital image IMG1 to a previous digital image IMG0. The previous digital image IMG0 was preferably acquired shortly before the first digital image IMG1 so as to determine the momentum at the time of the acquisition of the first digital image IMG1. Preferably, the previous digital image IMG0 is the one acquired less than 100 ms before the first image IMG1, preferably just before the first image IMG1, so as to have a momentum as low as possible.

[0048] In practice, each image IMG1, IMGO comprises a plurality of pixels, preferably with a gray level value. The value of each pixel in the motion image IMGA is determined by subtracting the value of a pixel in the previous digital image IMGO from the value of a pixel in the first digital image IMG1 at corresponding positions. Preferably, each image IMG1, IMGO has the same number of pixels.

[0049] With further reference to [Fig. 3], the method includes a step of determining E3 the pixels of the motion image IMGA having a value greater than a predetermined primary threshold SI, hereinafter referred to as primary pixels PIX1. In this example, the primary pixels PIX1 are shown on the motion image IMGA. The predetermined primary threshold SI is determined based on the noise level in the first digital image IMG1. In one aspect, the noise level is evaluated in a uniform area of ​​the first digital image IMG1. Indeed, in a uniform area, the variation in the gray levels of the pixels is considered to be solely due to noise. In practice, the mean and standard deviation of the gray levels in this area (Gaussian noise) are calculated, and the noise level is set at a multiple of the standard deviation (for example, 2 to 3 times).

[0050] Preferably, the method includes a step of determining E4 a plurality of cells from the pixels of the motion image IMGA. In practice, each motion image IMGA is divided into a plurality of cells. Preferably, the cells are rectangles or squares of predetermined size. In the figures, each cell is represented by a square. Each cell encompasses the same number of pixels.

[0051] Each cell is associated with a number NI of primary pixels PIX1 contained in said cell.

[0052] With reference to [Fig.3], the method includes a step of comparing the number of primary pixels PIX1 of a cell to a predetermined motion threshold SM, in particular, determined as a function of the number of pixels encompassed in the cell.

[0053] If the number of primary pixels PIX1 of a cell is greater than the predetermined motion threshold SM, the cell has motion and is hereafter designated "primary cell Cl" as illustrated in [Fig.4].

[0054] The primary cells Cl are determined from the motion image IMGA and are applied to the first image IMG1 in the manner of a reading window. This allows only a limited, but relevant, amount of information from the first image IMG1 to be processed. In other words, the primary cells Cl form elementary windows that allow only a portion of an image to be processed in order to accelerate processing.

[0055] According to an optional aspect, the method includes a step of evaluating the number of primary cells Cl relative to the total number of CTot cells and comparing it to a predetermined filtering threshold SF. If the number of primary cells Cl relative to the total number of CTot cells exceeds the predetermined filtering threshold SF, the first image IMG1 is rejected because it contains too much motion. The method then includes a step of acquiring a new first image IMG1' whose motion will also be tested.

[0056] Conversely, if the number of primary cells Cl relative to the total number of CTot cells is less than the predetermined filtering threshold SF, the first image IMG1 is retained for processing as illustrated in [Fig. 3]. This filtering step E41 allows for the early elimination of a first image IMG1 with an excessive amount of blur, which would interfere with the subsequent reading of the identification code ID.

[0057] The quantity of motion is localized on the ID identification codes since the nature of an ID identification code, in particular a barcode or a matrix code, generates primary pixels PIX1 due to the presence of black / white bars or a matrix of black / white squares, which can be conveniently highlighted in the motion image IMGA. In practice, the presence of text in the first image IMG1 also generates primary pixels PIX1.

[0058] With reference to [Fig. 5], the method includes a step of determining E5 a probability Y0-Y2 of the presence of an identification code ID in each primary cell Cl of the first image IMG1. Advantageously, this step is only carried out when the amount of blur is low, which allows the probability to be determined reliably (see preliminary step E31).

[0059] In practice, one or more mathematical correlation functions are applied to the primary cells Cl of the first image IMG1 so as to obtain a Y0-Y2 probability of presence for each primary cell Cl as illustrated in [Fig. 5]. Applying the mathematical correlation functions only to the primary cells Cl of the first image IMG1 reduces processing time and limits the computing resources required. Preferably, with reference to [Fig. 5], the mathematical correlation functions are applied to each primary cell CL

[0060] By way of example, with reference to [Fig. 6], the mathematical correlation function includes a measure of the intensity of a gradient along different angular orientations. In particular, a Ha-Hc histogram is determined, defining the amplitude of the gradient over an angular range from 0 to 360°. Depending on the category of the identification code or the absence of an identification code, the histogram is characterized- ristique.

[0061] In this example, the mathematical correlation function includes a comparison of the intensity measurement to at least one characteristic intensity profile of a category of an identification code so as to determine the probability of presence Y0-Y2 for each primary cell Cl for each category.

[0062] In the case of a one-dimensional barcode with black / white bars, the histogram Ha ([Fig. 6]) has two peaks 180° apart, corresponding to the orientation of the bars. For a two-dimensional barcode with a matrix of black / white squares, the histogram Hb ([Fig. 6]) has four peaks 90° apart, corresponding to the orientation of the matrix. Advantageously, when an identification code is present, the histogram has visible peaks that can be interpreted. Conversely, when text is present, the peaks are not precisely distributed, as illustrated in the histogram Hc.Advantageously, by comparing the intensity measurement to at least one characteristic intensity profile of a category of an identification code (e.g., histograms with two peaks at 180° or four peaks at 90°), a probability of presence Y0-Y2 can be determined for each category of ID identification code for each primary cell Cl.

[0063] With reference to [Fig. 5], three probability levels Y0, Y1, Y2 are shown for an identification code belonging to the "one-dimensional" category, namely: zero probability Y0, medium probability Y1, and high probability Y2. It is understood that the number of probability levels could be different. Similarly, probability levels could be presented for several categories of ID identification codes.

[0064] Preferably, the probability of presence Y0-Y2 is determined with respect to the height of the peaks of the histogram by comparison to a characteristic intensity profile for an identification code belonging to the "one dimension" category.

[0065] With reference to [Fig. 7], the process includes a step of determining E6 the primary cells Cl having a probability of presence Y0-Y2 greater than a predetermined secondary threshold S6, hereinafter referred to as secondary cells C2. In this example, all primary cells Cl having a medium probability of presence Y1 or a high probability of presence Y2 are considered secondary cells C2. The secondary cells C2 are associated with the category "one dimension". It is understood that in the presence of several categories, there could be secondary cells C2 for each category.

[0066] With reference to [Fig. 8], the method includes a step of determining E7 a relevant focus area Fp from said secondary cells C2 relating to each category. As illustrated in [Fig. 7], the relevant focus area Fp is a The shape encompasses the secondary C2 cells of the same category, preferably a rectangular shape. In this example, the secondary C2 cells already have a rectangular shape, and the relevant focus area Fp corresponds exactly to the secondary C2 cells. However, it goes without saying that this could be different. Preferably, the relevant focus area Fp is centered on the secondary C2 cells with the highest probability (probability Y2). The relevant focus area Fp is thus positioned at the most probable location of the ID code.

[0067] With reference to [Fig. 9], the method includes a step of acquiring E8 a second image IMG2 with the relevant focus area Fp. Unlike the acquisition of the first image IMG1 with a default focus area Fd located in the center of the first image IMG1, the second image IMG2 has a focus area positioned in the upper right corner and centered on the ID code. In other words, the second image IMG2 is acquired under optimal sharpness conditions for the ID code. It is important that the relevant focus area Fp be determined quickly since the second image IMG2 must be acquired very quickly compared to the first image IMG1 in order to understand a reduced amount of motion. Advantageously, the second image IMG2 is acquired less than 100 ms after the first image IMG1.

[0068] With reference to [Fig. 10], the method includes a step E9 of reading an identification code ID from the second image IMG2. Advantageously, since the identification code ID is clear, the reading can be performed optimally with a high processing time. Preferably, the second image IMG2 is processed remotely by the processing server 2, which allows the mobile phone 1 to receive the identification code ID2 and any other information related to the identification code ID, for example, a price, an identification label, a manufacturer, etc.

[0069] If several different identification codes are present in the same first image IMG1, a relevant focus area is preferably determined for each identification code. This allows the acquisition of several second images IMG2 with different relevant focus areas. Preferably, to enable a more robust determination of the relevant focus area Fp, the primary cells Cl are also labeled according to the orientation of the identification code ID (angular position of the peaks in the histograms) so as to facilitate the detection of different identification codes.

[0070] If several identification codes belonging to different categories are present in the same first image IMG1, a relevant focus area specific to each category is preferably determined for the acquisition of several second images IMG2 with relevant focus areas that may be different different depending on the categories. In other words, a relevant focus area can be determined individually for each identification code. When two relevant focus areas are close to each other, only one focus area is retained or determined in order to reduce the number of second image acquisitions. Processing server 2 is advantageously configured to determine several identification codes simultaneously in the same second image, IMG2.

Claims

Claims

1. A method for reading an identification code (ID) in a scene (SC), the method comprising steps of: • Acquire (El) a first digital image (IMG1) of the scene (SC) with a default focus area (Fd), • Determining (E2) a motion image (IMGA) by comparing the first digital image (IMG1) to a previous digital image (IMGO), each image (IMG1, IMGO) comprising a plurality of pixels having determined positions and each pixel having a value, the value of each pixel of the motion image (IMGA) being determined by subtracting the value of a pixel of the previous digital image (IMGO) from the value of a pixel of the first digital image (IMG1) at corresponding positions, • Determine (E3) in the motion image (IMGA) the pixels having a value greater than a predetermined primary threshold (SI), hereinafter referred to as primary pixels (PIX1), • Determine (E4) a plurality of primary cells (Cl) from the primary pixels (PIX1) in the motion image (IMGA), • Determine (E5) a probability of presence (Y0-Y2) of an identification code (ID) in each primary cell (Cl) of the first digital image (IMG1), • Determine (E6) the primary cells (Cl) having a probability of presence (Y0-Y2) greater than a predetermined secondary threshold (S6), hereinafter referred to as secondary cells (C2), • Determine (E7) a relevant focus area (Fp) from said secondary cells (C2), • Acquire (E8) a second image (IMG2) with the relevant focus area (Fp) and • Carry out (E9) a reading of an identification code (ID) in the second image (IMG2).

2. The method of claim 1, wherein the step of de- terminating (E4) a plurality of primary cells (Cl) comprises sub-steps consisting of: • Determining in the motion image (IMGA) a plurality of cells, each cell encompassing the same number of pixels, • Determining, for each cell, the number (NI) of primary pixels (PIX1) included in said cell, • Considering as a primary cell (Cl) only a cell having a number (NI) of primary pixels (PIX1) greater than a predetermined motion threshold (SM).

3. Method according to claim 2, comprising a step (E41) consisting of: • Determining the number of primary cells (Cl) relative to the total number of cells (CTot) and • Acquiring a new first image (IMG1') if the number of primary cells (Cl) relative to the total number of cells (CTot) is greater than a predetermined filtering threshold (SF).

4. Method according to one of claims 1 to 3, in which the relevant focus area (Fp) is determined by taking into account in a priority manner the secondary cells (C2) having the highest probability of presence (Y0-Y2).

5. Method according to one of claims 1 to 4, in which the probability of presence (Y0-Y2) of an identification code (ID) in a primary cell (Cl) is determined by means of at least one mathematical correlation function applied to the pixels of said primary cell (Cl).

6. A method according to claim 5, wherein the mathematical correlation function comprises an intensity measurement of a gradient along different angular orientations and a comparison of the intensity measurement to at least one intensity profile characteristic of a category of an identification code.

7. Method according to one of claims 1 to 6, an identification code (ID) being able to belong to at least two categories, the method comprising a step consisting of: • Determining (E5) a probability of presence (Y0-Y2) for each category of identification code (ID) in each primary cell (Cl) of the first digital image (IMG1), • For each category of identification code (ID), determining (E6) the primary cells (Cl) having a probability of presence (Y0-Y2) greater than a predetermined secondary threshold (S6), hereinafter referred to as secondary cells (C2), • For each category of identification code (ID), determining (E7) a relevant focus area (Fp) from said secondary cells (C2), • For each category of identification code (ID), acquiring (E8) a second image (IMG2) with the relevant focus area (Fp) and • For each category of identification code (ID), carrying out (E9) a reading of an identification code (ID) in the second image (IMG2).

8. Computer program type product, comprising at least one sequence of instructions stored and readable by a processor connected to an acquisition device and which, once read by this processor, the sequence of instructions causes the steps of the method according to one of claims 1 to 7 to be carried out.

9. A computer-readable medium comprising the computer program product of claim 8.

10. System (10) for reading an identification code (ID) in a scene (SC), the reading system (10) comprising an acquisition device (11) configured to acquire several digital images (IMGO, IMG1, IMG2) of the scene (SC) and a computer (2, 12) configured to: • Command the acquisition device (11) to acquire (El) a first digital image (IMG1) of the scene (SC) with a default focus area (Fd), • Determine (E2) a motion image (IMGA) by comparing the first digital image (IMG1) to a previous digital image (IMGO), each image (IMG1, IMGO) comprising a plurality of pixels having determined positions and a value, the value of each pixel of the image of motion (IMGA) being determined by subtracting the value of a pixel of the previous digital image (IMGO) from the value of a pixel of the first digital image (IMG1) at corresponding positions, Determining (E3) in the motion image (IMGA) the pixels having a value greater than a predetermined primary threshold (SI), hereinafter referred to as primary pixels (PIX1), Determining (E4) a plurality of primary cells (Cl) from the primary pixels (PIX1) in the motion image (IMGA), Determining (E5) a probability of presence (Y0-Y2) of an identification code (ID) in each primary cell (Cl) of the first digital image (IMG1), Determining (E6) the primary cells (Cl) having a probability of presence (Y0-Y2) greater than a predetermined secondary threshold (S6), hereinafter referred to as secondary cells (C2), Determine (E7) a relevant focus area (Fp) from said secondary cells (C2), Command the acquisition device (11) to acquire (E8) a second image (IMG2) with the relevant focus area (Fp) and Perform (E9) a reading of an identification code (ID) in the second image (IMG2).

Images