Mapping method and system for mapping real environment

By using multiple marker-organized unit identification and video sequences in the environment to be mapped, the coordinate transformation is calculated, and the problem of insufficient image quality of mobile cameras is solved, and the rapid and reliable mapping effect is achieved. It is suitable for Large environment and low-quality image processing.

CN111788537BActive Publication Date: 2024-11-26SR LABS
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN201880083252.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2017-12-22
Filing Date
2018-12-21
Publication Date
2024-11-26
Estimated Expiration
2038-12-21

AI Technical Summary

Technical Problem

The prior art is difficult to quickly and reliable mapping from low- and medium quality, insufficient clarity or incomplete focus images acquired by mobile cameras.

Method used

By storing reference images of the environment to be mapped, a number of different types of markings are organized to identify in the cell and identify these marks in a video sequence to calculate the perspective transformation of the coordinates of the homography, enabling fast and reliable mapping.

Benefits of technology

Ensure the reliability and speed of the mapping in a short time, be suitable for mapping in the large environment, and even if low-quality images are obtained, recognition and mapping can be achieved, and the system is simple and economical.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN111788537B_ABST
    Figure CN111788537B_ABST
Patent Text Reader

Abstract

A mapping method, comprising: - storing at least one reference image of an environment to be mapped, the environment to be mapped comprising a specific arrangement of a plurality of markers organized in units; wherein each unit is identified by a marker or a combination of markers; the plurality of markers comprises at least a plurality of different types of markers, the number of types being equal to the number of types; - detecting a video sequence by a mobile camera, wherein the environment to be mapped is at least partially framed; - identifying one or more units in at least one frame of the video sequence, the one or more units being part of the specific arrangement of markers in the reference image; - calculating at least one homography based on data relating to the identified units, the homography being used for perspective transformation of coordinates acquired in the video sequence into coordinates in the reference image, and vice versa.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This patent application claims the priority benefit of Italian Patent Application No. 102017000149372, filed on December 22, 2017, the disclosure of which is incorporated herein by reference in its entirety. Technical Field

[0003] The invention relates to a mapping method and system.

[0004] In particular, the present invention relates to a mapping method and a system for mapping at least one real environment to be mapped. Background Art

[0005] In different technical fields, there is an increasing need to map some real locations / elements. In practice, the results derived from the mapping are often used to evaluate the attitude of the user when interacting with these specific locations or elements, thereby identifying elements of interest in an image sequence, in order to determine their location and / or display items of information about them and / or interact with the elements of interest. Some known mapping methods involve mapping the real world starting from acquired images of the real world itself.

[0006] However, in order to make the acquisition of images easy and to affect the user's behavior as little as possible during the image acquisition, miniature cameras are used and built into mobile and / or wearable devices. The images acquired in this way are images in motion, which sometimes have the characteristics of medium / low quality, lack of clarity or incomplete focus.

[0007] There is currently no way to provide a fast and reliable mapping starting from such images acquired by a mobile camera. Summary of the invention

[0008] Therefore, the object of the present invention is to provide a mapping method which allows the user to overcome the above-mentioned disadvantages in a simple and economical way from a functional point of view and a construction point of view. In other words, the object of the present invention is to provide a fast mapping method which is at the same time able to provide reliable data.

[0009] According to these purposes, the present invention relates to a mapping method comprising the following steps:

[0010] - storing at least one reference image of an environment to be mapped, the environment to be mapped comprising a specific arrangement of a plurality of markers organized in cells; wherein each cell is identified by a marker or by a combination of markers; the plurality of markers comprises at least a plurality of different types of markers, the number of types being equal to the number of types;

[0011] - detecting a video sequence by a mobile camera, in which the environment to be mapped is at least partially framed;

[0012] - identifying one or more cells in at least one frame of the video sequence, the one or more cells being part of a specific arrangement of markers in a reference image;

[0013] - Based on the data relating to the identified cells, at least one homography is calculated, the homography being used for perspective transformation of coordinates acquired in the video sequence into coordinates in the reference image and vice versa.

[0014] The method according to the invention advantageously ensures reliable mapping in a short time. The use of a specific reference arrangement in which the markers are organized in cells allows the use of different types of markers (even particularly simple ones), thus making the recognition in the video sequence reliable and fast despite the fact that the video images are acquired by a moving camera and are sometimes characterized by medium / low quality, insufficient sharpness or incomplete focus.

[0015] Furthermore, due to the division of cells, the method according to the invention also allows mapping of very large environments. In fact, the method according to the invention allows mapping of very large environments even if only a portion of the environment of interest is framed.

[0016] Another object of the present invention is to provide a mapping system for mapping a real environment, which system is quick, easy and economical to manufacture and which system is capable of providing reliable data.

[0017] According to these objects, the present invention relates to a mapping system comprising:

[0018] a mobile camera configured to detect a video sequence in which the environment to be mapped is at least partially framed;

[0019] A plurality of tags are arranged in the environment to be mapped and organized in units according to a specific configuration; wherein each unit is identified by a tag or by a combination of tags; the plurality of tags at least comprises a plurality of different types of tags whose number of types is equal to the number of types;

[0020] A control device is provided with:

[0021] a memory module configured to store at least one reference image of an environment to be mapped, the environment to be mapped comprising a specific arrangement of a plurality of markers organized in cells;

[0022] A computing module is configured to: identify one or more units in at least one frame of a video sequence, the one or more units being part of a specific arrangement of markers; and based on data related to the identified units, calculate at least one homography, the homography being used to perspective transform coordinates obtained in the video sequence into coordinates in a reference image, and vice versa. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Other features and advantages of the present invention will be best understood by reading the following description of non-limiting embodiments of the present invention with reference to the accompanying drawings, in which:

[0024] – Figure 1 is a schematic diagram of a mapping system according to the present invention;

[0025] – Figure 2 yes Figure 1 Schematic diagram of the details of the system;

[0026] – Figure 3 yes Figure 1 A schematic block diagram of more details;

[0027] – Figure 4 is a schematic flow chart relating to a part of a mapping method according to the present invention;

[0028] – Figure 5 is a schematic diagram of the steps in the mapping method according to the present invention. DETAILED DESCRIPTION

[0029] exist Figure 1 , numeral 1 indicates a mapping system for mapping a real environment according to the present invention.

[0030] In the following, reference will be made to a mapping system and method applied to a study of the purchasing attitudes of consumers in a supermarket. Obviously, the mapping method and system according to the invention can be used for other applications, for example, in the retail sector, in the industrial sector and in warehouse management, for the display of information items about a specific viewing environment or a specific area of ​​interest, or for the application of augmented reality.

[0031] The mapping system 1 comprises a camera 2 and a control device 3, wherein the control device 3 is configured to receive data acquired by the camera 2 and perform mapping.

[0032] In the non-limiting embodiment shown and disclosed herein, the camera 2 is mounted on a pair of glasses 4 in order to acquire, in motion, a sequence of images related to what the person is looking at. Since the glasses 4 are a virtually non-intrusive device, the user is not affected when he is close to the environment and / or elements to be mapped. According to a variant, the camera is a built-in camera of a "smart device" (tablet, phone, watch, etc.).

[0033] In the non-limiting embodiment shown and disclosed herein, the environment to be mapped is a shelf in a supermarket displaying products that consumers can purchase.

[0034] Figure 2 A reference image 6 of an environment to be mapped 8 is shown. This reference image 6 is usually defined in the technical field as a "snapshot".

[0035] The reference image 6 is a preferably high definition image of the environment to be mapped 8. In the non-limiting embodiment shown and disclosed herein, the reference image 6 is an image showing a generally front view of the environment to be mapped 8.

[0036] As mentioned above, in the non-limiting embodiment shown and disclosed herein, the environment 8 to be mapped is a shelf in a supermarket.

[0037] The environment to be mapped 8 includes the components organized in units 11 (in Figure 2 In other words, the markers 10 are appropriately positioned by the user before proceeding with the mapping step. As described above, the markers 10 are arranged to define one or more cells 11. In this way, the markers 10 define a coordinate system in the environment 8 to be mapped.

[0038] The tag 10 is a passive tag and does not transmit any signal.

[0039] As discussed in more detail below, each unit 11 is identified by a marker 10 or a combination of markers 10. A combination of markers refers to a specific arrangement of markers (preferably different markers) that can identify a unit. In other words, a unit is defined by markers having, for example, a given spatial arrangement, orientation, etc.

[0040] The plurality of tags 10 includes a plurality of different types of tags, the number of types being equal to the number K of types.

[0041] The number K of types of different types of markings is equal to at least 2.

[0042] According to a variant not shown here, the number K of types of different types of markings is equal to at least three.

[0043] In the non-limiting embodiment shown and disclosed herein, the number of types K of different types of markings is equal to 4. In the non-limiting embodiment shown and disclosed herein, the markings 10 all have a circular shape and have a number of different colors equal to the number K of types.

[0044] More precisely, the markings 10 have K different colors.

[0045] For example, the markings 10 may be magenta, blue, green, and orange. Figure 2In FIG. 1 , the mark 10 is schematically shown in 4 different shades of grey.

[0046] The colors of the markings are preferably classified according to spatial LAB color coordinates.

[0047] Alternatively, the color of the indicia 10 is automatically identified by a suitable color classification algorithm.

[0048] In the non-limiting embodiment shown and disclosed herein, the markers 10 have the same diameter.

[0049] According to a variant not shown herein, the markings may have different diameters. For example, some cells may have markings with a larger diameter than the diameter of the remaining markings, so that these markings can also be easily detected at a long distance. In this way, some cells may be detected at a given distance, while other cells may be detected at different distances. Thus, the number of large markings may be minimized and the visual impact on the user may be maximized.

[0050] The markings 10 are preferably arranged on a white background. In this way, the colors (magenta, blue, green and orange) selected for the markings 10 are more easily recognizable. Obviously, for markings with different shapes and / or different colors, the background may be different.

[0051] In the non-limiting embodiment shown and disclosed herein, a cell 11 is defined by four markers 10 arranged at the vertices of a quadrilateral.

[0052] Figure 2 A specific arrangement of the units 11 (shown by dashed lines) is shown. In the non-limiting embodiments shown and disclosed herein, some of the units 11a may include other smaller units 11b. Obviously, the specific arrangement of the units 11 may vary depending on the type of application and / or the type of environment to be monitored. For example, the units 11 may be arranged relative to each other as needed. Figure 2 The configuration shown in is more or less close to each other.

[0053] Figure 3 A block diagram of the control device 3 is shown.

[0054] The control device 3 includes a memory module 15 and a calculation module 16 .

[0055] The memory module 15 stores the reference image 6 of the environment to be mapped 8 and information items (SNAPSHOTDATA) associated therewith, including for example the type of marking 10, the specific arrangement of the cells 11, etc., and other information items, such as predefined areas of interest (AOI).

[0056] The calculation module 16 is configured to receive a video sequence (VIDEO DATA) detected by the camera 2, in which the environment 8 to be mapped is at least partially framed, and to identify in at least one frame of the video sequence one or more cells 11 that are part of a specific arrangement of markings 10 of a reference image 6 contained in the data (SNAPSHOT DATA) from the memory module 15. As discussed in more detail below, the calculation module 16 is configured to determine mapping data (MAPPING DATA) starting from the video sequence (VIDEO DATA) detected by the camera 2 and starting from the data (SNAPSHOT DATA) from the memory module 15.

[0057] Optionally, the control device 3 may include a projection module 17 and / or a graphics module 18 and / or a statistics module 19, which are connected to the memory module 15 and / or the calculation module 16 via interfaces. These modules will be described in more detail below.

[0058] Reference Figure 4 , the computing module 16 is configured as follows:

[0059] - (step 100) identifying a candidate marker set MARKERSCANDIDATE SET in at least one frame of the video sequence VIDEO DATA;

[0060] - (Step 101) Eliminate possible false positive candidate markers MFP from the marker set MARKERS CANDIDATE SET to obtain a reliable candidate marker set MARKERS RELIABLE SET (marker reliable set);

[0061] - (step 102) identifying all combinations of candidate markers in a reliable set MARKERS RELIABLE SET that can be associated with cells 11 in a particular arrangement of the reference image 6 to obtain a candidate cell set CANDIDATE CELLS SET;

[0062] - (step 103) eliminating possible false positive cells CFP from the candidate cell set CANDIDATE CELLS SET to obtain a reliable candidate cell set CELLS RELIABLE SET (cell reliable set); and

[0063] - (step 104) Starting from the reliable candidate cell set CELLS RELIABLE SET and from the data SNAPSHOT DATA from the memory module 15, a homography (MAPPING DATA) is calculated, which is able to transform the coordinates of the cells in the reliable cell set CELLS RELIABLE acquired at a given viewing angle into coordinates that can be represented in the viewing angle of the reference image 6, and vice versa. In the following, a homography refers to a relationship between points of two spaces, so that each point of one space corresponds to a single point of the second space. This relationship is usually represented by means of a matrix (called a homography matrix) or a system of equations.

[0064] In particular, the step of identifying a candidate marker set MARKERSCANDIDATE SET in at least one frame of the video sequence VIDEO DATA comprises detecting all elements that can be associated with a marker and considering them as part of the candidate marker set MARKERS CANDIDATE SET.

[0065] The step of eliminating false positives (step 101) is performed by evaluating some parameters of candidate markers, which are part of the candidate marker set MARKERS CANDIDATE SET.

[0066] In the non-limiting embodiment shown and disclosed herein, the mark 10 is circular and has different colors; the elimination step is based on an evaluation of the area of ​​the candidate mark and / or the brightness of the candidate mark and / or the roundness of the candidate mark and / or the moment of inertia of the candidate mark and / or the convexity of the candidate mark and / or the color of the candidate mark, and / or based on a reference data set of classified false positives.

[0067] For example, a reference data set of false positives may be defined by an automatic learning algorithm and / or may be defined based on experimental data.

[0068] Figure 5 Some examples of parameters on which the false positive elimination step (step 101) is based are shown.

[0069] Point a) shows candidate markers with regions of different sizes.

[0070] Point b) shows candidate markers with different brightness, where the brightness represents the intensity of the binary image center of the candidate marker.

[0071] Point c) shows candidate markers with different roundness indicators.

[0072] Point d) shows candidate markers with different inertia indicators, where inertia refers to the dispersion of all pixels belonging to the candidate marker around its centroid.

[0073] Point e) shows candidate markers with different convexity indicators.

[0074] The evaluation of candidate marker colors is preferably based on spatial LAB color coordinates. The spatial LAB color coordinates are the following three: a, b, L. However, the evaluation of candidate marker colors is preferably performed by detecting the unique coordinates a, b of the candidate marker, and comparing them starting from the color assigned to the marker 10 at acceptable intervals of the coordinates a, b.

[0075] Obviously, as described above, the information item about the color assigned to the mark 10 is contained in the data SNAPSHOT DATA from the memory module 15 .

[0076] In this way, no comparison is made of the coordinates L that depend on the brightness. By doing so, elimination of candidate markers having the color of the marker 10 but a different brightness (a parameter that is overly dependent on the image sequence detection conditions) can be avoided.

[0077] Alternatively, the color of the indicia 10 is automatically identified by a suitable color classification algorithm.

[0078] At the end of the elimination step (step 101), there exists a reliable candidate marker set MARKERS RELIABLE SET.

[0079] The step of identifying a candidate cell set CELLS CANDIDATE SET (step 102 ) involves, starting from the markers that are part of the reliable candidate marker set MARKERS RELIABLE SET, detecting combinations of candidate markers that can be associated with a cell in the cells 11 that is part of a specific arrangement of the reference image 6 .

[0080] In the non-limiting embodiment shown and disclosed herein, the cells 11 that are part of a particular arrangement of the reference image 6 are quadrilateral. Therefore, the candidate cells are formed by candidate markers arranged at the vertices of the quadrilateral.

[0081] Once all candidate cells are identified, the step of eliminating false positive cells (step 103) is basically performed in four steps.

[0082] The first step involves eliminating all candidate cells that do not have at least one of the combinations of labels 10 of cell 11 .

[0083] In other words, candidate cells are all those cells formed by candidate markings having the same arrangement and color as one of the specific combinations defining the cell 11 in the reference image 6 .

[0084] The second step involves eliminating candidate cells that represent projections of cells 11 that are derived from assumed implausible and unrealistic framings. When these cells are considered false positives, they may produce implausible projections, i.e., projections that cannot be derived from actual framings. These cells can be eliminated a priori.

[0085] The third step involves eliminating candidate cells whose ACCAND / AMCAND ratio (cell area / area of ​​the marker defining the candidate cell) does not match the AC SNAPSHOT / AM SNAPSHOT The ratio (cell area / area of ​​the markings of cells 11 in a particular arrangement of the reference image 6 having the same marking combination as the candidate cell) corresponds.

[0086] The fourth step involves computing the homography starting from the data of all candidate cells.

[0087] If the homography computed by doing this corresponds to a reasonable projection of the particular arrangement of the reference image 6, the candidate cell is considered to be reliable.

[0088] If the homography does not correspond to a reasonable projection of a particular arrangement of the reference image 6, at least some candidate cells are eliminated.

[0089] At the end of the four cell selection steps, there exists a reliable cell set CELLS RELIABLE SET.

[0090] Starting from a reliable candidate cell set CELLS RELIABLE SET and from data SNAPSHOT DATA from a memory module 15, the method comprises: for at least some frames of a video sequence (VIDEO DATA), calculating a homography (MAPPING DATA) capable of transforming the coordinates of cells in the reliable cell set CELLS RELIABLE SET acquired at a given viewing angle into coordinates representable in the viewing angle of a reference image 6, and vice versa.

[0091] In other words, the method consists in calculating: a homography capable of obtaining a perspective transformation of coordinates acquired through a sequence of images into coordinates representable in a reference image, or vice versa, a homography capable of obtaining a perspective transformation of coordinates in a reference image into coordinates representable in the sequence of images.

[0092] As described above, homographies are computed for at least some frames of the video sequence VIDEO DATA.

[0093] The method according to the invention preferably comprises calculating at least one homography of at least one frame and determining the homographies of the remaining frames starting from this homography and the images of the respective frames, preferably by means of a computer vision algorithm (e.g. an “optical flow” algorithm) or by means of an interpolation algorithm.

[0094] On the other hand, sometimes, it is preferable to calculate multiple homographies and obtain an uncomputable homography between the computable homographies by the above method.

[0095] The homography obtained by doing this defines the mapping of the environment to be monitored (MAPPING DATA). These matrices can be used in a variety of applications.

[0096] In the non-limiting example disclosed and shown herein, the mapping data MAPPING DATA are preferably supplied to the projection module 17, which is configured to combine the mapping data MAPPING DATA from the calculation module 16 with possible data about the eye movements of the user who at least partially looks at the environment to be mapped 8. The data about the eye movements of the user are transmitted via the eye tracking device 30 (in Figure 1 The eye movement is detected by the eye movement calculation module 20 (schematically shown as part of the glasses 4) and is connected to the control device 3.

[0097] The eye tracking device 30 detects data related to the eye movements of the user who is at least partially looking at the environment to be mapped 8 during the video sequence acquisition, and the eye movement calculation module 20 processes the data acquired by the eye tracking device 30 and sends the gaze coordinates GAZES ON VIDEO on the video sequence to the projection module 17.

[0098] According to a variant not shown herein, the eye movement calculation module 20 is built into the eye tracking device 30 and the module is not part of the control device 3. The projection module 17 is configured to process the gaze coordinates GAZES ON VIDEO from the eye movement calculation module 20 based on the homography (MAPPING DATA) determined by the calculation module 16. In this way, the projection module 17 calculates the gaze point coordinates GAZES ON SNAPSHOT on the reference image 6.

[0099] Optionally, the control device 3 may further comprise a graphics module 18 configured to represent the unit 11 identified by the calculation module 16 on the video sequence and / or the reference image 6 and, if necessary, also display gaze coordinates GAZES ON VIDEO from the eye movement calculation module 20 on the reference image 6 .

[0100] The graphics module 18 is also configured to process the input data and, if necessary, represent heat maps (usually defined as HEAT MAPS) and / or GAZE PLOTS (gaze maps indicating the sample locations of gaze points or fixations, the order of these samples or fixations, and the time spent viewing) and / or GAZE OPACITYMAPS (gaze opacity maps showing only the area where the user's attention is focused, while the rest of the area is masked or opaque) on the reference image.

[0101] Optionally, the control device 3 may also include a statistical module 19 configured to process data from the projection module 17 (GAZES ON SNAPHOT) and data from the memory module 15 (SNAPSHOT DATA, AOI) to provide statistical processing of the data (METRICS, metrics) useful for different applications and evaluations.

[0102] The mapping method and system according to the invention advantageously allow mapping the real environment to be monitored in a simple and reliable manner starting from video sequences detected in motion, even under conditions where it is determined that low-quality images are acquired.

[0103] The mapping data (MAPPING DATA) obtained by the mapping method and system according to the present invention can be used for statistical processing (performed by the statistical module 19) and / or for data in applications (not described and shown herein) that allow the representation of important data and / or involve the interaction of the user with the mapped environment / elements.

[0104] All steps of the method according to the invention are performed automatically. This allows "real-time" mapping.

[0105] Furthermore, the method and system according to the present invention are sufficiently reliable and robust. For example, they are not affected by the acquisition of low-quality images, the acquisition of partial images of the environment to be mapped, the acquisition of partially obstructed parts of the environment to be mapped, and the viewing angle of the acquired images.

[0106] Furthermore, the mapping data obtained by the method and system according to the present invention are accurate and are able to provide an indication of the distance from which the sequence of images of the mapped environment / element was acquired.

[0107] Finally, the system according to the invention is simple and economical, also due to the fact that it preferably uses passive markers (ie inactive markers, such as markers that emit a signal) and does not require special maintenance.

[0108] Finally, it is obvious that the mapping system and method described herein may be modified and varied without thereby exceeding the scope of protection of the present disclosure.

Claims

1. A mapping method, comprising the following steps: - positioning a plurality of passive markers (10) in an environment to be mapped (8); the plurality of markers (10) at least comprises a plurality of different types of markers (10) whose number of types is equal to the number of types (K); the plurality of markers (10) are organized in units (11); wherein each unit (11) is identified by a combination of markers (10); the combination of markers (10) is a specific arrangement of markers capable of identifying the unit (11); - storing at least one reference image (6) of an environment (8) to be mapped, said environment (8) to be mapped comprising said plurality of markings (10) organized in cells (11); - detecting a video sequence by a mobile camera (2), in which the environment (8) to be mapped is at least partially framed; - identifying in at least one frame of the video sequence one or more cells (11) which are part of the specific arrangement of markers (10) in the reference image (6); - Based on the data relating to the identified cells, at least one homography is calculated, the homography being used to perspective transform the coordinates acquired in the video sequence into coordinates in the reference image (6), and vice versa.

2. The method according to claim 1, wherein the number (K) of different types of markings (10) is equal to at least 2.

3. The method according to claim 1, wherein the plurality of markings (10) comprises markings (10) having a circular shape.

4. The method according to claim 1, wherein the plurality of marks (10) includes marks (10) having a plurality of different colors, the number of types of the plurality of different colors being equal to the number of types (K).

5. The method according to claim 1, wherein identifying one or more cells (11) in at least one frame of the video sequence comprises at least: identifying a set of candidate tags in at least one frame of the video sequence; Eliminating false positive potential candidate markers MFP from the candidate marker set to obtain a reliable candidate marker set; identifying all combinations of candidate labels in the reliable candidate label set that can be associated with the unit (11) of the reference image (6) to obtain a candidate unit set; Eliminate false positive potential cells CFP from the candidate cell set to obtain a reliable candidate cell set.

6. The method according to claim 5, wherein the step of eliminating false positive potential candidate markers MFP from the candidate marker set comprises: Eliminate the candidate marker based on an evaluation of the area of ​​the candidate marker and / or the brightness of the candidate marker and / or the roundness of the candidate marker and / or the moment of inertia of the candidate marker and / or the convexity of the candidate marker and / or the color of the candidate marker, and / or based on a reference data set of classified false positives.

7. The method according to claim 1, comprising the steps of: While the video sequence is being acquired, a data sequence at least relating to eye movements of a user who is at least partially looking at the environment (8) to be mapped is detected by an eye tracking device (30).

8. The method according to claim 7, comprising the steps of: The data sequence acquired by the eye tracking device (30) is processed to obtain the gaze point coordinates on the video sequence.

9. The method according to claim 8, comprising the steps of: The gaze point coordinates on the video sequence are processed based on the homography to obtain reference gaze point coordinates on the reference image (6).

10. A mapping system comprising: a mobile camera (2) configured to detect a video sequence in which the environment to be mapped (8) is at least partially framed; A plurality of passive markers (10) are arranged in the environment to be mapped (8) and organized in units (11); each unit (11) is identified by a combination of markers (10); the plurality of markers (10) at least comprises markers of a plurality of different types, the number of which is equal to the number of types (K); A control device (3) is provided with: A memory module (15) configured to store at least one reference image (6) of the environment to be mapped (8), the environment to be mapped (8) comprising the plurality of markings (10) organized in cells (11); A calculation module (16) is configured to: identify one or more units (11) in at least one frame of the video sequence; and calculate at least one homography based on data related to the identified units, the homography being used to perspective transform coordinates obtained in the video sequence into coordinates in the reference image (6), and vice versa.

11. The system according to claim 10, wherein the number (K) of different types of markings in the markings (10) is equal to at least 2.

12. The system of claim 10, wherein the plurality of markers (10) comprises markers (10) having a circular shape.

13. The system according to claim 10, wherein the plurality of markers (10) includes markers (10) having a plurality of different colors, the number of types of the plurality of different colors being equal to the number of types (K).

14. The system according to claim 10, wherein the computing module (16) is configured to: identifying a set of candidate tags in at least one frame of the video sequence; Eliminating false positive potential candidate markers MFP from the candidate marker set to obtain a reliable candidate marker set; identifying all combinations of candidate labels in the reliable candidate label set that can be associated with the unit (11) of the reference image (6) to obtain a candidate unit set; Eliminate false positive potential cells CFP from the candidate cell set to obtain a reliable candidate cell set.

15. A system according to claim 14, wherein the computing module (16) is configured to eliminate the false positive potential candidate marker MFP from the candidate marker set based on an evaluation of the area of ​​the candidate marker and / or the brightness of the candidate marker and / or the roundness of the candidate marker and / or the moment of inertia of the candidate marker and / or the convexity of the candidate marker and / or the color of the candidate marker, and / or based on a reference data set of classified false positives.

16. The system according to claim 10, comprising: An eye tracking device (30) is configured to detect, while the video sequence is being acquired, a data sequence at least related to eye movements of a user who is at least partially looking at the environment to be mapped (8).

17. The system according to claim 16, comprising: The eye movement calculation module (20) is configured to process the data sequence acquired by the eye tracking device (30) to obtain the gaze point coordinates on the video sequence.

18. The system according to claim 17, comprising: A projection module (17) is configured to process the gaze point coordinates on the video sequence based on the homography to obtain reference gaze point coordinates on the reference image (6).