Magnetic field mapping method and device based on implicit expression

Through the implicitly expressed magnetic field mapping method, a high-resolution magnetic field map is generated using decoding functions and MLP networks, solving the problem of time-consuming construction of magnetic field maps in geometric repetitive environments, and achieving efficient and accurate magnetic field positioning, which is suitable for mobile robots Global positioning.

CN119006290AActive Publication Date: 2024-11-22HUAZHONG UNIV OF SCI & TECH
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
CN202411037335.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-31
Publication Date
2024-11-22
Estimated Expiration
2044-07-31

AI Technical Summary

Technical Problem

In GPS-constrained, geometric repetitive environments, the process of building high-resolution magnetic field maps is time-consuming and inefficient. Traditional methods such as linear interpolation and Gaussian process regression fail to effectively consider the inherent characteristics of the magnetic field, and deep learning methods lack accurate measurements due to the lack of accurate measurements in the deep learning method. Inefficient in location open source datasets.

Method used

Using the magnetic field mapping method based on implicit expression, the low-resolution magnetic field map is converted into a low-resolution magnetic field image, and the decoding function fθ is used to map it into a 2D feature map, and the high-resolution magnetic field image is predicted through the MLP network. RGB values ​​generate high-resolution magnetic field maps, use raster diagrams to express magnetic field data, and optimize formulas to ensure the selection and overlap of hidden codes, achieving continuous magnetic field image representation.

Benefits of technology

It significantly improves the efficiency and accuracy of magnetic field map construction, reduces positioning errors in geometric repetition environments, achieves effective promotion in invisible locations, and shows robust performance at higher resolutions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119006290A_ABST
    Figure CN119006290A_ABST
Patent Text Reader

Abstract

The invention belongs to the related technical field of mobile cabinet robot perception, and discloses a magnetic field mapping method and device based on implicit expression, and the method comprises the following steps: S1, converting a low-resolution magnetic field map into a low-resolution magnetic field image; s2, the encoder # imgabs0 # maps the obtained low-resolution magnetic field image into a 2D feature map; s3, inputting the obtained 2D feature map and the center coordinate of each pixel in the high-resolution magnetic field image into a decoding function f theta, wherein the decoding function f theta predicts the RGB value of each pixel position in the high-resolution image; wherein the decoding function f theta is adopted to implicitly represent the continuation magnetic field image; and S4, generating a high-resolution magnetic field image based on the RGB value of each pixel position in the high-resolution magnetic field image, and then obtaining a high-resolution magnetic field map based on the high-resolution magnetic field image. According to the method, the problems of time consumption and low efficiency of construction of an existing dense magnetic field map are solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field related to mobile robot perception, and more specifically, to a magnetic field mapping method and device based on implicit expression. Background Art

[0002] Accurate and reliable global localization is key to enabling various mobile robot functions such as mapping, navigation, and intelligence. Accurate global localization remains a challenging problem for mobile robots operating in GPS-restricted, geometrically repetitive environments, such as service robots in offices and hotel corridors, security robots in parking lots, and industrial robots in warehouses. These geometrically repetitive environments often lack unique geometric features due to their similar structures. When local localization fails or the robot is disoriented or abducted, it can be difficult to estimate the robot's initial pose (i.e., initialization) or recover its state. Magnetic fields have distinct characteristics at different indoor locations and are not affected by the typical repetitive geometric features of such environments. Therefore, they can be used as an alternative to GPS signals for localization and navigation in indoor environments. For robots facing the challenges of repetitive indoor environments, using magnetic field information for localization is an effective strategy. Previous studies have verified the feasibility of using magnetic fields for indoor localization and demonstrated the stability of magnetic fields over long periods of time. Typically, the construction of magnetic field-based indoor localization systems requires the construction of grid-based magnetic field maps, and the quality of these maps is critical to the accuracy of localization.

[0003] Due to the limited measurement range of the magnetometer and the vast operating environment of the robot, it is impossible for the robot to collect magnetic field data at every location. Because of the need for intensive data collection and the limitations of the measurement range of current sensors, building large-scale dense magnetic field maps faces great challenges. Each data point in a high-resolution magnetic field map (such as a magnetic field map with a resolution of 1 cm) requires the magnetic field intensity to be measured separately at precise intervals on the x-axis and y-axis. This requirement requires thousands of measurements over a fairly large area, which is a labor-intensive and time-consuming process. Therefore, it is crucial to model the magnetic field and build a dense map based on the collected sparse magnetic field data. Traditionally, people use methods such as linear interpolation to construct a complete magnetic field map. However, the linear interpolation method does not take into account the inherent characteristics of the magnetic field. Some people also use the Gaussian process regression (GPR) method to model the three coordinate axes of the magnetic field by using separate Gaussian processes, but this method does not consider the connection between the magnetic field intensity in each axis direction at the same location.

[0004] In recent years, learning-based methods have become increasingly popular for predicting surrounding environments. However, collecting comprehensive indoor magnetic field data is time-consuming and laborious. This complexity makes deep learning methods for magnetic field mapping inefficient because such methods require a large amount of training data. In addition, the lack of open-source magnetic field datasets with precise measurement locations makes it more difficult to train deep learning networks to predict magnetic fields. Summary of the invention

[0005] In view of the above defects or improvement needs of the prior art, the present invention provides a magnetic field mapping method and device based on implicit expression, which aims to solve the problem that the construction of existing dense magnetic field maps is time-consuming and inefficient.

[0006] To achieve the above object, according to one aspect of the present invention, a magnetic field mapping method based on implicit expression is provided, the method comprising the following steps:

[0007] S1, converting low-resolution magnetic field maps into low-resolution magnetic field images;

[0008] S2, encoder Mapping the obtained low-resolution magnetic field image into a 2D feature map;

[0009] S3, input the obtained 2D feature map and the center coordinates of each pixel in the high-resolution magnetic field image into the decoding function f θ , decoding function f θ Predict the RGB value of each pixel position in the high-resolution image; the decoding function f is used θ To implicitly represent the continuous magnetic field image;

[0010] S4, generating a high-resolution magnetic field image based on the RGB value of each pixel position in the high-resolution magnetic field image, and then obtaining a high-resolution magnetic field map based on the high-resolution magnetic field image; wherein the resolution of the high-resolution magnetic field image is greater than the resolution of the low-resolution magnetic field image, and the resolution of the high-resolution magnetic field map is greater than the resolution of the low-resolution magnetic field map.

[0011] Furthermore, the magnetic field map is expressed using a grid map. The size of each grid in the grid map is defined by the resolution of the magnetic field map. The resolution is represented by the side length r of each grid. The overall size of the magnetic field map is: length L x and width L y , each grid in the magnetic field map contains specific information: the grid center coordinates (x, y) and the corresponding magnetic field intensity component B x ,B y ,B z ; The magnetic field database is represented in the form of a magnetic field image. The obtained magnetic field image is constructed with a resolution of H×W pixels, where H and W are directly determined by the map size L x, L y And the resolution r is obtained, the calculation formula is:

[0012]

[0013] The symbol Represents the floor function.

[0014] Furthermore, the pixel coordinates (p x ,p y ) is mapped from the grid coordinates (x, y) in the magnetic field map, and the corresponding formula is:

[0015]

[0016] The magnetic field strength B is normalized using the following formula: x ,B y ,B z Normalize to fit the RGB color model range (0-255):

[0017]

[0018] Among them B x,min ,B x,max ,B y,min ,B y,max ,B z,min and B z,max It is the B on the entire magnetic field map x ,B y ,B z The minimum and maximum values ​​of the RGB values ​​are assigned to the corresponding pixel positions in the magnetic field image (p x ,p y ).

[0019] Further, each magnetic field image is represented as B (i) , the pixel data in the low-resolution magnetic field image is mapped into a two-dimensional feature map All magnetic field images share a decoding function f θ , decoding function f θ For the MLP network, the formula used in the decoding process is:

[0020] s=f θ (z,x);

[0021] Where z is a vector, x∈X is the two-dimensional coordinate in the continuous magnetic field image domain, and s∈S is the predicted signal, i.e., the RGB value representing the magnetic field intensity.

[0022] Furthermore, the feature vector z of the two-dimensional feature map M(i) is referred to as the hidden code. For the continuous magnetic field image B(i), the coordinate x qThe RGB values ​​are:

[0023] B (i) (x q ) = f θ (z * ,x q -v * ); (1)

[0024] where z * is the distance x q The latest hidden code, in M (i) Medium * It is hidden code z * Coordinates in the magnetic field image domain.

[0025] Furthermore, using the decoding function f θ Representing continuous magnetic field images as two-dimensional feature maps Conceptualized as a latent array uniformly distributed in a 2D spatial region, a two-dimensional feature map M (i) Each hidden code z in * Both contain a local segment information of the continuous magnetic field image.

[0026] Furthermore, formula (1) is optimized as follows:

[0027]

[0028] in are the nearest hidden codes in the upper left, upper right, lower left, and lower right subspaces respectively, yes The coordinates of S t is x q and The area of ​​the rectangle between t′ and t′ is the diagonal of t, and the weight is S = ∑ t S t Normalize.

[0029] The present invention also provides a magnetic field mapping system based on implicit expression, the system comprising a memory and a processor, the memory storing a computer program, and the processor executing the magnetic field mapping method based on implicit expression as described above when executing the computer program.

[0030] The present invention also provides a computer-readable storage medium, which stores machine-executable instructions. When the machine-executable instructions are called and executed by a processor, the machine-executable instructions prompt the processor to implement the magnetic field mapping method based on implicit expression as described above.

[0031] The present invention also provides an application of the magnetic field mapping method based on implicit expression as described above in the global positioning of a mobile robot.

[0032] In general, compared with the prior art, the above technical solutions conceived by the present invention mainly have the following beneficial effects:

[0033] 1. The present invention uses coordinate x hr Query on the 2D feature map, x hr is the center coordinate of each pixel in the high-resolution magnetic field image; then the function f θ Predict the signal value s of each coordinate based on the 2D feature map hr (RGB value representing magnetic field intensity), generates a high-resolution magnetic field image by predicting the RGB value of each pixel position in the high-resolution image, and then converts the generated high-resolution magnetic field image into a high-resolution magnetic field map, realizing the mapping of the magnetic field map, providing a continuous expression, which can be effectively generalized to unseen positions when used, and saves time.

[0034] 2. The low-resolution magnetic field image is converted into a high-resolution magnetic field map through a two-dimensional feature map and a decoding function, and a continuous magnetic field image B is proposed. (i) , coordinate x q The calculation formula of RGB value.

[0035] 3. Optimizing formula (1) to formula (2) solves the problem when x q When moving in a two-dimensional domain, the hidden code z * In order to solve the problem that the choice of the latent code suddenly switches from one to another (i.e., the choice of the most recent latent code changes), we allow the local fragment represented by a single latent code (the feature in the feature map) to overlap with its neighboring segments. This overlap ensures that each specific coordinate is associated with four different latent codes, each of which independently contributes to the signal prediction at that position; at the same time, this method helps to achieve smooth and continuous transitions at the changed coordinates, thereby ensuring the continuity of the prediction output across spatial domains. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is a schematic diagram of continuous expression of magnetic field images involved in the present invention;

[0037] Figure 2 It is a schematic diagram of generating a high-resolution magnetic field image involved in the present invention. DETAILED DESCRIPTION

[0038] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0039] See also Figure 1 and Figure 2 The present invention provides a magnetic field mapping method based on implicit expression, and the magnetic field mapping method mainly includes the following steps:

[0040] S1, converting low-resolution magnetic field maps into low-resolution magnetic field images.

[0041] Typically, the magnetic field map used for robot positioning consists of two parts: the two-dimensional position coordinates (x, y) and the three-dimensional magnetic field vector B. The magnetic field vector B consists of three components: B x , B y and B z , representing the strength in the X, Y and Z directions respectively. The magnetic field map is expressed using a grid map. The size of each grid in the grid map is defined by the resolution of the magnetic field map. The resolution is represented by the side length r of each grid. The smaller the resolution, the higher the quality of the magnetic field map. The overall size of the magnetic field map is: length L x and width L y Each grid in the magnetic field map contains specific information: the grid center coordinates (x, y) and the corresponding magnetic field intensity component B x ,B y ,B z .

[0042] Generally speaking, the magnetic field intensity changes continuously in space and has a unique spatial distribution. Color information also has spatial distribution characteristics in the image. Each pixel has a different RGB value, forming an overall color map, which shows a certain spatial distribution law, which is helpful for the location and identification of the target. In addition, both the magnetic field intensity information and the color information contain three components. Therefore, the magnetic field database is represented in the form of a magnetic field image. The obtained magnetic field image is constructed with a resolution of H×W pixels, where H and W are directly determined by the map size L. x , L y And the resolution r is obtained, the calculation formula is:

[0043]

[0044] The symbol represents the floor function, which rounds a real number down to the previous largest integer. Each pixel in the magnetic field image corresponds to a unit in the magnetic field map, and the pixel coordinates in the magnetic field image (px ,p y ) is mapped from the grid coordinates (x, y) in the magnetic field map, and the corresponding formula is:

[0045]

[0046] The magnetic field strength B is normalized using the following formula: x ,B y ,B z Normalize to fit the RGB color model range (0-255):

[0047]

[0048] Among them B x,min ,B x,max ,B y,min ,B y,max ,B z,min and B z,max It is the B on the entire magnetic field map x ,B y ,B z The obtained RGB values ​​are then assigned to the corresponding pixel positions (p x ,p y ), thereby effectively converting the magnetic field data into an image format. In this way, the magnetic field map can also be processed into a magnetic field image, and a corresponding relationship between the magnetic field map and the magnetic field image is established.

[0049] S2, encoder The obtained low-resolution magnetic field image is mapped into a 2D feature map.

[0050] Using a neural network (decoding function f θ ) is used to predict the RGB value (representing the magnetic field intensity) of the center coordinate of each pixel in the high-resolution magnetic field image as an implicit representation of the continuous magnetic field image. The decoding function f θ To implicitly express the continuous magnetic field image; each magnetic field image is represented as B (i) , the pixel data in the low-resolution magnetic field image is mapped into a two-dimensional feature map All magnetic field images share a decoding function f θ , decoding function f θ For the MLP network, the formula used in the decoding process is:

[0051] s=f θ (z,x);

[0052] Where z is a vector, x∈X is the two-dimensional coordinate in the continuous magnetic field image domain, and s∈S is the predicted signal, that is, the RGB value representing the magnetic field intensity. With the decoding function f already definedθ , each vector z can be viewed as a function f that maps coordinates to RGB values θ (z,·):X→S. This embodiment assumes that the feature vector z (referred to as hidden code) of the two-dimensional feature image M(i) is uniformly distributed in the two-dimensional space of the continuous magnetic field image B(i) (e.g. Figure 1 As shown in ), and assign a two-dimensional coordinate to each hidden code z. For the continuous magnetic field image B(i), the coordinate x q The RGB values ​​are:

[0053] B (i) (x q ) = f θ (z * ,x q -v * ); (1)

[0054] where z * is the distance x q The latest hidden code, in M (i) Medium * It is hidden code z * Coordinates in the magnetic field image domain. Figure 1 For example, is the current coordinate x q The hidden code, v * It is a hidden code The image domain coordinates.

[0055] Using the decoding function f θ Representing continuous magnetic field images as two-dimensional feature maps Conceptualized as a latent array uniformly distributed in a 2D spatial region. (i) Each hidden code z in * They all contain a local fragment of the continuous magnetic field image; the specific function of the latent code is to predict the RGB values ​​of a set of defined coordinates close to it.

[0056] Due to x q The predicted signal value at is obtained by querying the two-dimensional feature map M (i) The nearest hidden code z * Information is obtained when x q When moving in a two-dimensional domain, the hidden code z * The choice of x suddenly switches from one to another (i.e. the choice of the most recent hidden code changes). For example, when x q and Figure 1 This happens when the dashed lines in intersect. Around those coordinates that intersect with the dashed lines, different hidden codes will be used to predict the signal values ​​of two infinitely close coordinates, which is not continuous enough.

[0057] To solve this problem, Figure 1 As shown, formula (1) is expanded to:

[0058]

[0059] in are the nearest hidden codes in the upper left, upper right, lower left, and lower right subspaces respectively, yes The coordinates of S t is x q and The area of ​​the rectangle between t′ and t′, where t′ is the diagonal of t (i.e. 00 to 11, 10 to 01), and the weight is S = ∑ t S t Normalize.

[0060] The present invention allows the local segments represented by a single latent code (feature in the feature map) to overlap with its neighboring segments. This overlap ensures that each specific coordinate is associated with four different latent codes, each of which independently contributes to the signal prediction at that location; the corresponding four predictions are then integrated through a consensus mechanism that uses a voting system weighted by a standardized confidence score. The confidence score is calculated proportionally to the area of ​​the rectangle formed between the query point and the diagonal corresponding point of the nearest latent code, and its value increases as the query coordinates get closer to the latent code. This approach helps to achieve smooth and continuous transitions at coordinates that change, such as Figure 1 This is shown by the dotted line in , thereby ensuring the continuity of the prediction output across the spatial domain.

[0061] S3, input the obtained 2D feature map and the center coordinates of each pixel in the high-resolution magnetic field image into the decoding function f θ , decoding function f θ Predict the RGB value of each pixel position in the high-resolution image; the decoding function f is used θ To implicitly represent the continuous magnetic field image.

[0062] Based on the proposed decoding function f θ A high-resolution magnetic field image can be rendered from the two-dimensional feature map represented by the low-resolution magnetic field image. Figure 2 As shown, the input low-resolution magnetic field image is generated by a magnetic field map with a sparse grid. This embodiment uses ESDR as an encoder Encoder The input low-resolution magnetic field image is mapped to a 2D feature map, which has the same size as the input low-resolution magnetic field image; then the coordinate x is used hr Query on the 2D feature map, x hr is the center coordinate of each pixel in the high-resolution magnetic field image; then the function fθ Predict the signal value s of each coordinate based on the 2D feature map hr (RGB value representing magnetic field strength). Decoding function f θ It is a 5-layer MLP with ReLU activation layers and 256 hidden dimensions. It generates a high-resolution magnetic field image by predicting the RGB value of each pixel position in the high-resolution image, and then converts the generated high-resolution magnetic field image into a high-resolution magnetic field map to achieve mapping of the magnetic field map.

[0063] S4, generating a high-resolution magnetic field image based on the RGB value of each pixel position in the high-resolution magnetic field image, and then obtaining a high-resolution magnetic field map based on the high-resolution magnetic field image; wherein the resolution of the high-resolution magnetic field image is greater than the resolution of the low-resolution magnetic field image, and the two are relative; similarly, the resolution of the high-resolution magnetic field map is greater than the resolution of the low-resolution magnetic field map.

[0064] The performance of three different magnetic field mapping methods (bilinear interpolation (BI), Gaussian process regression (GPR) and the method proposed in this invention) were evaluated on three data sets with different grid resolutions (20 cm, 15 cm and 5 cm). The evaluation indicators included the average error (Ave), the mean square error (MSE) and the error range (maximum and minimum values).

[0065] For 20 cm grid mapping, the proposed method shows significantly lower average error (Ave = 0.031 G) and RMSE (0.0443 G) compared to BI (Ave = 0.0324 G, RMSE = 0.0477 G) and GPR (Ave = 0.1178 G, RMSE = 0.1355 G). Similarly, the maximum error recorded by the proposed method (maximum = 0.2136 G) is significantly smaller than the error observed using GPR (maximum = 0.5036 G).

[0066] In the 15 cm grid scenario, the proposed method outperforms other methods, with the lowest RMSE (0.0377G), and the average error (Ave = 0.0247G) is also very competitive, which is almost the same as BI (Ave = 0.0238G, RMSE = 0.0443G), and is significantly better than GPR (Ave = 0.115G, RMSE = 0.0169G). The maximum error of the proposed method is 0.23G, which is significantly lower than BI (0.1764G) and GPR (0.48G), indicating that it has robust performance even at finer grid resolutions.

[0067] The most demanding 10 cm grid mapping further verifies the effectiveness of the proposed method. Compared with BI (average = 0.0144G, RMSE = 0.0282G) and GPR (average = 0.1094G, RMSE = 0.1208G), the proposed method achieves the lowest RMSE (0.0264G) and a very competitive average error (average = 0.0152G). It is worth noting that the proposed method also records the smallest maximum error (maximum = 0.2154G), highlighting its accuracy in processing high-resolution data.

[0068] The present invention also provides a magnetic field mapping system based on implicit expression, the system comprising a memory and a processor, the memory storing a computer program, and the processor executing the magnetic field mapping method based on implicit expression as described above when executing the computer program.

[0069] The present invention also provides a computer-readable storage medium, which stores machine-executable instructions. When the machine-executable instructions are called and executed by a processor, the machine-executable instructions prompt the processor to implement the magnetic field mapping method based on implicit expression as described above.

[0070] The present invention also provides an application of the magnetic field mapping method based on implicit expression as described above in the global positioning of a mobile robot.

[0071] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A magnetic field mapping method based on implicit expression, characterized in that: The method comprises the following steps: S1, converting low-resolution magnetic field maps into low-resolution magnetic field images; S2, encoder Mapping the obtained low-resolution magnetic field image into a 2D feature map; S3, input the obtained 2D feature map and the center coordinates of each pixel in the high-resolution magnetic field image into the decoding function f θ , decoding function f θ Predict the RGB value of each pixel position in the high-resolution image; the decoding function f is used θ To implicitly represent the continuous magnetic field image; S4, generating a high-resolution magnetic field image based on the RGB value of each pixel position in the high-resolution magnetic field image, and then obtaining a high-resolution magnetic field map based on the high-resolution magnetic field image; wherein the resolution of the high-resolution magnetic field image is greater than the resolution of the low-resolution magnetic field image, and the resolution of the high-resolution magnetic field map is greater than the resolution of the low-resolution magnetic field map.

2. The magnetic field mapping method based on implicit expression according to claim 1, characterized in that: The magnetic field map is expressed using a grid map. The size of each grid in the grid map is defined by the resolution of the magnetic field map. The resolution is represented by the side length r of each grid. The overall size of the magnetic field map is: length L x and width L y , each grid in the magnetic field map contains specific information: the grid center coordinates (x, y) and the corresponding magnetic field intensity component B x ,B y ,B z ; The magnetic field database is represented in the form of a magnetic field image. The obtained magnetic field image is constructed with a resolution of H×W pixels, where H and W are directly determined by the map size L x , L y And the resolution r is obtained, the calculation formula is: The symbol Represents the floor function.

3. The magnetic field mapping method based on implicit expression according to claim 2, characterized in that: The pixel coordinates in the magnetic field image (p x ,p y ) is mapped from the grid coordinates (x, y) in the magnetic field map, and the corresponding formula is: The magnetic field strength B is normalized using the following formula: x ,B y ,B z Normalize to fit the RGB color model range (0-255): Among them B x,min ,B x,max ,B y,min ,B y,max ,B z,min and B z,max It is the B on the entire magnetic field map x ,B y ,B z The minimum and maximum values ​​of the RGB values ​​are assigned to the corresponding pixel positions in the magnetic field image (p x ,p y ).

4. The magnetic field mapping method based on implicit expression according to claim 1, characterized in that: Denote each magnetic field image as B (i) , the pixel data in the low-resolution magnetic field image is mapped into a two-dimensional feature map All magnetic field images share a decoding function f θ , decoding function f θ For the MLP network, the formula used in the decoding process is: s=f θ (z,x); Where z is a vector, x∈X is the two-dimensional coordinate in the continuous magnetic field image domain, and s∈S is the predicted signal, i.e., the RGB value representing the magnetic field intensity.

5. The magnetic field mapping method based on implicit expression according to claim 4, characterized in that: Two-dimensional feature map M (i) The characteristic vector z is referred to as the hidden code. (i) , coordinate x q The RGB values ​​are: B (i) (x q )=f θ (z * ,x q -v * ); (1) where z * is the distance x q The latest hidden code, in M (i) Medium * It is hidden code z * Coordinates in the magnetic field image domain.

6. The magnetic field mapping method based on implicit expression according to claim 5, characterized in that: Using the decoding function f θ Representing continuous magnetic field images as two-dimensional feature maps Conceptualized as a latent array uniformly distributed in a 2D spatial region, a two-dimensional feature map M (i) Each hidden code z in * Both contain a local segment information of the continuous magnetic field image.

7. The magnetic field mapping method based on implicit expression according to claim 5, characterized in that: Formula (1) is optimized as follows: in are the nearest hidden codes in the upper left, upper right, lower left, and lower right subspaces respectively, yes The coordinates of S t is x q and The area of ​​the rectangle between t′ and t′ is the diagonal of t, and the weight is S = ∑ t S t Normalize.

8. A magnetic field mapping system based on implicit expression, characterized by: The system includes a memory and a processor, the memory stores a computer program, and the processor executes the magnetic field mapping method based on implicit expression according to any one of claims 1 to 7 when executing the computer program.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores machine-executable instructions. When the machine-executable instructions are called and executed by a processor, the machine-executable instructions prompt the processor to implement the magnetic field mapping method based on implicit expression according to any one of claims 1 to 7.

10. An application of the magnetic field mapping method based on implicit expression according to any one of claims 1 to 7 in global positioning of a mobile robot.

Citation Information

Patent Citations

  • RGB image guided depth map super-resolution method based on joint implicit image function

    CN113850719A

  • Model training and geomagnetic chart optimization method and device, medium and equipment

    CN116721316A

  • Unknown environment real-time dense map construction method based on RGB-D camera

    CN117011482A

  • Augmented reality map construction system and method, electronic equipment and storage medium

    CN118196313A

  • Image compression apparatus and method

    EP3839874A1