Methods and systems for editing digital maps

Abstract

There is provided a computer-implemented method for editing a raw digital map of a geographical area. The method may comprise providing a digital map of a geographical area; utilising a deep learning model built to perform object detection to identify one or more objects on the digital map; presenting to a user via a user interface, the digital map with labels annotating one or more identified objects, wherein the user interface comprises a user interaction interface configured to receive user input; receiving one or more inputs from the user amending the digital map with labels to generate a refined digital map with labels annotating one or more objects.

Description

METHODS AND SYSTEMS FOR EDITING DIGITAL MAPSTECHNICAL FIELD

[0001] The disclosure relates to the editing of digital maps. Specifically, the disclosure relates to a computer-implemented method for editing a raw digital map of a geographical area which incorporates a deep learning model.BACKGROUND

[0002] A map editor is a software that allows users to correct the robot-generated map. Robots generate an initial map by scanning the environment. However, the initial map often contain noisy points, dynamic objects, and distance errors. To further improve the map accuracy, users often need to identify and solve such issues in map editors manually. Furthermore, as the size of the map increases, it would take significant human effort to classify and clean up each area on the map. At the same time, the tedious manual process may cause human errors (i.e., missing certain regions, etc.).SUMMARY

[0003] It is the object of the disclosure to leverage deep learning technology to identify objects within a map, therefore potentially increasing efficiency and accuracy in map editing, thereby potentially reducing the probability of human errors and accelerating the map editing process.

[0004] The object is achieved by the subject matter of the independent claims. Preferred embodiments are subject matter of the dependent claims.

[0005] It shall be noted that all embodiments of the present disclosure concerning a method or a series of performed steps might be carried out with the order of the steps as described, nevertheless this has not to be the only and essential order of the steps of the method. The herein presented methods or series of performed steps can be carried out with another order of the disclosed steps without departing from the respective method embodiment, unless explicitly mentioned to the contrary hereinafter.

[0006] To solve the above technical problems, the present disclosure provides a computer- implemented method for editing a raw digital map of a geographical area, the method comprising: providing a digital map of a geographical area, wherein the digital map is a raw digital map comprising raw data captured by one or more sensors; utilising a deep learning model built to perform object detection, identifying one or more objects on the digital map; presenting to a user via a user interface, the digital map with labels annotating one or more identified objects, wherein the user interface comprises a user interaction interface configured to receive user input; and receiving one or more inputs from the user amending the digital map with labels to generate a refined digital map with labels annotating one or more objects.

[0007] The computer-implemented method of the present disclosure is advantageous over known methods due to the incorporation of a deep learning model built to perform object detection and annotation to identify and label one or more objects on a raw digital map, thereby potentially increasing the accuracy, efficiency, and coverage of object identification within a digital map of a geographical area. The deep learning model may be able to cover a large geographical area which may not be practical for a human to cover manually. In addition, the deep learning model may be able to identify objects more accurately, effectively, or efficiently within the digital map. Although a human may be able to identify objects in two-dimensional map data, this would be relatively time consuming when scaled to larger scales. In addition, a human may have difficulty in identifying objects in complex three-dimensional map data. Identification of objects in complex three-dimensional map data may require the human to rotate the map to view objects one by one, which may be very time consuming and prone to mistakes. On the other hand, a deep learning model is able to accurately process a large area and easily process higher dimensional map data. The method is also advantageous as it involves humans to increase the accuracy of the predictions of the deep learning model to generate a more accurate refined digital map with labels for subsequent applications. In addition, the deep learning model may be trained on data from multiple highly skilled map editors, which is advantageous as such a deep learning model may learn from the experience of multiple human map editors and may eventually surpass human map editors in most scenarios, which may allow low-skilled map editors to easilyedit maps with the assistance of deep learning models that were trained on data from highly skilled map editors.

[0008] A preferred computer-implemented method of the present disclosure is a computer- implemented method as described above or as described above as preferred, wherein the raw digital map is generated by one or more sensors, preferably one or more LIDAR sensors, mounted on a mobile robot configured to navigate and map a geographical area.

[0009] The above-described aspect of the present disclosure has the advantage that performing object detection and annotation on a digital map of a geographical area generated by one or more sensors, preferably LIDAR sensors, mounted on a mobile robot configured to navigate and map a geographical area may provide more accurate object detection over a large geographical area as compared to performing object detection and annotation on multiple camera frames which are combined to generate a map or floor plan. Although object detection may be carried out on multiple camera frames, with multiple camera frames aggregated or accumulated to generate a global map, the camera frames may require transformation with additional calculation to generate a map (or floor plan), which may potentially introduce noise.

[0010] A preferred computer-implemented method of the present disclosure is a computer- implemented method as described above or as described above as preferred, wherein the deep learning model is a convolutional neural network.

[0011] The above-described aspect of the present disclosure has the advantage that a convolutional neural network (CNN) has high accuracy in object recognition and can automatically filter images and extract or detect features in images (or higher dimension sensor data) without any human supervision.

[0012] A preferred computer-implemented method of the present disclosure is a computer- implemented method as described above or as described above as preferred, wherein the one or more inputs from the user amending the digital map with labels comprises amending one or more labels, removing one or more labels, adding one or more labels, removing one or more objects, and / or removing one or more types of objects.

[0013] The above-described aspect of the present disclosure has the advantage that receiving the different types of inputs from a user, a more accurate digital map with labels may be generated for use in subsequent applications.

[0014] A preferred computer-implemented method of the present disclosure is a computer- implemented method as described above or as described above as preferred, further comprising removing noise, duplicates, and / or dynamic objects from the digital map.

[0015] The above-described aspect of the present disclosure has the advantage that the removal of noise, duplicates, and / or dynamic objects may result in a digital map with errors and temporary objects removed, therefore generating a digital map that may be more accurate and / or representative of the physical environment for use in subsequent applications.

[0016] A preferred computer-implemented method of the present disclosure is a computer- implemented method as described above or as described above as preferred, further comprising storing the refined digital map with labels annotating one or more objects on a database and re-training the deep learning model using a plurality of refined digital map with labels annotating one or more objects stored on the database.

[0017] The above-described aspect of the present disclosure has the advantage the deep learning model may be further refined through receiving human feedback, which may potentially increase the accuracy of the deep learning model and / or number of object classes that can be detected by the deep learning model. The increase in accuracy of the deep learning model and / or object classes that can be detected by the deep learning model may lead to a reduction in human effort in future raw digital map editing for new geographical areas.

[0018] The above-described advantageous aspects of a computer-implemented method of the disclosure also hold for all aspects of a below-described digital map of the disclosure. All below-described advantageous aspects of a digital map of the disclosure also hold for all aspects of an above-described computer-implemented method of the disclosure.

[0019] The disclosure also relates to a digital map generated by a computer-implemented method according to the present disclosure.

[0020] The above-described advantageous aspects of a computer-implemented method or digital map of the disclosure also hold for all aspects of a below-described computing system of the disclosure. All below-described advantageous aspects of a computing system of the disclosure also hold for all aspects of an above-described computer-implemented method or digital map of the disclosure.

[0021] The disclosure also relates to a computing system comprising a display, an input device, one or more processors and a memory that stores executable instructions for execution by the one or more processors, the executable instructions comprising instructions for performing a computer-implemented method according to the present disclosure.

[0022] The above-described advantageous aspects of a computer-implemented method, digital map, or computing system of the disclosure also hold for all aspects of a below- described computer program, machine-readable storage medium, or data carrier signal of the disclosure. All below-described advantageous aspects of a computer program, machine- readable storage medium, or data carrier signal of the disclosure also hold for all aspects of an above-described computer-implemented method, digital map, or computing system of the disclosure.

[0023] The disclosure also relates to a computer program, a machine-readable storage medium, or a data carrier signal that comprises instructions, that upon execution on a computing system, cause the computing system to perform the steps of a computer- implemented method according to the present disclosure. The machine-readable medium may include any medium and / or mechanism for storing or transmitting information in a form readable by a machine (e.g., a computing device). The machine-readable medium may be any medium, such as for example, read-only memory (ROM); random access memory (RAM); a universal serial bus (USB) stick; a compact disc (CD); a digital video disc (DVD); a data storage device; a hard disk; electrical, acoustical, optical, or other forms of propagated signals (e.g., digital signals, data carrier signal, carrier waves), or any other medium on which a program element as described above can be transmitted and / or stored.

[0024] As used in this summary, in the description below, in the claims below, and in the accompanying drawings, the term “sensor” includes any sensor that detects, senses, measures, and / or images objects in an environment, including light sensors, lidar, radar, or sonar.

[0025] As used in this summary, in the description below, in the claims below, and in the accompanying drawings, the term “sensor data” means the output or data of any sensor.

[0026] As used in this summary, in the description below, in the claims below, and in the accompanying drawings, the term “robot” refers to any agent capable of movement, including vehicles. Examples of agents include mobile robots, vehicles, articulated robots, humanoid robots, cobots, or hybrids.BRIEF DESCRIPTION OF THE DRAWINGS

[0027] These and other features, aspects, and advantages will become better understood with regard to the following description, appended claims, and accompanying drawings where:

[0028] Fig. 1 is a schematic illustration of a method of editing a raw digital map of a geographical area, in accordance with embodiments of the present disclosure;

[0029] Fig. 2 is a flowchart of examples of visualised digital maps generated at one or more steps of a method of editing a raw digital map of a geographical area, in accordance with embodiments of the present disclosure; and

[0030] Fig. 3 is a schematic illustration of a computing system for implementing some embodiments of the present disclosure, in accordance with embodiments of the present disclosure.

[0031] In the drawings, like parts are denoted by like reference numerals.

[0032] It should be appreciated by those skilled in the art that any block diagrams herein represent conceptual views of illustrative systems embodying the principles of the present subject matter. Similarly, it will be appreciated that any flow charts, flow diagrams, state transition diagrams, pseudo code, and the like represent various processes which may be substantially represented in computer readable medium and executed by a computer or processor, whether or not such computer or processor is explicitly shown.DETAILED DESCRIPTION

[0033] In the summary above, in this description, in the claims below, and in the accompanying drawings, reference is made to particular features (including method steps)of the disclosure. It is to be understood that the disclosure of the disclosure in this specification includes all possible combinations of such particular features. For example, where a particular feature is disclosed in the context of a particular aspect or embodiment of the disclosure, or a particular claim, that feature can also be used, to the extent possible, in com-bination with and / or in the context of other particular aspects and embodiments of the disclosure, and in the disclosures generally.

[0034] In the present document, the word “exemplary” is used herein to mean “serving as an example, instance, or illustration.” Any embodiment or implementation of the present subject matter described herein as “exemplary” is not necessarily be construed as preferred or advantageous over other embodiments.

[0035] While the disclosure is susceptible to various modifications and alternative forms, specific embodiment thereof has been shown by way of example in the drawings and will be described in detail below. It should be understood, however that it is not intended to limit the disclosure to the forms disclosed, but on the contrary, the disclosure is to cover all modifications, equivalents, and alternative falling within the scope of the disclosure.

[0036] The present disclosure is directed to computer-implemented methods, computing systems, computer programs, machine-readable storage media, and data carrier signals for editing a raw digital map of a geographical area. The editing of raw digital map of a geographical area leverages deep learning to potentially increase efficiency, accuracy and speed of map labelling and / or annotation, while reducing the human effort required for map labelling and / or annotation.

[0037] Fig. 1 is a schematic illustration of a method of editing a raw digital map of a geographical area, in accordance with embodiments of the present disclosure. Method 100 for editing a raw digital map of a geographical area may be used to edit raw digital maps generated by one or more sensors. The geographical area may be any geographical area, whether indoor or outdoor, including a building, a warehouse, a factory, etc. Preferably, the geographical area may be an indoor area.

[0038] According to some embodiments, method 100 may comprise step S101 wherein a digital map of a geographical area is provided, wherein digital map is a raw digital map comprising raw data captured by one or more sensors. The one or more sensors may be anysensor, including LIDAR, sonar, camera or radar. In some embodiments, the raw digital map data may be stored on and retrieved from a database. In some embodiments, the raw digital map of a geographical area may be generated by one or more LIDAR sensors mounted on a mobile robot configured to navigate and map a geographical area. In some embodiments, the raw digital map may be generated using known methods for robot mapping of physical areas, such as Google’s Cartographer disclosed in “Real-Time Loop Closure in 2D LIDAR SLAM” by Hess et. al., and SLAM Toolbox disclosed in “SLAM Toolbox: SLAM for the dynamic world” by Macenski and Jambrecic. In some embodiments, the raw digital map may be derived from a variety of sources, with each source providing different amounts and types of information, wherein each individual source of raw digital map data compiled and stored in a database.

[0039] According to some embodiments, the raw digital map may be an occupancy map or occupancy grid map, which is a map that represents an environment as a discrete grid and each cell of the grid is marked with a likelihood that the cell contains an object. The occupancy map may be two or more dimensional maps, and preferably a 2-dimensional (2D) occupancy grid map. In some embodiments, the raw digital map may be Octomap or Mapnet. In some embodiments, the raw digital map may be open VDB or neural VDB which allow optimisation while reducing CPU and memory usage.

[0040] In some embodiments, the raw digital map may be visualised on a display to a user. In some embodiments, the raw digital map may be displayed with a user interface or a graphical user interface to a user. Visualisation of the raw digital map may be done using any known methods, including using known software such as Matplotlib for Python available at https: / / matplotlib.org / or D3.js for web-based applications available at https: / / d3js.org / .

[0041] According to some embodiments, method 100 may comprise step SI 04 wherein a deep learning model built to perform object detection and annotation is utilised to identify one or more objects on the raw digital map. Object detection or identification may comprise identifying the identity and / or class of an object.

[0042] Any known deep learning model built for object detection and annotation may be used. Preferably, the deep learning model may be a convolutional neural network (CNN). A convolutional neural network (CNN) is a multi-layered feed-forward neural network, madeby stacking many hidden layers on top of each other in sequence. The sequential design may allow convolutional neural networks to learn hierarchical features. The hidden layers are typically convolutional layers followed by activation layers, some of them followed by pooling layers. The CNN may be configured to identify pattens in data. The convolutional layer may include convolutional kernels, that are used to look for patterns across the input data. The convolutional kernel may return a large positive value for a portion of the input data that matches the kernel’s pattern or may return a smaller value for another portion of the input data that does not match the kernel’s pattern. The architecture of the deep learning model may vary based on the type of raw digital map to be edited.

[0043] In some embodiments, where the raw digital map is a two-dimensional map, the deep learning model may be Mask R-CNN, UNet or DeepLab. Mask R-CNN is disclosed in “Mask R-CNN” by He et. al., wherein an example of the architecture may be found at least in Fig. 1 and Section 3 and an example of the training may be found at least in Section 3. UNet is disclosed in “U-Net: Convolutional Networks for Biomedical Image Segmentation” by Ronneberger et. al., wherein an example of the architecture may be found at least in Fig. 1 and Section 2, and wherein an example of the training may be found at least in Section 3. DeepLab is disclosed in “DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs” by Chen et. al., wherein an example of the architecture may be found at least in Figs. 2 to 4 and Section 3, and an example of the training may be found at least in Section 4.

[0044] In some embodiments, where the raw digital map is a three-dimensional map, the deep learning model may be 3D Fully Convolutional Networks, 3D U-Net, or Octree-based Convolutional Neural Networks. Examples of deep learning models for 3D input include PointNet, Minowski Convolutional Neural Network, PointCNN, and Yolov8. PointNet is disclosed in “PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation” by Qi et. al., wherein an example of the architecture may be found at least in Fig. 2 and Section 4 of the main paper, and Figure 9 and 10, and Section C and D of the Supplementary Material, and an example of the training may be found at least in Section C of the Supplementary Material. Minkowski Convolutional Neural Network is disclosed in “4D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networks” by Choy et. al., wherein an example of the architecture may be found at least in in Figs. 4 and 5, and Sections 5 and 6 of the main paper, as well as the Sections 1 and 2 of the SupplementaryMaterial, and an example of the training may be found at least in Section 7. PointCNN is disclosed in “PointCNN: Convolution On % -Transformed Points”, wherein an example of the architecture may be found at least in Fig. 4 and Section 3 of the main paper, and Fig. 1 and Section 1 of the Supplementary Material, and an example of the training may be found in Section 2 of the Supplementary Material.

[0045] According to some embodiments, method 100 may comprise step SI 08, wherein the digital map with labels annotating one or more identified objects is presented to a user via a user interface, wherein the user interface comprises user interaction interface configured to receive user input. Examples of labels may include object labels that identify the object, and other labels such as object classifications (e.g., static / dynamic objects, noise, duplicate, etc). The user may provide input through the user interaction interface using one or more input devices, such as a keyboard, mouse, touchscreen, or joystick. In some embodiments, the user may manipulate with the visualized digital map with labels, including rotating, moving, zooming in and / or zooming out.

[0046] According to some embodiments, method 100 may comprise step SI 12, wherein one or more inputs are received from the user amending the digital map with labels to generate a refined digital map with labels annotating one or more objects. The user may provide one or more inputs amending the digital map with labels, such as amending one or more labels, adding one or more labels, removing one or more labels, and removing one or more objects, and / or removing one or more types of objects. For example, the user may amend or correct an incorrectly labelled object. For example, the user may add a label for an object that was not identified or labelled by the deep learning model. For example, the user may remove a label from an object that the deep learning model has mistakenly labelled twice. For example, the user may amend a class of the label of an object, labelling an object as static, dynamic, or noise. For example, the user may remove types of objects from the digital map with labels, such as removing all objects identified to be noise or dynamic objects. In some embodiments, where the user removes one or more objects and / or types of objects from the digital map, a pop-up notification may be presented to the user to request for a reason that the user is removing the object and / or type of object.

[0047] According to some embodiments, method 100 may comprise step SI 16 wherein noise, duplicates, and / or dynamic objects are removed from the digital map. In someembodiments, noise, duplicates and / or dynamic objects may be removed based on the labels and / or input received from the user.

[0048] According to some embodiments, method 100 may comprise step S120 wherein the refined digital map with labels annotating one or more objects is stored in a database, and the deep learning model is re-trained using a plurality of refined digital map with labels annotating one or more objects stored on the database.

[0049] Fig. 2 is a flowchart of examples of visualised digital maps generated at one or more steps of method 100 of editing a raw digital map of a geographical area, in accordance with embodiments of the present disclosure. It is emphasised that although Fig. 2 illustrates the maps as two-dimensional maps, method 100 may also be implemented with maps of higher dimension (e.g., three-dimensional maps). In some embodiments, a raw digital map 244 of a geographical area is provided in step S 101. The raw digital map 244 may be visualised to a user on display and / or graphical user interface. As illustrated in Fig. 2, certain regions of the raw digital map may be visualised as black to indicate a high probability that the region contains an object.

[0050] According to some embodiments, a deep learning model 248 built to perform object detection and annotation is utilised to identify one or more objects on the raw digital map in step SI 04. In some embodiments, a digital map with labels annotating one or more objects 252 may be presented to the user via a user interface in step S108. As illustrated in Fig. 2, the objects in the digital map were identified and labelled as “box”, “rack”, “table”, and “noise”.

[0051] According to some embodiments, the user may amend the digital map with labels 252 in step SI 12 to generate a refined digital map 256 with labels annotating one or more objects. As illustrated in Fig. 2, the user has labelled one of the objects previously labelled as “box” as a “duplicated measurement” and has corrected the label of “table” to “forklift”. In some embodiments, noise, duplicates and / or dynamic objects (e.g., forklift) may be removed from the digital map in step 116, as illustrated in the generation of digital map 260 wherein the objects labelled as “duplicated measurement”, “noise”, and “forklift” were removed.

[0052] According to some embodiments, the deep learning model 248 built to perform object detection and annotation may be an untrained deep learning model that is trained and refined through multiple iterations of method 100, wherein the refined digital maps with labels annotating one or more objects based on the received one or more inputs from the user generated in step SI 16 are saved onto a database in step SI 20 to generate a training dataset to train deep learning model 248 through interactive deep learning. In some embodiments, the deep learning model built to perform object detection and annotation may be trained on a training dataset comprising between 100 to 1000 digital maps with labels annotating 3 to 20 classes of objects with primary object categories including noise, dynamic objects, and static objects, wherein the refined digital maps with labels annotating one or more objects based on the received one or more inputs from the user generated in step SI 16 are used to re-train and further refine deep learning model 248 in step S120 to increase the accuracy of the deep learning model 248.

[0053] Fig. 3 is a schematic illustration of a computing system for implementing some embodiments of the present disclosure, in accordance with embodiments of the present disclosure. Computing system 300 can be used, for example, for one or more steps of method 100. System 300 can be a computer connected to a network. System 300 can be a client or a server. As shown in Fig. 3, system 300 can be any suitable type of processor-based system, such as a personal computer, workstation, server, handheld computing device (portable electronic device) such as a phone or tablet, or an embedded system or other dedicated device. The system 300 can include, for example, one or more processors 310, one or more memory 314, one or more of input device 320, one or more of output device 330, a graphical user interface (GUI) 334, storage 340, and communication device 360. Input device 320 and / or output device 330 can generally either be connectable or integrated with the computing system 300. In some embodiments, storage 340 may comprise a database storing one or more digital maps 244, 252, 256, and / or 260. In some embodiments, storage 340 may store the deep learning model 248.

[0054] Input device 320 can be any suitable device that provides input, such as a touch screen, keyboard or keypad, mouse, gesture recognition component of a virtual / augmented reality system, or voice-recognition device. Output device 330 can be or include any suitable device that provides output, such as a display, touch screen, haptics device, virtual / augmented reality display, or speaker.

[0055] Storage 340 can be any suitable device that provides storage, such as an electrical, magnetic, or optical memory including a RAM, cache, hard drive, removable storage disk, or other non-transitory computer readable medium. Communication device 360 can include any suitable device capable of transmitting and receiving signals over a network, such as a network interface chip or device. The components of the computing system 300 can be connected in any suitable manner, such as via a physical bus or wirelessly.

[0056] Processor(s) 310 can be any suitable processor or combination of processors, including any of, or any combination of, a central processing unit (CPU), graphics processing unit (GPU), field programmable gate array (FPGA), and application-specific integrated circuit (ASIC). Software 350, which can be stored in storage 340 and executed by one or more processors 310, can include, for example, the programming that embodies the functionality or portions of the functionality of the present disclosure (e.g., as embodied in the devices as described above). For example, software 350 can include one or more programs for execution by one or more processor(s) 310 for performing one or more of the steps of method 100.

[0057] Software 350 can also be stored and / or transported within any non-transitory computer-readable storage medium for use by or in connection with an instruction execution system, apparatus, or device, such as those described above, that can fetch instructions associated with the software from the instruction execution system, apparatus, or device and execute the instructions. In the context of this disclosure, a computer-readable storage medium can be any medium, such as storage 340, that can contain or store programming for use by or in connection with an instruction execution system, apparatus, or device.

[0058] Software 350 can also be propagated within any transport medium for use by or in connection with an instruction execution system, apparatus, or device, such as those described above, that can fetch instructions associated with the software from the instruction execution system, apparatus, or device and execute the instructions. In the context of this disclosure, a transport medium can be any medium that can communicate, propagate or transport programming for use by or in connection with an instruction execution system, apparatus, or device. The transport computer readable medium can include, but is not limited to, an electronic, magnetic, optical, electromagnetic, or infrared wired or wireless propagation medium.

[0059] System 300 may be connected to a network, which can be any suitable type of interconnected communication system. The network can implement any suitable communications protocol and can be secured by any suitable security protocol. The network can comprise network links of any suitable arrangement that can implement the transmission and reception of network signals, such as wireless network connections, T1 or T3 lines, cable networks, DSL, or telephone lines.

[0060] System 300 can implement any operating system suitable for operating on the network. Software 350 can be written in any suitable programming language, such as C, C++, Java, or Python. In various embodiments, application software embodying the functionality of the present disclosure can be deployed in different configurations, such as in a client / server arrangement or through a Web browser as a Web-based application or Web service, for example.

[0061] Finally, the language used in the specification has been principally selected for readability and instructional purposes, and it may not have been selected to delineate or circumscribe the inventive subject matter. It is therefore intended that the scope of the disclosure be limited not by this detailed description, but rather by any claims that issue on an application based here on. Accordingly, the embodiments of the present disclosure are intended to be illustrative, but not limiting, of the scope of the disclosure, which is set forth in the following claims.

Claims

CLAIMS1. A computer-implemented method for editing a raw digital map of a geographical area, the method comprising: providing a digital map of a geographical area, wherein the digital map is a raw digital map comprising raw data captured by one or more sensors; utilising a deep learning model built to perform object detection, identifying one or more objects on the digital map; presenting to a user via a user interface, the digital map with labels annotating one or more identified objects, wherein the user interface comprises a user interaction interface configured to receive user input; and receiving one or more inputs from the user amending the digital map with labels to generate a refined digital map with labels annotating one or more objects.

2. The computer-implemented method of claim 1, wherein the raw digital map is generated by one or more sensors, preferably one or more LIDAR sensors, mounted on a mobile robot configured to navigate and map a geographical area.

3. The computer-implemented method of any of the preceding claims, wherein the deep learning model is a convolutional neural network.

4. The computer-implemented method of any of the preceding claims, wherein the one or more inputs from the user amending the digital map with labels comprises amending one or more labels, removing one or more labels, adding one or more labels, removing one or more objects, and / or removing one or more types of objects.

5. The computer-implemented method of any of the preceding claims, further comprising removing noise, duplicates, and / or dynamic objects from the digital map.

6. The computer-implemented method of any of the preceding claims, further comprising storing the refined digital map with labels annotating one or more objects on a database and re-training the deep learning model using a plurality of refined digital map with labels annotating one or more objects stored on the database.

7. A digital map generated by a computer-implemented method according to any of the preceding claims.

8. A computing system comprising a display, an input device, one or more processors and a memory that stores executable instructions for execution by the one or more processors, the executable instructions comprising instructions for performing a computer- implemented method according to any of the preceding claims 1 to 6.

9. A computer program, a machine-readable storage medium, or a data carrier signal that comprises instructions, that upon execution on a computing system, cause the computing system to perform the steps of a computer-implemented method according to any of the preceding claims 1 to 6.

Images