CN114648598A

CN114648598A - Global map construction method and device

Info

Publication number: CN114648598A
Application number: CN202210190958.6A
Authority: CN
Inventors: 王发平; 李夏威; 姜波
Original assignee: Shenzhen Haixing Zhijia Technology Co Ltd
Current assignee: Shenzhen Haixing Zhijia Technology Co Ltd
Priority date: 2022-02-28
Filing date: 2022-02-28
Publication date: 2022-06-21

Abstract

The invention provides a global map construction method and a device, wherein the method comprises the following steps: acquiring a first local map constructed by a first target object and a collected first environment image, and extracting a first visual feature from the first environment image; receiving a second local map and second visual features sent by other target objects; and respectively fusing the first local map and the second local maps corresponding to the rest of target objects based on the relationship between the first visual feature and the second visual features corresponding to the rest of target objects to obtain the global map corresponding to the first target object. Therefore, the cooperative construction of the global map is realized through the fusion of the multi-target object local map and the local map of the first target object, the construction cost of the global map is reduced while the map construction efficiency is improved, and the accuracy of the final global map is further ensured by utilizing the visual characteristics in the original environment image acquired by each target object.

Description

Global map construction method and device

Technical Field

The invention relates to the technical field of artificial intelligence, in particular to a global map construction method and a global map construction device.

Background

With the development of engineering machinery industry in recent years, the requirements for intellectualization, unmanned operation and digitization are more and more urgent, and the intellectualization and the unmanned operation firstly need to comprehensively sense and model the operation environment of the engineering machinery, so that the following tasks of automatic driving, operation path planning and the like of engineering vehicles are facilitated. The working and driving environment of the engineering machinery is often in a larger global range and has coherent stability in a geographic area, and the characteristics of the working and driving environment provide challenges for environment modeling of a whole scene.

At present, environment modeling is focused on research and use of single vehicle mapping, a single vehicle is additionally provided with a plurality of environment sensing devices to model and position the surrounding environment in the continuous driving process of the vehicle, a relatively mature SLAM (simultaneous localization and mapping) system based on laser radar is successfully applied to various scenes, and an unmanned aerial vehicle or an indoor scene robot can perform autonomous mapping and path planning through a visual SLAM system. For the field of engineering machinery, errors can be continuously accumulated along with the running of vehicles in the single-vehicle environment modeling, in a large scene of the engineering machinery, the single-vehicle environment modeling cannot well obtain a stable diagram building effect, and the stability of the system can be tested. The cooperative mapping method based on the laser radar is only suitable for multi-vehicle direct cooperative mapping with environment data collected by various vehicles being laser point clouds, the laser radar with high cost is not suitable for being carried on each engineering vehicle, and the investment cost of the engineering vehicle can be greatly increased when the laser radar is applied to an engineering operation scene.

Disclosure of Invention

In view of this, embodiments of the present invention provide a global map construction method and apparatus, so as to overcome a problem that it is difficult to consider both the map construction accuracy and the economy in a panoramic map construction manner in the prior art.

According to a first aspect, an embodiment of the present invention provides a global map construction method, which is applied to any one target object in a collaborative map construction system, where the collaborative map construction system includes a plurality of target objects, and the method includes:

acquiring a first local map constructed by a first target object and a collected first environment image, and extracting a first visual feature from the first environment image;

receiving a second local map and second visual features sent by other target objects, wherein the second visual features are visual features extracted from a second environment image acquired by the current target object;

and respectively fusing the first local map and second local maps corresponding to the rest of target objects based on the relationship between the first visual feature and the second visual features corresponding to the rest of target objects to obtain a global map corresponding to the first target object.

Optionally, the fusing the first local map and the second local maps corresponding to the remaining target objects respectively based on the relationship between the first visual feature and the second visual features corresponding to the remaining target objects includes:

performing feature point matching on the first visual feature and a second visual feature corresponding to the current target object;

processing the first local map and a second local map corresponding to the current target object based on the feature point matching result to obtain a current global map of the first target object and the current target object;

and obtaining a global map corresponding to the first target object based on the current global maps of the first target object and all the other target objects.

Optionally, the processing the first local map and the second local map based on the feature point matching result to obtain a current global map of the first target object and the current target object includes:

judging whether the feature point matching result meets the preset feature point matching requirement or not;

and when the feature point matching result meets the preset feature point matching requirement, performing map fusion of overlapping positions on the first local map and the second local map to obtain a global map corresponding to the current target object.

Optionally, the method further comprises:

and when the feature point matching result does not meet the preset feature point matching requirement, deleting the second visual feature and the second local map corresponding to the current target object, and continuing to perform feature point matching on the first visual feature and the second visual feature corresponding to the next target object.

Optionally, before performing feature point matching on the first visual feature and a second visual feature corresponding to the current target object, the method further includes:

acquiring a first visual bag-of-words model corresponding to the first visual characteristic and a second visual bag-of-words model corresponding to a second visual characteristic corresponding to the current target object;

calculating a first distance between the second visual bag-of-words model and the first visual bag-of-words model;

judging whether the first distance is smaller than a distance threshold value;

and when the first distance is smaller than a distance threshold value, performing feature point matching on the first visual feature and a second visual feature corresponding to the current target object.

Optionally, the method further comprises:

and when the first distance is not less than a distance threshold value, map splicing of different positions is carried out on the first local map and a second local map corresponding to the current target object, so that a current global map of the first target object and the current target object is obtained.

Optionally, the extracting the first visual feature from the first environment image includes:

a first visual feature is extracted from the first environmental image using a full convolution network.

According to a second aspect, an embodiment of the present invention provides a global map building apparatus, which is applied to any one target object in a collaborative mapping system, where the collaborative mapping system includes a plurality of target objects, and the apparatus includes:

the acquisition module is used for acquiring a first local map constructed by a first target object and a collected first environment image and extracting a first visual feature from the first environment image;

the first processing module is used for receiving a second local map and second visual features sent by other target objects, wherein the second visual features are visual features extracted from a second environment image acquired by a current target object;

and the second processing module is used for fusing the first local map and second local maps corresponding to the rest of target objects respectively based on the relationship between the first visual feature and the second visual features corresponding to the rest of target objects to obtain a global map corresponding to the first target object.

According to a third aspect, embodiments of the present invention provide a computer-readable storage medium storing computer instructions which, when executed by a processor, implement the method of the first aspect of the present invention and any one of its alternatives.

According to a fourth aspect, an embodiment of the present invention provides an electronic device, including:

a memory and a processor, the memory and the processor being communicatively coupled to each other, the memory having stored therein computer instructions, the processor being configured to execute the computer instructions to perform the method of the first aspect of the present invention and any one of the alternatives thereof.

The technical scheme of the invention has the following advantages:

the embodiment of the invention provides a global map construction method and a global map construction device, which are applied to any one target object in a collaborative map construction system, wherein the collaborative map construction system comprises a plurality of target objects, a first local map constructed by a first target object and a collected first environment image are obtained, and a first visual feature is extracted from the first environment image; receiving a second local map and second visual features sent by other target objects; and respectively fusing the first local map and the second local maps corresponding to the rest of target objects based on the relationship between the first visual feature and the second visual features corresponding to the rest of target objects to obtain the global map corresponding to the first target object. Therefore, the position relation between the local maps constructed by the other target objects and the local map of the first target object is determined by utilizing the relation between the visual features in the environment images collected by the target objects, the cooperative construction of the global map of the first target object is realized by fusing the local maps of the multiple target objects and the local map of the first target object, the construction efficiency of the global map is improved, meanwhile, the laser radar does not need to be loaded, the construction cost of the global map is reduced, and the accuracy of the final global map is further ensured by utilizing the visual features in the original environment images collected by the target objects.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.

FIG. 1 is a flow chart of a global map construction method in an embodiment of the present invention;

FIG. 2 is a schematic structural diagram of a global map building apparatus according to an embodiment of the present invention;

fig. 3 is a schematic structural diagram of an electronic device in an embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings, and are only for convenience of description and simplicity of description, but do not indicate or imply that the device or element being referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus, should not be construed as limiting the present invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.

In the description of the present invention, it should be noted that, unless otherwise explicitly specified or limited, the terms "mounted," "connected," and "connected" are to be construed broadly, e.g., as meaning either a fixed connection, a removable connection, or an integral connection; can be mechanically or electrically connected; the two elements may be directly connected or indirectly connected through an intermediate medium, or may be communicated with each other inside the two elements, or may be wirelessly connected or wired connected. The specific meanings of the above terms in the present invention can be understood by those skilled in the art according to specific situations.

The technical features mentioned in the different embodiments of the invention described below can be combined with each other as long as they do not conflict with each other.

At present, environment modeling is focused on research and use of single vehicle mapping, a single vehicle is additionally provided with a plurality of environment sensing devices to model and position the surrounding environment in the continuous driving process of the vehicle, the current mature SLAM system based on the laser radar is successfully applied to various scenes, and an unmanned aerial vehicle or an indoor scene robot can automatically map and plan a path through a visual SLAM system. For the field of engineering machinery, errors can be continuously accumulated along with the running of a vehicle in single-vehicle environment modeling, stable diagram building effects cannot be well obtained in the single-vehicle environment modeling under a large engineering machinery scene, and the stability of a system can be tested; and the SLAM system based on the laser radar is usually high in cost and poor in economical efficiency.

Based on the above problems, embodiments of the present invention provide a global map construction method, which is applied to any one target object in a collaborative map construction system, where the collaborative map construction system includes a plurality of target objects, and in the embodiments of the present invention, the collaborative map construction system is formed by using a plurality of engineering vehicles in the same large scene as the target objects, and it can be understood that the engineering vehicles referred to herein include not only the engineering vehicles in the traditional sense, such as a forklift, a dump truck, a flatbed, a tractor, etc., but also the engineering mechanical devices, such as an excavator, a paver, a mixer, a grader, etc. Each vehicle serves as one of the nodes, and a complete single-node system runs inside the nodes: the system comprises an independent VSLAM module for modeling the peripheral local environment of a single vehicle, a communication link for communication between nodes, a module for establishing a map by cooperation of adjacent nodes and the like. The vehicle is provided with a binocular or depth vision sensor for acquiring image information of the surrounding environment of the vehicle and the like. In practical application, the cooperative mapping system may also be formed by other vehicles or intelligent devices that have a requirement for cooperative mapping, such as an intelligent device arranged at a field end, such as an edge server, and the present invention is not limited thereto.

As shown in fig. 1, the global map construction method is applied to a controller of any target object in a collaborative map construction system, and specifically includes the following steps:

step S101: the method comprises the steps of obtaining a first local map constructed by a first target object and a collected first environment image, and extracting a first visual feature from the first environment image.

The first target object is an engineering vehicle with a global map building requirement, the first environment image is a key frame of an environment image acquired by a vision sensor carried by the engineering vehicle, and the number of the key frames can be flexibly set according to modeling precision. In addition, in practical application, the construction mode of each local map of the engineering vehicle can be obtained by adopting other map construction methods based on a single vehicle in the prior art, for example, if a certain engineering vehicle is equipped with laser radar equipment, the local map can also be constructed based on laser point cloud data collected by the laser radar equipment.

Specifically, a SLAM that uses only a camera as an external perception sensor is called a visual SLAM or vSLAM. The camera has the advantages of rich visual information, low hardware cost and the like, and a classical vSLAM system generally comprises four main parts, namely a front-end visual odometer, rear-end optimization, closed-loop detection and composition. The main implementation process is Visual Odometry (Visual Odometry): pose estimation with visual input only; back-end Optimization (Optimization): the rear end receives the camera poses measured by the visual odometer at different moments and the closed-loop detection information, and the camera poses and the closed-loop detection information are optimized to obtain globally consistent tracks and maps; closed loop assay (loopcloning): in the map building process, the robot detects whether track closed loop occurs through sensor information such as vision and the like, namely whether the robot enters the same historical place is judged; mapping (Mapping): and establishing a map corresponding to the task requirement according to the estimated track. For a detailed process of how to utilize the VSLAM module to construct the local map, reference may be made to related descriptions in the prior art, and details are not described herein again.

Specifically, because the conventional ORB (organized Fast and Rotated brief) feature extraction method cannot well extract deep semantic features of vision, when loop detection, that is, the closed-loop detection is performed, image features based on ORB cannot be well matched, and especially in a multi-vehicle collaborative environment modeling scene, environmental factors such as viewing angles and distances of different vehicles to the same place are different, which causes a large error of loop detection based on ORB feature matching, thereby affecting the accuracy of a final global map. In an embodiment of the invention, a full convolution network is used to extract a first visual feature from a first environmental image. By using the full convolution network, the image feature extraction process is adaptive to image rotation invariance, translation invariance and scale invariance, deep semantic features of the image are better extracted, unnecessary ineffective extraction of low-level image features and interference on subsequent matching calculation are reduced, and the accuracy of finally constructing the global map is further improved.

Step S102: and receiving the second local map and the second visual characteristics sent by the rest target objects.

And the other target objects are other engineering vehicles in the collaborative mapping system, and the second visual feature is a visual feature extracted by the current target object from a second environment image acquired by a visual sensor of the current target object.

Specifically, each engineering vehicle in the collaborative mapping system carries out local mapping and positioning by operating the VSLAM module and utilizing the key frame of the acquired environment image to obtain a local map, and extracting corresponding visual features from the acquired environment image by using the visual feature extraction mode in the process, and sending the local map and the visual features to other engineering vehicles through a communication link, so that each engineering vehicle can obtain local mapping and visual characteristics of other engineering vehicles, the global map modeling of any engineering vehicle can be realized through the construction of the local map by the engineering vehicle and the extraction of the depth image in the environment image key frame, and the distributed data processing mode greatly reduces the calculation processing amount, reduces the unnecessary consumption of calculation resources and further improves the efficiency of global map construction.

Step S103: and respectively fusing the first local map and the second local maps corresponding to the rest of target objects based on the relationship between the first visual feature and the second visual features corresponding to the rest of target objects to obtain the global map corresponding to the first target object.

Specifically, the visual features are extracted from the key frames of the environment images acquired by the engineering vehicles, the relationship between the visual features corresponding to different engineering vehicles reflects the position relationship between different engineering vehicles, and the local maps corresponding to the engineering vehicles can be fused based on the position relationship to obtain a global map with a certain engineering vehicle as the center.

By executing the steps, the global map construction method provided by the embodiment of the invention determines the position relationship between the local maps constructed by the other target objects and the local map of the first target object by using the relationship between the visual features in the environment images acquired by the target objects, realizes the cooperative construction of the global map of the first target object by fusing the local maps of the multiple target objects and the local map of the first target object, improves the map construction efficiency, does not need to load a laser radar, reduces the construction cost of the global map, and further ensures the accuracy of the final global map by using the visual features in the original environment images acquired by the target objects.

Specifically, in an embodiment, the step S103 specifically includes the following steps:

step S201: and performing feature point matching on the first visual features and the second visual features corresponding to the current target object.

Step S202: and processing the first local map and a second local map corresponding to the current target object based on the feature point matching result to obtain a current global map of the first target object and the current target object.

Specifically, in an embodiment, the step S202 specifically includes the following steps:

step S301: and judging whether the feature point matching result meets the preset feature point matching requirement or not.

Specifically, the preset feature point matching requirement refers to the number of feature point matches, and the specific number requirement may be flexibly set according to the global map construction accuracy requirement and the speed requirement, which is not limited in the present invention.

Step S302: and when the feature point matching result meets the preset feature point matching requirement, performing map fusion of the overlapping position on the first local map and the second local map to obtain a global map corresponding to the current target object.

Specifically, the overlapped coordinates of the feature points in the world coordinate system are obtained by obtaining the coordinates of the successfully matched feature points in the corresponding environment images, utilizing the conversion relation between the visual sensor coordinate system on the corresponding engineering vehicle and the world coordinate system, and then the overlapped coordinates of the two successfully matched feature points in the world coordinate system are utilized to perform map fusion of the overlapped positions of the two local maps.

Step S303: and when the feature point matching result does not meet the preset feature point matching requirement, deleting the second visual feature and the second local map corresponding to the current target object, and continuing to perform feature point matching on the first visual feature and the second visual feature corresponding to the next target object.

Specifically, as the local maps of the two engineering vehicles have the overlapping area of a certain area only when the matching number of the feature points reaches a certain number value, if the matching number of the feature points is not met, the overlapping area is too small, the number of the matched feature points is small, fusion errors easily occur, and the fusion accuracy of the local maps is difficult to guarantee.

Step S203: and obtaining a global map corresponding to the first target object based on the current global maps of the first target object and all the other target objects.

Specifically, all the current global maps are established on the basis of the local map of the first target object, that is, all the current global maps include the local map of the first target object, and the global map corresponding to the first target object can be obtained by directly merging all the current global maps. Similarly, the global map of any target object, namely any engineering vehicle position in the collaborative map building system can be obtained by adopting the process.

Specifically, in an embodiment, before executing the step S201, the global map building method provided in the embodiment of the present invention further includes the following steps:

step S401: and acquiring a first visual bag-of-words model corresponding to the first visual feature and a second visual bag-of-words model corresponding to the second visual feature corresponding to the current target object.

Specifically, the visual features of the key frames of the environment image are represented by adopting a visual bag-of-words model. In practical application, the visual bag-of-words model of each target object may be obtained by processing each target object, and then sent to the target object undergoing global map modeling through a communication link, or the target object undergoing global map modeling may obtain a corresponding visual bag-of-words model by using the received visual features, which is not limited in the present invention.

Step S402: and calculating a first distance between the second visual bag-of-words model and the first visual bag-of-words model.

Specifically, the first distance is obtained by calculating the euclidean distance between two visual bag-of-words models.

Step S403: and judging whether the first distance is smaller than a distance threshold value.

The distance threshold may be dynamically configured, the distance threshold is normalized, the matching severity may be dynamically adjusted by adjusting the size of the distance threshold in the test process to obtain the best mapping effect, and the initial threshold may be set to be between 0.3 and 0.5.

Specifically, when the first distance is smaller than the distance threshold, the above step S201 is performed. And when the first distance is not less than the distance threshold, performing map splicing of different positions on the first local map and a second local map corresponding to the current target object to obtain a current global map of the first target object and the current target object.

In practical application, if the distance between two visual bag-of-words models is larger, the possibility that the local maps corresponding to the two visual bag-of-words models are overlapped is smaller, more computing resources are consumed due to continuous matching calculation of the feature points, the two local maps which are not overlapped at all can be directly spliced according to respective positions by setting a distance threshold, and feature point matching is not needed. Therefore, the visual bag-of-words model is adopted to pre-calculate the feature matching of different target objects, so that unnecessary specific matching calculation is reduced, unnecessary consumption of calculation resources is reduced, and the global map modeling efficiency is further improved.

The global map construction method provided by the embodiment of the present invention will be described in detail below with reference to specific application examples.

And representing the controller of each engineering vehicle in the collaborative mapping system by using an Agent.

Each Agent sending end operates a VSLAM system by accessing a visual sensor, internal storage consumption is reduced by limiting the number of key frames in the VSLAM system, a visual odometer is operated after image input is obtained, feature extraction is carried out on the key frames, a full convolution network is adopted to replace ORB feature extraction in the feature extraction process, meanwhile, a visual bag-of-words model is adopted to represent the features of the key frames, and the visual features and bag-of-words vectors of the key frames are cached locally.

And the Agent performs mapping and positioning of a local map based on the single-vehicle VSLAM system at a transmitting end, stores the local map and the global map after coordinate conversion, and has the same data before effective communication with other agents.

The Agent establishes communication with an adjacent Agent through a communication module, and sends key frame word bag vectors to other adjacent agents through a communication link so as to be used for pre-calculating the visual features of the key frame images and simultaneously sends corresponding key frame visual feature point data and local map information.

And the target Agent receives key frame word bag vectors of other agents at a receiving end through a communication link, performs key frame visual feature pre-matching calculation by calculating Euclidean distance with the locally cached key frame word bag vectors, acquires whether a set distance threshold result is met, performs further feature point matching calculation on key frame visual feature data sent by other agents and key frame visual feature data cached by the node if the set threshold is met, and otherwise, directly performs map splicing of the global map at different positions.

And the target Agent performs feature point matching at a receiving end, destroys the visual features of the key frames of other agents cached in the previous step and the corresponding local map if the set threshold of the matching number of the feature points is not met, and triggers closed-loop detection to perform map fusion of the global map at the overlapping position if the set threshold is met, wherein the closed-loop detection is the prior art, and the specific implementation process is not repeated herein.

Each Agent dynamically maintains and updates a local global map, and when the retrace of a single Agent repeated place and the fusion of the same place of a multi-Agent map are carried out, a closed loop detection mechanism is triggered.

How to utilize the technical scheme of the invention to carry out collaborative environment modeling in a large scene by a plurality of engineering machines is explained by a specific embodiment. In the unmanned forklift goods transportation scene, a plurality of unmanned forklifts need to load goods at the point A in a relatively fixed park, unload the goods at the point B through a section of intermediate driving stage, and then continue to return to the point A to circulate the tasks of the steps; each forklift is an independent Agent, the image information of the surrounding environment is acquired by additionally arranging a visual sensor, the vehicle-mounted controller provides a computing unit and a storage unit, a 5G module of the controller provides a communication function, and a collaborative mapping system runs in each forklift and comprises a sending end and a receiving end; at a sending end, independently operating the VSLAM system based on the full convolution network in the technical scheme, locally modeling a local map and a global map, continuously modeling the surrounding environment of the forklift, and continuously updating the local map and the global map; at a receiving end, synchronously operating environment collaborative modeling of different forklift nodes, establishing connection through a communication module when other nearby forklift nodes exist in the forklift, performing mutual communication of the different forklift nodes, performing key frame feature matching precalculation on each forklift node through visual word bag vectors which are communicated with each other, and then judging whether further image visual feature matching calculation is needed or not according to precalculation results; and if not, directly splicing maps of the multi-forklift global map at different position points on the receiving end of each forklift. If the local map is required to be matched with the multi-forklift global map, carrying out specific visual feature matching calculation by using the received image visual feature data, and if the set threshold is met according to the set threshold, carrying out map fusion of the multi-forklift global map at the same position point, and if the set threshold is not met, deleting the received image visual feature and the local map; through the collaborative environment modeling of different forklifts, a single forklift can obtain map information of an area which is not passed by, and meanwhile, the long-time long-distance accumulated error of a single forklift is reduced in a multi-forklift collaborative mode. Through the visual VSLAM system and multi-target object collaborative environment modeling, the configuration cost is reduced, the global map building efficiency is improved, the single-vehicle accumulated error is reduced, and the over-the-horizon global map information is quickly built in different target objects.

An embodiment of the present invention further provides a global map building apparatus, as shown in fig. 2, the global map building apparatus specifically includes:

the obtaining module 101 is configured to obtain a first local map constructed by a first target object and a collected first environment image, and extract a first visual feature from the first environment image. For details, refer to the detailed description of step S101 above, and are not repeated herein.

The first processing module 102 is configured to receive the second local map and the second visual features sent by the other target objects, where the second visual features are visual features extracted by the current target object from the second environment image acquired by the current target object. For details, refer to the detailed description of step S102, which is not repeated herein.

The second processing module 103 is configured to fuse the first local map and the second local maps corresponding to the remaining target objects, respectively, based on a relationship between the first visual feature and the second visual feature corresponding to each of the remaining target objects, so as to obtain a global map corresponding to the first target object. For details, refer to the detailed description of step S103, which is not repeated herein.

Through the cooperative cooperation of the above components, the global map construction device provided in the embodiment of the present invention determines the position relationship between the local maps constructed by the other target objects and the local map of the first target object by using the relationship between the visual features in the environment images acquired by the target objects, and realizes the cooperative construction of the global map of the first target object by fusing the local maps of the multiple target objects and the local map of the first target object, thereby improving the map construction efficiency without loading a lidar, reducing the global map construction cost, and further ensuring the accuracy of the final global map by using the visual features in the original environment images acquired by the target objects.

As shown in fig. 3, an embodiment of the present invention further provides an electronic device, which may include a processor 901 and a memory 902, where the processor 901 and the memory 902 may be connected through a bus or in another manner, and fig. 3 takes the connection through the bus as an example.

Processor 901 may be a Central Processing Unit (CPU). The Processor 901 may also be other general purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs) or other Programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components, or combinations thereof.

The memory 902, which is a non-transitory computer readable storage medium, may be used to store non-transitory software programs, non-transitory computer executable programs, and modules, such as program instructions/modules corresponding to the methods in the embodiments of the present invention. The processor 901 executes various functional applications and data processing of the processor, i.e., implements the above-described method, by executing non-transitory software programs, instructions, and modules stored in the memory 902.

The memory 902 may include a storage program area and a storage data area, wherein the storage program area may store an application program required for operating the device, at least one function; the storage data area may store data created by the processor 901, and the like. Further, the memory 902 may include high speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, the memory 902 may optionally include memory located remotely from the processor 901, which may be connected to the processor 901 via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.

One or more modules are stored in the memory 902 and when executed by the processor 901 perform the methods described above.

The specific details of the server may be understood by referring to the corresponding related descriptions and effects in the above method embodiments, and are not described herein again.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program, and the implemented program can be stored in a computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. The storage medium may be a magnetic Disk, an optical Disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a Flash Memory (Flash Memory), a Hard Disk (Hard Disk Drive, abbreviated as HDD) or a Solid State Drive (SSD), etc.; the storage medium may also comprise a combination of memories of the kind described above.

Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations fall within the scope defined by the appended claims.

Claims

1. A global map construction method is applied to any one target object in a collaborative map construction system, the collaborative map construction system comprises a plurality of target objects, and the method is characterized by comprising the following steps:

2. The method of claim 1, wherein the fusing the first local map and the second local maps corresponding to each of the remaining target objects respectively based on the relationship between the first visual feature and the second visual features corresponding to the remaining target objects comprises:

3. The method of claim 2, wherein the processing the first local map and the second local map based on the feature point matching result to obtain a current global map of the first target object and a current target object comprises:

and when the feature point matching result meets the preset feature point matching requirement, performing map fusion of the overlapping position on the first local map and the second local map to obtain a global map corresponding to the current target object.

4. The method of claim 3, further comprising:

5. The method of claim 2, wherein prior to feature point matching the first visual feature and the second visual feature corresponding to the current target object, the method further comprises:

judging whether the first distance is smaller than a distance threshold value;

6. The method of claim 5, further comprising:

7. The method of claim 1, wherein said extracting a first visual feature from the first environmental image comprises:

8. A global map construction device is applied to any one target object in a collaborative map construction system, the collaborative map construction system comprises a plurality of target objects, and the device is characterized by comprising:

9. An electronic device, comprising:

a memory and a processor, the memory and the processor being communicatively coupled to each other, the memory having stored therein computer instructions, the processor being configured to execute the computer instructions to perform the method of any of claims 1-7.

10. A computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions that, when executed by a processor, implement the method of any one of claims 1-7.