WO2023070441A1

WO2023070441A1 - Movable platform positioning method and apparatus

Info

Publication number: WO2023070441A1
Application number: PCT/CN2021/127050
Authority: WO
Inventors: 赵峰
Original assignee: 深圳市大疆创新科技有限公司
Priority date: 2021-10-28
Filing date: 2021-10-28
Publication date: 2023-05-04
Also published as: CN117940739A

Abstract

A movable platform positioning method and apparatus, the method comprising: when a movable platform executes a first motion operation in a target area, acquiring images which are taken by a photographing apparatus on the movable platform (S11); establishing an online map of the target area on the basis of the images which are taken by the photographing apparatus at different moments (S12); acquiring an offline map of the target area established when the movable platform has executed a historical motion operation in the target area, wherein the historical motion operation is an operation prior to the first motion operation (S13); and positioning the movable platform on the basis of the image which is taken by the photographing apparatus at the current moment, the online map and the offline map, so as to obtain a first positioning result of the movable platform (S14). By means of the method, when visual repositioning is performed, the online map and the offline map complement each other, so as to cover an operation scenario as much as possible, thereby improving the success rate and accuracy of positioning the movable platform.

Description

Positioning method and device for a movable platform

technical field

The present application relates to the field of positioning, and in particular to a positioning method and device for a movable platform.

Background technique

Without the assistance of external positioning devices (such as GPS/RTK, etc.), the mobile platform needs to rely on its own sensors for positioning (such as visual odometer), but as time goes by, the positioning accuracy of the mobile platform It will gradually decline. At this time, it is necessary to reposition the movable platform to reduce the positioning error.

In the case of no external positioning device (such as GPS) assistance, local visual positioning technology (such as visual odometer) is relatively mature, but the global visual relocalization technology is only in the laboratory research stage, and the current relocalization technology has positioning The accuracy is not high, the robustness is poor and so on. Traditional relocation methods are generally based on offline maps or online maps. Offline maps generally have a certain time interval from establishment to use. During this process, the environment may have changed, which makes offline maps unable to reflect changes in the current environment in real time, especially The availability of offline maps across a long period of time (such as the time dimension of spring, summer, autumn and winter) will decrease rapidly, which will gradually reduce the success rate of relocation; limited by computing power and real-time requirements, online maps can generally only create three-dimensional sparse points The map, and the position of the sparse point is generally not accurate enough, and the positioning through the map is easy to make the relocation error larger.

Contents of the invention

The present application provides a positioning method and device for a movable platform.

Specifically, this application is achieved through the following technical solutions:

In the first aspect, the embodiment of the present application provides a positioning method for a mobile platform, including:

When the movable platform executes the first motion operation in the target area, acquiring an image captured by the photographing device on the movable platform;

Establishing an online map of the target area according to the images captured by the photographing device at different times;

Acquiring an offline map of the target area established when the movable platform executes a historical motion operation in the target area, the historical motion operation is an operation before the first motion operation;

The movable platform is positioned according to the image captured by the photographing device at the current moment, the online map and the offline map, and a first positioning result of the movable platform is obtained.

In a second aspect, an embodiment of the present application provides a positioning device for a movable platform, the device comprising:

a storage device for storing program instructions; and

One or more processors, calling the program instructions stored in the storage device, when the program instructions are executed, the one or more processors are individually or jointly configured to implement the first aspect described method.

In a third aspect, the embodiment of the present application provides a mobile platform, including:

fuselage; and

In the positioning device according to the second aspect, the positioning device is arranged on the fuselage.

In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the method described in the first aspect is implemented.

According to the technical solution provided by the embodiment of the application, the application does not rely solely on the online map or the offline map when performing visual relocation, but positions the mobile platform by combining the online map and the offline map. In addition, it can cover the operation scene as much as possible, and improve the success rate of positioning of the mobile platform; and, integrate the effective information of the online map and the offline map, and improve the positioning accuracy of the mobile platform.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings that need to be used in the description of the embodiments.

FIG. 1 is a schematic flowchart of a positioning method for a mobile platform in an embodiment of the present application;

Fig. 2 is a schematic diagram of an implementation process for locating a movable platform and obtaining a first positioning result of the movable platform according to an image captured by a photographing device at a current moment, an online map, and an offline map in an embodiment of the present application;

FIG. 3 is a schematic flowchart of a positioning method for a mobile platform in another embodiment of the present application;

Fig. 4 is a schematic diagram of a movable platform performing trajectory repetition in the first target area A and the second target area B in an embodiment of the present application;

Fig. 5 is an embodiment of the application according to the second position information of the second feature point in the online map that matches the first feature point in the image at each current moment and the offline map with each current moment The third position information of the third feature point matched with the first feature point in the image, the movable platform is positioned, and the schematic diagram of the implementation process of obtaining the first positioning result of the movable platform;

Fig. 6 is a schematic diagram of a movable platform moving in a target area in an embodiment of the present application;

Fig. 7 is a schematic structural diagram of a positioning device for a movable platform in an embodiment of the present application.

Detailed ways

The technical solutions in the embodiments of the present application will be described below with reference to the drawings in the embodiments of the present application.

It should be noted that, in the case of no conflict, the features in the following embodiments and implementation manners can be combined with each other.

In this application, "at least one" means one or more, and "multiple" means two or more. "And/or" describes the association relationship of associated objects, indicating that there can be three types of relationships, for example, A and/or B, which can mean: A exists alone, A and B exist at the same time, and B exists alone, where A, B can be singular or plural. The character "/" generally indicates that the contextual objects are an "or" relationship. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (unit) of a, b, or c can represent: a, b, c, a and b, a and c, b and c, or a and b and c, wherein a, b, c can be single or multiple.

The mobile platform in this embodiment of the present application may include unmanned aerial vehicles (such as drones), robots, agricultural machinery, and the like.

The positioning method of the embodiment of the present application can be applied to fields such as autonomous return and trajectory replay of movable platforms, such as autonomous return of drones, automatic inspection of robot warehouses, and automatic agricultural operations.

Fig. 1 is a schematic flow chart of a positioning method for a mobile platform in an embodiment of the present application; the execution body of the positioning method in the embodiment of the present application may include a mobile platform or an independent controller provided on the mobile platform or may A combination of a mobile platform and the aforementioned stand-alone controller. Please refer to FIG. 1 , a method for positioning a mobile platform provided in an embodiment of the present application may include steps S11 to S14.

Wherein, in S11, when the movable platform executes the first motion operation in the target area, an image captured by the photographing device on the movable platform is acquired.

The movable platform may include single or multiple photographing devices, wherein, when the movable platform includes multiple photographing devices, the multiple photographing devices face different directions of the movable platform, and are used to obtain images of the environment in different directions of the movable platform image.

In S12, an online map of the target area is established according to the images captured by the photographing device at different times.

For example, when the movable platform includes a single camera, an online map of the target area is established based on the images captured by the single camera at different times; The images captured at all times are used to establish an online map of the target area, and multiple shooting devices correspond to multiple online maps.

Feature extraction can be performed on the images captured by the shooting device at different times to obtain the position information of the first feature point in each image and the feature expression of each first feature point; then, the same The feature points are associated with the feature points, combined with the pose of the shooting device when shooting each image, the second position information of the second feature point is obtained, and combined with the feature expression, an online map is obtained.

The characteristic expression may include color and/or texture; of course, the content of the characteristic expression may also include other environmental characteristic expressions.

It should be noted that, when the movable platform executes the first movement operation in the target area, the establishment of the online map continues.

Among them, the feature expression can be represented by such as orb/sift/cnn feature, and can also be represented by other image features. Feature expression can be in vector form or in other ways.

The first location information is two-dimensional location information, and the second location information is three-dimensional location information.

In S13, the offline map of the target area established when the movable platform executes the historical motion operation in the target area is acquired, and the historical motion operation is the operation before the first motion operation.

In some embodiments, the historical motion operation is the previous operation of the first motion operation. In this way, the time interval between the historical motion operation and the first motion operation will not be too long to prevent the geographical environment of the target area from being affected due to the long time interval. The variation is large, so that the offline map cannot accurately reflect the characteristics of the target area when the first motion operation is performed.

The way to build the offline map is similar to the way to build the online map. The offline map is built when the movable platform performs historical motion operations in the target area. An offline map has been created during a motion operation.

The offline map can be stored in the storage device of the mobile platform, or in an external storage device (such as the storage device of the remote control terminal of the mobile platform). In S13, when acquiring the offline map of the target area established when the mobile platform performs historical motion operations in the target area, specifically, the offline map is acquired from the storage device of the mobile platform or an external storage device.

In S14, the movable platform is positioned according to the image captured by the photographing device at the current moment, the online map and the offline map, and a first positioning result of the movable platform is obtained.

It should be understood that the image taken at the current moment is acquired in S11.

In the embodiment of the present application, the first positioning result is used to indicate the pose of the movable platform.

Fig. 2 is a schematic diagram of an implementation process for positioning a movable platform and obtaining a first positioning result of the movable platform according to an image captured by a photographing device at the current moment, an online map, and an offline map in an embodiment of the present application; As shown in FIG. 2 , an implementation process of locating a movable platform and obtaining a first positioning result of the movable platform according to an image captured by a photographing device at the current moment, an online map, and an offline map may include S21-S24.

In S21, the first position information of the first feature point of the image at the current moment, the second position information of the second feature point of the online map, and the third position information of the third feature point of the offline map;

Wherein, the first position information is two-dimensional position information, the second position information is three-dimensional position information, and the third position information is also three-dimensional position information.

In the embodiment of the present application, the first location information, the second location information, and the third location information are all coordinates of the real world.

In S22, according to the first position information and the second position information, the image at the current moment is relocated with the online map, and the second feature point matching the first feature point in the online map is determined;

The matching of the first feature point with the second feature point means that the first feature point is associated with the second feature point, and the first position information of the first feature point and the second position information of the second feature point are used to indicate the location of the target area. the same location.

S23. According to the first position information and the third position information, relocate the image at the current moment and the offline map, and determine the third feature point matching the first feature point in the offline map;

Matching of the first feature point with the third feature point means that the first feature point is associated with the third feature point, and the first position information of the first feature point and the third position information of the third feature point are used to indicate the location of the target area. the same location.

It should be understood that S22 and S23 may be executed synchronously, or may be executed sequentially, for example, S22 is executed first, and then S23 is executed, or S23 is executed first, and then S22 is executed.

S24. According to the second position information of the second feature point matching the first feature point in the online map, and the third position information of the third feature point matching the first feature point in the offline map, the movable The platform is positioned, and the first positioning result of the movable platform is obtained.

The offline map and the online map complement each other, which can cover the operation scene as much as possible and improve the success rate of positioning of the mobile platform; The precision of the relocation.

When implementing S24, different fusion positioning methods can be used. For example, in some embodiments, the second position information of the movable platform is determined according to the second position information of the second feature point matching the first feature point in the online map. positioning result; according to the third position information of the third feature point matching the first feature point in the offline map, determine the third positioning result of the movable platform; fuse the second positioning result and the third positioning result to determine the The first location result for mobile platforms.

Wherein, both the second positioning result and the third positioning result are used to indicate the pose of the movable platform.

In some embodiments, the ransac pnp algorithm is used to process the second position information of the second feature point matching the first feature point in the online map to obtain the second positioning result of the movable platform; of course, it can also be used Other algorithms process the second position information of the second feature point matching the first feature point in the online map to obtain the second positioning result of the movable platform.

In some embodiments, the ransac pnp algorithm is used to obtain the third positioning result of the movable platform for the third position information of the third feature point matching the first feature point in the offline map; of course, other algorithms can also be used A third positioning result of the movable platform is obtained for third position information of a third feature point matching the first feature point in the offline map.

Optionally, the ransac pnp algorithm is used to process the second position information of the second feature point matched with the first feature point in the online map to obtain the second positioning result of the movable platform, and the ransac pnp algorithm is used to process the offline The third position information of the third feature point matching the first feature point in the map is used to obtain the third positioning result of the movable platform.

Before fusing the second positioning result and the third positioning result to determine the first positioning result of the movable platform, it is necessary to determine the first weight of the online image and the second weight of the offline image. When fusing the second positioning result and the third positioning result to determine the first positioning result of the movable platform, specifically, fusing the second positioning result and the third positioning result according to the first weight and the second weight, A first positioning result of the movable platform is determined.

The calculation formula of the first positioning result is as follows:

p=w ₁ p ₁ +w ₂ p ₂ (1);

In the formula (1), p is the first positioning result, p ₁ and p ₂ are the second positioning result and the third positioning result respectively, w ₁ and w ₂ are the first weight and the second weight respectively.

The methods for determining the first weight and the second weight may include various methods. For example, in some embodiments, determining the first weight of an online image and the second weight of an offline image may include the following steps:

(1), according to the first quantity of the second feature point matching with the first feature point in the online map, determine the first weight;

(2) Determine the second weight according to the second quantity of the third feature points matching the first feature point in the offline map.

Wherein, steps (1) and (2) can be executed synchronously or sequentially, for example, (1) is executed first, and then (2), or (2) is executed first, and then (1) is executed.

In this embodiment, the first weight is positively correlated with the first quantity, that is, the larger the first quantity, the greater the first weight; the smaller the first quantity, the smaller the first weight. Correspondingly, the second weight is positively correlated with the second quantity, that is, the larger the second quantity, the greater the second weight; the smaller the second quantity, the smaller the second weight. With such a setting, the offline map and the online map can complement each other, so as to cover the operation scene as much as possible, and improve the success rate and positioning accuracy of the mobile platform positioning.

In some embodiments, determining the first weight of the online image and the second weight of the offline image may include the following steps:

(1), reprojecting the second feature point matched with the first feature point in the online map to the image at the current moment to obtain the fourth feature point;

(2), determine the first pixel distance between the fourth feature point and the corresponding first feature point on the image at the current moment;

(3), according to the first pixel distance, determine the first weight;

(4), reprojecting the third feature point matched with the first feature point in the offline map to the image at the current moment to obtain the fifth feature point;

(5), determining the second pixel distance between the fifth feature point and the corresponding first feature point on the image at the current moment;

(6) Determine a second weight according to the second pixel distance.

The first pixel distance and the second pixel distance can be used to characterize the reprojection error.

In this embodiment, the first weight is negatively correlated with the first pixel distance, that is, the larger the first pixel distance is, the larger the reprojection error is, and the smaller the first weight is; the smaller the first pixel distance is, the larger the reprojection error is. The smaller the value, the greater the first weight. Correspondingly, the second weight is negatively correlated with the second pixel distance, that is, the larger the second pixel distance is, the larger the reprojection error is, and the smaller the second weight is; the smaller the second pixel distance is, the smaller the reprojection error is. The greater the weight of the second. With such a setting, the offline map and the online map can complement each other, so as to cover the operation scene as much as possible, and improve the success rate and positioning accuracy of the mobile platform positioning.

In another embodiment, the process of realizing the fusion positioning method of S24 may include: according to the second position information of the second feature point matched with the first feature point in the online map and the second position information of the second feature point matched with the first feature point in the offline map The third position information of the third feature point forms the first data group; the position information in the first data group is processed by ransac pnp algorithm to obtain the first positioning result of the movable platform. This fusion method uses the fusion at the feature level, which improves the success rate and positioning accuracy of the mobile platform positioning.

It should be understood that the first data set is the second position information of the second feature point matching the first feature point in the online map and the third position information of the third feature point matching the first feature point in the offline map collection of information. In addition, the ransac pnp algorithm can also be replaced by other relocation algorithms.

The positioning method of the embodiment of the present application does not rely solely on online maps or offline maps when performing visual relocation. The offline maps have high precision but are not real-time, and the online maps have low precision but are available in real time. This application combines online maps and offline maps. Mobile platform for positioning, online map and offline map complement each other, can cover the operation scene as much as possible, improve the success rate of mobile platform positioning; and integrate the effective information of online map and offline map to improve the positioning accuracy of mobile platform .

FIG. 3 is a schematic flowchart of a positioning method for a movable platform in another embodiment of the present application; referring to FIG. 3 , the positioning method in an embodiment of the present application may further include S31-S34.

Wherein, in S31, the first position information of the first feature point of the image at the current moment and the third position information of the third feature point of the offline map are obtained;

In S32, according to the first position information and the third position information, the image at the current moment is relocated with the offline map, and the third feature point matching the first feature point in the offline map is determined;

In S33, the third feature point matched with the first feature point in the offline map is re-projected onto the image at the current moment to obtain the fifth feature point;

After executing S33, it is necessary to extract the feature expression of the fifth feature point in the image at the current moment.

In S34, the feature expression corresponding to the third feature point in the offline map is updated according to the feature expression of the fifth feature point.

Updates in this embodiment of the application may include additions or deletions.

For example, in some embodiments, when the feature expression of the fifth feature point is inconsistent with the feature expression of the corresponding third feature point in the offline map, the fifth feature is added to the feature expression of the corresponding third feature point in the offline map feature representation of points. The update of the offline map is not simply to replace the original feature expression, but to establish a multi-modal feature expression. Offline maps generally have problems that cannot cope with long-term changes. For example, the characteristics of spring, summer, autumn and winter in the current area are completely different. For example, a tree in the target area is green in summer and yellow in autumn. In this application, for the same scene, when the feature expression of the fifth feature point is inconsistent with the feature expression of the corresponding third feature point of the offline map, it is considered that a feature expression of a modality is added and saved, which improves the performance of the offline map. Long-term availability and robustness.

For example, the feature expression is represented by a vector, and the feature expression of the fifth feature point is inconsistent with the feature expression of the corresponding third feature point of the offline map may include: the feature expression of the fifth feature point and the feature expression of the corresponding third feature point of the offline map The difference is greater than the preset threshold. That is, when the difference between the feature expression of the fifth feature point and the feature expression of the corresponding third feature point of the offline map is greater than the preset threshold, it is considered that the feature expression of the fifth feature point is different from the feature expression of the corresponding third feature point of the offline map Inconsistent. In addition, when the difference between the feature expression of the fifth feature point and the feature expression of the corresponding third feature point of the offline map is less than or equal to the preset threshold, it is considered that the feature expression of the fifth feature point is different from the corresponding third feature point of the offline map The characteristics are consistent. Wherein, the size of the preset threshold can be set as required.

Each feature expression of the third feature point in the offline map is the corresponding feature expression of the corresponding fifth feature point matched last time, that is, the feature expression of each mode of the third feature point in the offline map Both are the feature expressions of the corresponding modes of the corresponding fifth feature point matched last time, which improves the long-term usability and robustness of the offline map.

In some embodiments, the feature expression of the fifth feature point is matched with the feature expression of the corresponding third feature point in the offline map; when the feature expression of the third feature point in the offline map is not matched to the corresponding fifth feature When the duration of the characteristic expression of the point is greater than or equal to the preset duration, the corresponding characteristic expression is deleted. That is, when the feature expression of a certain modality has not been successfully matched for a long time, it is considered that the environment has undergone irreversible changes (for example, a certain number of trees in the target area has been cut down), and the feature expression of the modality is directly deleted. Wherein, the size of the preset duration can be set as required.

The above positioning method can be tested, for example, see Figure 3, under the condition of no external positioning source (such as GPS/TK, etc.), the environmental features in the first target area A are simplified by wavy lines, and in the second target area B The environmental features at the same position as the environmental features of the first target area A are also simplified by wavy lines, and the environmental features at positions different from the environmental features of the first target area A in the second target area B are represented by straight lines, and the dotted lines Used to divide different local areas. The movable platform moved in the first target area A and built an offline map, and then went to the second target area B for the first trajectory replay. It can be found that the movable platform can move accurately at the wavy line in the second target area B Repeat, gradually drifting in the straight line in the second target area B. After the first trajectory replay, the movable platform performs the second trajectory replay in the second target area B, and it can be found that the movable platform can be accurately replayed, which shows that the movable platform has been replayed during the first trajectory replay. Updated offline maps. After the second trajectory replay, the movable platform moves into the first target area A, and the trajectory replay can still be performed, which means that the offline map stores the two modes of the first target area A and the second target area B. characteristic expression.

The embodiment of the present application also provides a positioning method, which may include:

(1), controlling the movable platform to execute the first motion operation in the first target area and the second motion operation in the second target area in sequence;

Wherein, the first partial area (A1) of the first target area and the second partial area (B1) of the second target area are in the same location area and have the same environmental characteristics. The third partial area (A2) of the first target area and the fourth partial area (B2) of the second target area are in the same location area and have different environmental characteristics. The first partial area (A1) is adjacent to the third partial area (A2), and the second partial area (B1) is adjacent to the fourth partial area (B2).

(2) Acquiring the first motion trajectory of the movable platform in the first partial area (A1) and the second motion trajectory in the third partial area (A2) during the process of performing the first motion operation in the first target area;

(3) Obtaining the third motion track in the second partial area (B2) and the fourth motion track in the fourth partial area (B2) when the movable platform performs the second motion operation in the second target area.

The first motion track is the same as the third motion track, and the second motion track is different from the fourth motion track.

(4) Control the movable platform to perform the third motion operation in the second target area, and obtain the fifth motion trajectory and the fifth motion trajectory in the second partial area (B1) of the movable platform in the second target area during the execution of the third motion operation in the second target area. The sixth motion trajectory of the four local areas (B2);

The third motion track is the same as the fifth motion track, and the fourth motion track is the same as the sixth motion track.

(5) Control the movable platform to perform the fourth motion operation in the first target area, and obtain the seventh motion track and the second motion track in the first local area (A1) of the movable platform in the process of executing the fourth motion operation in the first target area The eighth motion trajectory of the three local areas (A2).

The first motion track is the same as the seventh motion track, and the eighth motion track is the same as the second motion track.

Because of the existence of offline maps, when the movable platform enters the same scene, it can generate basically the same navigation trajectory. The online map created in real time by performing motion operations, if the features of the online map do not match some of the feature expressions of the offline maps of the similar location areas that have been experienced before, it can be converted into feature expressions of different modalities of the offline map for storage.

The traditional positioning method is based on a single shooting device, and the positioning success rate and accuracy are low. For this, in the embodiment of the present application, the success rate and accuracy of relocation can be significantly improved by relocating the feature points of the multi-channel shooting device respectively.

The movable platform in this embodiment includes a plurality of photographing devices facing different directions of the movable platform for acquiring images of environments in different directions of the movable platform. Exemplarily, the movable translation platform may include a front-view camera device arranged at the front of the fuselage, a left-view camera device arranged at the left side of the fuselage, and a right-view camera device arranged at the right side of the fuselage.

Referring to Fig. 4, in S14, the movable platform is positioned according to the image captured by the shooting device at the current moment, the online map and the offline map, and the implementation process of obtaining the first positioning result of the movable platform may include:

S51. Obtain the first position information of the first feature point of the image of each shooting device at the current moment;

S52. Obtain the second position information of the second feature point of the online map and the third position information of the third feature point of the offline map;

It should be noted that in S52, for each photographing device, there is an online map corresponding to each photographing device. The online map corresponding to each photographing device is based on the images captured at different times when the photographing device performs the first movement operation in the target area. established.

S53. According to the first position information and the second position information, respectively relocate the image at each current moment and the online map, and determine the second feature point in the online map that matches the first feature point in the image at each current moment. Feature points;

S53 specifically, according to the first location information and the second location information, respectively relocate the image at each current moment and the online map, and determine that the corresponding online map matches the first feature point in the image at each current moment In this step, there is a one-to-one correspondence between the shooting device, the image at the current moment and the online map.

S54. According to the first position information and the third position information, respectively relocate the image at each current moment and the offline map, and determine the third feature point in the offline map that matches the first feature point in the image at each current moment. Feature points;

It should be noted that in S54, for each photographing device, there is an offline map corresponding to each photographing device. The offline map corresponding to each photographing device is based on the images captured at different times when the photographing device performs historical motion operations in the target area. established.

S53 specifically, according to the first location information and the third location information, respectively relocate the image at each current moment and the offline map, and determine that the corresponding offline map matches the first feature point in the image at each current moment In this step, there is a one-to-one correspondence between the shooting device, the image at the current moment and the offline map.

S55, according to the second position information of the second feature point matching the first feature point in the image at each current moment in the online map and the matching with the first feature point in the image at each current moment in the offline map The third position information of the third feature point is used to locate the movable platform, and obtain the first positioning result of the movable platform.

When implementing S55, different multi-photographing device fusion positioning methods can be used. For example, in some embodiments, according to the first feature point of the second feature point in the online map that matches the first feature point in the image at each current moment 2. Position information and the third position information of the third feature point in the offline map that matches the first feature point in the image at each current moment, to determine the fourth positioning result of the movable platform corresponding to each shooting device; The fourth positioning results of the movable platform respectively corresponding to the plurality of photographing devices are fused to obtain the first positioning result of the movable platform. That is, repositioning is performed on each of the multiple photographing devices, and then the repositioning results of the multiple photographing devices are fused to obtain the first positioning result of the movable platform.

Wherein, the fourth positioning result is used to indicate the pose of the movable platform.

The ransac pnp algorithm can be used to analyze the second position information of the second feature point in the online map of each shooting device that matches the first feature point in the image of the current moment captured by the shooting device and the offline map of the shooting device The third position information of the third feature point matched with the first feature point in the image captured by the camera at the current moment determines the fourth positioning result of the movable platform corresponding to the camera; of course, the ransac pnp algorithm also Can be replaced by other relocation algorithms.

When fusing the fourth positioning results of the movable platform corresponding to the plurality of photographing devices to obtain the first positioning result of the movable platform, specifically, according to the preset third weights respectively corresponding to the plurality of photographing devices, the The fourth positioning results of the movable platform respectively corresponding to the plurality of photographing devices are weighted and averaged to obtain the first positioning result of the movable platform.

Wherein, the third weights corresponding to the plurality of photographing devices can be determined according to the relocation error of each photographing device. For example, for a certain photographing device, the greater the relocation error, the smaller the third weight corresponding to the photographing device; The smaller the error, the larger the third weight of the shooting device. The relocation error can be determined according to the number of matching feature points and/or the reprojected pixel distance, as described in the corresponding part of the above embodiment, and will not be repeated here.

In another embodiment, the process of implementing the multi-camera fusion positioning method of S55 may include: according to the second position information of the second feature point in the online map that matches the first feature point in the image at each current moment And the third position information of the third feature point in the offline map matched with the first feature point in the image at each current moment to form the second data set; use ransac pnp algorithm or bundle adjustment algorithm to the second data set The location information is processed to obtain the first positioning result of the movable platform. This fusion method uses feature-level fusion for multi-directional fusion relocation, rather than simply calculating a relocation result for each direction and then performing fusion. The multi-directional fusion relocation method of the embodiment of the present application can Greatly improve the success rate and positioning accuracy of the movable platform positioning.

For example, referring to FIG. 6 , the movable platform is provided with a first sensor 10 and a second sensor 20 , and the first sensor 10 and the second sensor 20 face different orientations of the fuselage of the movable platform.

(1), control the first movement operation of the movable platform in the target area to obtain the offline map of the target area, and obtain the movement track of the first operation;

(2), block the first sensor 10 or the second sensor 20, control the movable platform to perform the secondary motion operation in the target area, control the unblocked sensor to collect the image of the target area, and obtain the secondary motion based on the image and offline map positioning The trajectory of the job;

The relocation of the movable platform in step (2) is very inaccurate.

(3), control the movable platform to perform three motion operations in the target area, control the first sensor 10 or the second sensor 20 to collect images of the target area, and obtain the motion trajectory of the three operations based on the image and offline map positioning;

The relocation of the movable platform in step (3) is more accurate than the relocation of the movable platform in step (2).

The trajectory deviation between the trajectory of the secondary operation and the trajectory of the first operation is greater than the deviation between the trajectory of the three operations and the trajectory of the first operation, which shows that the mobile platform uses the multi-camera device fusion relocation method in the above embodiment .

Assuming that the UAV uses the downward-looking shooting device and the forward-looking shooting device for fusion positioning, and without an external positioning source (such as GPS/TK, etc.), an environment with poor forward-looking and downward-looking is constructed. Covering a certain shooting device, the UAV cannot be repositioned through another shooting device in a single direction, and cannot complete accurate trajectory replay. As long as the downward-looking camera device and the front-view camera device are not blocked, relocation can be performed, which shows that the UAV uses the multi-camera device fusion relocation method in the above-mentioned embodiment.

Corresponding to the positioning method of the movable platform in the foregoing embodiments, an embodiment of the present application further provides a positioning device of the movable platform. Referring to FIG. 7 , the device for positioning a mobile platform according to the embodiment of the present application may include a storage device and one or more processors.

Wherein, the storage device is used for storing program instructions. The storage device stores an executable instruction computer program of the positioning method of the movable platform, and the storage device may include at least one type of storage medium, and the storage medium includes a flash memory, a hard disk, a multimedia card, a card-type memory (for example, SD or DX memory, etc.), Random Access Memory (RAM), Static Random Access Memory (SRAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), Programmable Read Only Memory (PROM ), magnetic storage, magnetic disks, optical disks, etc. Furthermore, the positioning means of the movable platform may cooperate with network storage means performing the storage function of the memory through a network connection. The memory may be an internal storage unit of the positioning device of the movable platform, such as a hard disk or a memory of the positioning device of the movable platform. Memory also can be the external storage device of the positioning device of movable platform, for example the plug-in type hard disk equipped on the positioning device of movable platform, smart memory card (Smart Media Card, SMC), secure digital (Secure Digital, SD) card , Flash Card (Flash Card) and so on. Further, the storage may also include both an internal storage unit of the positioning device of the movable platform and an external storage device. Memory is used to store computer programs and other programs and data needed by the device. The memory can also be used to temporarily store data that has been output or will be output.

One or more processors call the program instructions stored in the storage device, and when the program instructions are executed, the one or more processors are individually or jointly configured to implement the positioning method in the above embodiments.

The processor of this embodiment can implement the positioning method of the embodiment shown in FIG. 1 or FIG. 2 or FIG. 3 or FIG. 5 of this application. The positioning device of this embodiment can be described with reference to the positioning method of the above embodiment.

The processor can be a central processing unit (Central Processing Unit, CPU), and can also be other general-purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), on-site Programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor, or the processor may be any conventional processor, or the like.

An embodiment of the present application provides a movable platform. The movable platform may include a fuselage and the positioning device in the above embodiment, and the positioning device is arranged on the fuselage.

In addition, an embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the method described in the first aspect is implemented.

The computer-readable storage medium may be an internal storage unit of the removable platform described in any of the foregoing embodiments, such as a hard disk or a memory. The computer-readable storage medium can also be an external storage device of a removable platform, such as a plug-in hard disk equipped on the device, a smart memory card (Smart Media Card, SMC), an SD card, a flash memory card (Flash Card) wait. Further, the computer-readable storage medium may also include both an internal storage unit of the removable platform and an external storage device. The computer-readable storage medium is used to store the computer program and other programs and data required by the removable platform, and can also be used to temporarily store outputted or to-be-exported data.

Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be implemented through computer programs to instruct related hardware, and the programs can be stored in a computer-readable storage medium. During execution, it may include the processes of the embodiments of the above-mentioned methods. Wherein, the storage medium may be a magnetic disk, an optical disk, a read-only memory (Read-Only Memory, ROM) or a random access memory (Random Access Memory, RAM), etc.

The above disclosures are only part of the embodiments of the present application, and certainly cannot be used to limit the scope of the present application. Therefore, equivalent changes made according to the claims of the present application still fall within the scope of the present application.

Claims

A positioning method for a movable platform, characterized in that it comprises:

When the movable platform executes the first motion operation in the target area, acquiring an image captured by the photographing device on the movable platform;

Establishing an online map of the target area according to the images captured by the photographing device at different times;

Acquiring an offline map of the target area established when the movable platform executes a historical motion operation in the target area, the historical motion operation is an operation before the first motion operation;

The movable platform is positioned according to the image captured by the photographing device at the current moment, the online map and the offline map, and a first positioning result of the movable platform is obtained.
The method according to claim 1, wherein the movable platform is positioned according to the image captured by the photographing device at the current moment, the online map and the offline map, and the movable platform is obtained. The top positioning results for mobile platforms, including:

Acquiring the first position information of the first feature point of the image at the current moment, the second position information of the second feature point of the online map, and the third position information of the third feature point of the offline map;

According to the first location information and the second location information, relocate the image at the current moment with the online map, and determine a second feature in the online map that matches the first feature point point;

According to the first location information and the third location information, relocate the image at the current moment with the offline map, and determine a third feature matching the first feature point in the offline map point;

According to the second position information of the second feature point matching the first feature point in the online map, and the third feature point matching the first feature point in the offline map Position information, positioning the movable platform, and obtaining a first positioning result of the movable platform.
The method according to claim 2, characterized in that, according to the second position information of the second feature point in the online map that matches the first feature point, and, in the offline map that matches the first feature point, The third position information of the third feature point matching the first feature point is used to locate the movable platform and obtain the first positioning result of the movable platform, including:

determining a second positioning result of the movable platform according to second position information of a second feature point matching the first feature point in the online map;

determining a third positioning result of the movable platform according to third position information of a third feature point matching the first feature point in the offline map;

The second positioning result and the third positioning result are fused to determine the first positioning result of the movable platform.
The method according to claim 3, characterized in that before the fusing of the second positioning result and the third positioning result, before determining the first positioning result of the movable platform, further comprising:

determining a first weight for the online image and a second weight for the offline image;

The fusing the second positioning result and the third positioning result to determine the first positioning result of the movable platform includes:

According to the first weight and the second weight, the second positioning result and the third positioning result are fused to determine a first positioning result of the movable platform.
The method according to claim 4, wherein said determining the first weight of the online image and the second weight of the offline image comprises:

determining the first weight according to a first number of second feature points matching the first feature point in the online map;

The second weight is determined according to a second number of third feature points matching the first feature point in the offline map.
The method according to claim 5, wherein the first weight is positively correlated with the first quantity, and the second weight is positively correlated with the second quantity.
The method according to claim 4, wherein said determining the first weight of the online image and the second weight of the offline image comprises:

Reprojecting the second feature point matching the first feature point in the online map onto the image at the current moment to obtain a fourth feature point;

determining a first pixel distance between the fourth feature point and the corresponding first feature point on the image at the current moment;

determining the first weight according to the first pixel distance;

Reprojecting a third feature point matching the first feature point in the offline map onto the image at the current moment to obtain a fifth feature point;

determining a second pixel distance between the fifth feature point and the corresponding first feature point on the image at the current moment;

Determine the second weight according to the second pixel distance.
The method according to claim 7, wherein the first weight is negatively correlated with the first pixel distance, and the second weight is negatively correlated with the second pixel distance.
The method according to claim 3, wherein the second position information of the movable platform is determined according to the second position information of the second feature point matching the first feature point in the online map. The positioning result includes: using the ransac pnp algorithm to process the second position information of the second feature point matching the first feature point in the online map to obtain a second positioning result of the movable platform; and /or,

The determining the third positioning result of the movable platform according to the third position information of the third feature point matched with the first feature point in the offline map includes: using ransac pnp algorithm for the offline The third position information of the third feature point in the map that matches the first feature point is used to obtain a third positioning result of the movable platform.
The method according to claim 2, characterized in that, according to the second position information of the second feature point matching the first feature point in the online map and the second feature point matching the first feature point in the offline map The third position information of the third feature point matched with one feature point is used to locate the movable platform and obtain the first positioning result of the movable platform, including:

According to the second position information of the second feature point matching the first feature point in the online map and the third position information of the third feature point matching the first feature point in the offline map , forming the first data group;

The position information in the first data group is processed by using the ransac pnp algorithm to obtain a first positioning result of the movable platform.
The method according to claim 1, further comprising:

Acquiring the first position information of the first feature point of the image at the current moment and the third position information of the third feature point of the offline map;

According to the first location information and the third location information, relocate the image at the current moment with the offline map, and determine a third feature matching the first feature point in the offline map point;

Reprojecting a third feature point matching the first feature point in the offline map onto the image at the current moment to obtain a fifth feature point;

The feature expression corresponding to the third feature point in the offline map is updated according to the feature expression of the fifth feature point.
The method according to claim 11, wherein, according to the feature expression of the fifth feature point, updating the feature expression of the corresponding third feature point in the offline map includes:

When the feature expression of the fifth feature point is inconsistent with the feature expression of the corresponding third feature point in the offline map, adding the fifth feature point to the feature expression of the corresponding third feature point in the offline map characteristic expression.
The method according to claim 12, wherein the feature expression is represented by a vector, and the feature expression of the fifth feature point is inconsistent with the feature expression of the corresponding third feature point of the offline map comprising: the first The difference between the feature expression of the five feature points and the feature expression of the corresponding third feature point of the offline map is greater than a preset threshold.
The method according to claim 12, wherein each feature expression of the third feature point in the offline map is a corresponding feature expression of the corresponding fifth feature point matched last time.
The method of claim 11, wherein the characteristic expression comprises color and/or texture.
The method according to claim 11, wherein the updating of the feature expression of the corresponding third feature point in the offline map according to the feature expression of the fifth feature point includes:

matching the feature expression of the fifth feature point with the feature expression of the corresponding third feature point in the offline map;

When the feature expression of the third feature point in the offline map does not match the feature expression of the corresponding fifth feature point for a duration greater than or equal to a preset duration, the corresponding feature expression is deleted.
The method according to claim 1, wherein the movable platform includes a plurality of photographing devices, and the plurality of photographing devices are oriented in different directions of the movable platform, and are used to capture images in different directions of the movable platform. images of the environment;

The positioning of the movable platform according to the image captured by the shooting device at the current moment, the online map and the offline map, and obtaining the first positioning result of the movable platform includes:

Acquiring the first position information of the first feature point of the image of each shooting device at the current moment;

Acquiring second position information of a second feature point of the online map and third position information of a third feature point of the offline map;

According to the first location information and the second location information, relocate the image at each current moment and the online map respectively, and determine the first feature in the online map and the image at each current moment point matching the second feature point;

According to the first location information and the third location information, relocate the image at each current moment and the offline map respectively, and determine the first feature in the offline map and the image at each current moment point matching the third feature point;

According to the second position information of the second feature point matching the first feature point in the image at each current moment in the online map and the first feature point in the image at each current moment in the offline map The third location information of the matched third feature point is used to locate the movable platform, and obtain a first positioning result of the movable platform.
The method according to claim 17, characterized in that, according to the second position information of the second feature point in the online map that matches the first feature point in the image at each current moment and the offline The third position information of the third feature point in the map that matches the first feature point in the image at each current moment is used to locate the movable platform and obtain the first positioning result of the movable platform, including :

According to the second position information of the second feature point matching the first feature point in the image at each current moment in the online map and the first feature point in the image at each current moment in the offline map Matching the third position information of the third feature point to determine the fourth positioning result of the movable platform corresponding to each shooting device;

The fourth positioning results of the movable platform respectively corresponding to the plurality of photographing devices are fused to obtain the first positioning result of the movable platform.
The method according to claim 18, characterized in that merging the fourth positioning results of the movable platform respectively corresponding to a plurality of the photographing devices to obtain the first positioning result of the movable platform, include:

According to the preset third weights respectively corresponding to the multiple shooting devices, the weighted average is performed on the fourth positioning results of the movable platform respectively corresponding to the multiple shooting devices to obtain the first position of the movable platform. positioning results.
The method according to claim 17, characterized in that, according to the second position information of the second feature point in the online map that matches the first feature point in the image at each current moment and the offline The third position information of the third feature point in the map that matches the first feature point in the image at each current moment is used to locate the movable platform and obtain the first positioning result of the movable platform, including :

According to the second position information of the second feature point matching the first feature point in the image at each current moment in the online map and the first feature point in the image at each current moment in the offline map matching third position information of the third feature point to form a second data set;

The position information of the second data group is processed by using a ransac pnp algorithm or a bundle adjustment algorithm to obtain a first positioning result of the movable platform.
A positioning device for a movable platform, characterized in that the device comprises:

a storage device for storing program instructions; and

One or more processors that invoke program instructions stored in the memory device, and when the program instructions are executed, the one or more processors are individually or collectively configured to implement:

When the movable platform executes the first motion operation in the target area, acquiring an image captured by the photographing device on the movable platform;

Establishing an online map of the target area according to the images captured by the photographing device at different times;

Acquiring an offline map of the target area established when the movable platform executes a historical motion operation in the target area, the historical motion operation is an operation before the first motion operation;

The movable platform is positioned according to the image captured by the photographing device at the current moment, the online map and the offline map, and a first positioning result of the movable platform is obtained.
The device according to claim 21, wherein the one or more processors perform an operation on the movable platform according to the image captured by the capturing device at the current moment, the online map, and the offline map. When performing positioning and obtaining the first positioning result of the movable platform, it is further configured to implement individually or jointly:

Acquiring the first position information of the first feature point of the image at the current moment, the second position information of the second feature point of the online map, and the third position information of the third feature point of the offline map;

According to the first location information and the second location information, relocate the image at the current moment with the online map, and determine a second feature in the online map that matches the first feature point point;

According to the first location information and the third location information, relocate the image at the current moment with the offline map, and determine a third feature matching the first feature point in the offline map point;

According to the second position information of the second feature point matching the first feature point in the online map, and the third feature point matching the first feature point in the offline map Position information, positioning the movable platform, and obtaining a first positioning result of the movable platform.
The apparatus according to claim 22, wherein said one or more processors are based on second location information of a second feature point matching said first feature point in said online map, and, The third position information of the third feature point matching the first feature point in the offline map is used to locate the movable platform, and when the first positioning result of the movable platform is obtained, separately or collectively are further configured to implement:

determining a second positioning result of the movable platform according to second position information of a second feature point matching the first feature point in the online map;

determining a third positioning result of the movable platform according to third position information of a third feature point matching the first feature point in the offline map;

The second positioning result and the third positioning result are fused to determine the first positioning result of the movable platform.
The device according to claim 23, wherein the one or more processors fuse the second positioning result and the third positioning result to determine the first positioning result of the movable platform Previously, individually or collectively, it was also configured to implement:

determining a first weight for the online image and a second weight for the offline image;

When the one or more processors fuse the second positioning result and the third positioning result to determine the first positioning result of the movable platform, they are further configured, individually or jointly, for Implementation:

According to the first weight and the second weight, the second positioning result and the third positioning result are fused to determine a first positioning result of the movable platform.
The apparatus of claim 24, wherein the one or more processors are further configured individually or collectively when determining the first weight of the online image and the second weight of the offline image into the implementation:

determining the first weight according to a first number of second feature points matching the first feature point in the online map;

The second weight is determined according to a second number of third feature points matching the first feature point in the offline map.
The apparatus according to claim 25, wherein the first weight is positively correlated with the first quantity, and the second weight is positively correlated with the second quantity.
The apparatus of claim 24, wherein the one or more processors are further configured individually or collectively when determining the first weight of the online image and the second weight of the offline image into the implementation:

Reprojecting the second feature point matching the first feature point in the online map onto the image at the current moment to obtain a fourth feature point;

determining a first pixel distance between the fourth feature point and the corresponding first feature point on the image at the current moment;

determining the first weight according to the first pixel distance;

Reprojecting a third feature point matching the first feature point in the offline map onto the image at the current moment to obtain a fifth feature point;

determining a second pixel distance between the fifth feature point and the corresponding first feature point on the image at the current moment;

Determine the second weight according to the second pixel distance.
The apparatus according to claim 27, wherein the first weight is negatively correlated with the first pixel distance, and the second weight is negatively correlated with the second pixel distance.
The apparatus according to claim 23, wherein said one or more processors determine said first feature point based on second position information of a second feature point matching said first feature point in said online map. When the second positioning result of the movable platform is used, it is further configured individually or jointly for implementation: using the ransac pnp algorithm for the second feature point matching the first feature point in the online map Processing the second position information to obtain a second positioning result of the movable platform; and/or,

When the one or more processors determine the third positioning result of the movable platform according to the third position information of the third feature point matching the first feature point in the offline map, independently Or jointly be further configured to implement: using the ransac pnp algorithm to obtain the third position information of the third feature point matching the first feature point in the offline map, and obtain the third position of the movable platform positioning results.
The device according to claim 22, wherein said one or more processors are based on the second location information of a second feature point matched with said first feature point in said online map and said The third position information of the third feature point matching the first feature point in the offline map is used to locate the movable platform, and when the first positioning result of the movable platform is obtained, individually or jointly is further configured to implement:

According to the second position information of the second feature point matching the first feature point in the online map and the third position information of the third feature point matching the first feature point in the offline map , forming the first data group;

The position information in the first data group is processed by using the ransac pnp algorithm to obtain a first positioning result of the movable platform.
The apparatus according to claim 21, wherein the one or more processors, individually or collectively, are further configured to implement a process comprising:

Acquiring the first position information of the first feature point of the image at the current moment and the third position information of the third feature point of the offline map;

According to the first location information and the third location information, relocate the image at the current moment with the offline map, and determine a third feature matching the first feature point in the offline map point;

Reprojecting a third feature point matching the first feature point in the offline map onto the image at the current moment to obtain a fifth feature point;

The feature expression corresponding to the third feature point in the offline map is updated according to the feature expression of the fifth feature point.
The device according to claim 31, wherein the one or more processors update the feature expression of the corresponding third feature point in the offline map according to the feature expression of the fifth feature point , are further configured individually or collectively to implement:

When the feature expression of the fifth feature point is inconsistent with the feature expression of the corresponding third feature point in the offline map, adding the fifth feature point to the feature expression of the corresponding third feature point in the offline map characteristic expression.
The device according to claim 32, wherein the feature expression is represented by a vector, and the feature expression of the fifth feature point is inconsistent with the feature expression of the corresponding third feature point of the offline map comprising: the first The difference between the feature expression of the five feature points and the feature expression of the corresponding third feature point of the offline map is greater than a preset threshold.
The device according to claim 32, wherein each feature expression of the third feature point in the offline map is a corresponding feature expression of the corresponding fifth feature point matched last time.
The apparatus of claim 31, wherein the characteristic expression comprises color and/or texture.
The device according to claim 31, wherein when the one or more processors update the feature expression of the corresponding third feature point in the offline map according to the feature expression of the fifth feature point , individually or collectively are further configured to implement:

matching the feature expression of the fifth feature point with the feature expression of the corresponding third feature point in the offline map;

When the feature expression of the third feature point in the offline map does not match the feature expression of the corresponding fifth feature point for a duration greater than or equal to a preset duration, the corresponding feature expression is deleted.
The device according to claim 21, wherein the movable platform includes a plurality of photographing devices, and the plurality of photographing devices are directed towards different directions of the movable platform, and are used to capture images in different directions of the movable platform. images of the environment;

The one or more processors locate the movable platform according to the image captured by the photographing device at the current moment, the online map and the offline map, and obtain the first positioning of the movable platform The results, individually or collectively, are further configured for implementation:

Acquiring the first position information of the first feature point of the image of each shooting device at the current moment;

Acquiring second position information of a second feature point of the online map and third position information of a third feature point of the offline map;

According to the first location information and the second location information, relocate the image at each current moment and the online map respectively, and determine the first feature in the online map and the image at each current moment point matching the second feature point;

According to the first location information and the third location information, relocate the image at each current moment and the offline map respectively, and determine the first feature in the offline map and the image at each current moment point matching the third feature point;

According to the second position information of the second feature point matching the first feature point in the image at each current moment in the online map and the first feature point in the image at each current moment in the offline map The third location information of the matched third feature point is used to locate the movable platform, and obtain a first positioning result of the movable platform.
The apparatus according to claim 37, wherein the one or more processors are configured according to the second feature point in the online map that matches the first feature point in the image at each current moment. 2. Location information and third location information of a third feature point in the offline map that matches the first feature point in the image at each current moment, positioning the movable platform, and obtaining the movable platform The first positioning results, individually or collectively, are further configured for implementation:

According to the second position information of the second feature point in the online map that matches the first feature point in the image at each current moment and the first feature point in the offline map with the image at each current moment Matching the third position information of the third feature point to determine the fourth positioning result of the movable platform corresponding to each shooting device;

The fourth positioning results of the movable platform respectively corresponding to the plurality of photographing devices are fused to obtain the first positioning result of the movable platform.
The device according to claim 38, wherein the one or more processors fuse the fourth positioning results of the movable platforms respectively corresponding to a plurality of the photographing devices to obtain the movable The platform's first positioning results, individually or collectively, are further configured to implement:

According to the preset third weights respectively corresponding to the multiple shooting devices, the weighted average is performed on the fourth positioning results of the movable platform respectively corresponding to the multiple shooting devices to obtain the first position of the movable platform. positioning results.
The apparatus according to claim 37, wherein the one or more processors are configured according to the second feature point in the online map that matches the first feature point in the image at each current moment. 2. Location information and third location information of a third feature point in the offline map that matches the first feature point in the image at each current moment, positioning the movable platform, and obtaining the movable platform The first location results, individually or collectively, are further configured for implementation:

According to the second position information of the second feature point matching the first feature point in the image at each current moment in the online map and the first feature point in the image at each current moment in the offline map matching third position information of the third feature point to form a second data set;

The position information of the second data group is processed by using a ransac pnp algorithm or a bundle adjustment algorithm to obtain a first positioning result of the movable platform.
A mobile platform, characterized in that it comprises:

fuselage; and

The positioning device according to any one of claims 21 to 40, the positioning device is arranged on the fuselage.
A computer-readable storage medium, on which a computer program is stored, wherein, when the program is executed by a processor, the method according to any one of claims 1 to 20 is realized.