A method, device and storage medium for generating a map model

Abstract

The present application provides a method, device and storage medium for generating a map model, which relates to the field of communications and is used to improve the accuracy of a map model (such as a three-dimensional model). The method includes: acquiring multiple first images and multiple first positions, the multiple first images include a first object to be rescued and a scene object, the first position is a position where the first image is collected, and the multiple first positions correspond to the multiple first images. Based on the multiple first images and the multiple first positions, a first map model is generated, the first map model is used to reflect the disaster scene, and the first map model includes scene objects. The location information of the first object to be rescued is obtained. The first map model is processed according to the location information of the first object to be rescued to generate a second map model, and the second map model includes the location information of the first object to be rescued, the first object to be rescued and the scene object.

Description

Technical Field

[0001] The present application relates to the field of communications, and in particular to a method, device and storage medium for generating a map model. Background Art

[0002] In recent years, with the development of drone technology, drones can be used in a variety of scenarios. For example, after disasters such as earthquakes, floods, and mudslides, rescue workers can understand the on-site situation through videos collected by drones.

[0003] At present, when a drone is at a disaster site, it needs to first collect videos and transmit the collected videos back to the terminal. Afterwards, rescuers can understand the situation at the disaster site and make a rescue plan based on the video content played by the terminal. However, in the above technical solution, the drone only collects videos of the disaster site, and the amount of information collected is relatively small, which seriously affects the rescuers' judgment of the disaster site and delays the rescue time. Summary of the invention

[0004] The present application provides a method, device and storage medium for generating a map model, which are used to improve the accuracy of a map model (such as a three-dimensional model).

[0005] In order to achieve the above objectives, this application adopts the following technical solutions:

[0006] According to a first aspect of the present application, a method for generating a map model is provided. The method comprises:

[0007] A map model generation device (which may be referred to as a "generation device") obtains a plurality of first images and a plurality of first positions, wherein the plurality of first images include a first object to be rescued and a scene object, the first position is a position where the first image is collected, and the plurality of first positions correspond to the plurality of first images. The generation device generates a first map model based on the plurality of first images and the plurality of first positions, wherein the first map model is used to reflect the disaster scene, and the first map model includes a scene object. The generation device obtains the position information of the first object to be rescued. The generation device processes the first map model according to the position information of the first object to be rescued, and generates a second map model, wherein the second map model includes the position information of the first object to be rescued, the first object to be rescued, and the scene object.

[0008] Optionally, the method for generating a map model further includes: the generating device acquires a plurality of second images, the plurality of second images being all the collected images. The generating device acquires a plurality of first images, including: the generating device determines a plurality of first images from a plurality of second images according to a first position corresponding to a target image and a plurality of first positions, the target image being any second image, and a distance between the first positions corresponding to the plurality of first images and the first position corresponding to the target image being less than a preset distance threshold.

[0009] Optionally, the method for generating a map model further includes: the generating device generates a plurality of second positions according to the plurality of first positions and the plurality of first images, the reprojection error of the first image acquired at the second position is smaller than the reprojection error of the first image acquired at the first position, and the plurality of second positions correspond to the plurality of first images. The generating device generates the first map model according to the plurality of second positions and the plurality of first images.

[0010] Optionally, the map model generation method further includes: the generation device processes the multiple second positions and the multiple first images through a target preset algorithm to determine a dense point cloud, the target preset algorithm is used to determine a coordinate set corresponding to the pixel points of the multiple first images, and the dense point cloud is determined by the coordinate set corresponding to the pixel points of the multiple first images. The generation device processes the dense point cloud to generate a triangulated network model. The generation device renders the triangulated network model through the multiple first images to generate the first map model.

[0011] Optionally, the plurality of first images include the second object to be rescued, and the second object to be rescued includes the first object to be rescued. The method for generating the map model further includes: the generating device performs deduplication processing on the second object to be rescued to obtain the first object to be rescued. The generating device determines the position information of the first object to be rescued based on the plurality of first positions and the plurality of target position relationships, and the target position relationship is the relationship between the first position and the position of the first object to be rescued.

[0012] According to a second aspect of the present application, a device for generating a map model is provided, the device comprising an acquisition module and a processing module.

[0013] The acquisition module is used to acquire multiple first images and multiple first positions, wherein the multiple first images include a first object to be rescued and a scene object, the first position is a position where the first image is collected, and the multiple first positions correspond to the multiple first images. The processing module is used to generate a first map model based on the multiple first images and the multiple first positions, wherein the first map model is used to reflect the disaster scene, and the first map model includes the scene object. The acquisition module is also used to acquire the position information of the first object to be rescued. The processing module is specifically used to process the first map model according to the position information of the first object to be rescued to generate a second map model, wherein the second map model includes the position information of the first object to be rescued, the first object to be rescued, and the scene object.

[0014] Optionally, the acquisition module is further used to acquire multiple second images, where the multiple second images are all the acquired images. The processing module is further used to determine multiple first images from the multiple second images according to the first position corresponding to the target image and the multiple first positions, where the target image is any second image, and the distance between the first position corresponding to the multiple first images and the first position corresponding to the target image is less than a preset distance threshold.

[0015] Optionally, the processing module is further used to generate a plurality of second positions according to the plurality of first positions and the plurality of first images, the reprojection error of the first image collected at the second position is smaller than the reprojection error of the first image collected at the first position, and the plurality of second positions correspond to the plurality of first images. The processing module is further used to generate a first map model according to the plurality of second positions and the plurality of first images.

[0016] Optionally, the processing module is further used to process the plurality of second positions and the plurality of first images through a target preset algorithm to determine a dense point cloud, the target preset algorithm is used to determine a coordinate set corresponding to the pixel points of the plurality of first images, and the dense point cloud is determined by the coordinate set corresponding to the pixel points of the plurality of first images. The processing module is further used to process the dense point cloud to generate a triangulated network model. The processing module is further used to render the triangulated network model through the plurality of first images to generate a first map model.

[0017] Optionally, the plurality of first images include the second object to be rescued, and the second object to be rescued includes the first object to be rescued. The processing module is further used to perform deduplication processing on the second object to be rescued to obtain the first object to be rescued. The processing module is further used to determine the position information of the first object to be rescued based on the plurality of first positions and the plurality of target position relationships, wherein the target position relationship is the relationship between the first position and the position of the first object to be rescued.

[0018] According to a third aspect of the present application, a map model generation device is provided, the device comprising: a processor and a memory. The processor and the memory are coupled. The memory is used to store one or more programs, the one or more programs including computer executable instructions, and when the map model generation device is running, the processor executes the computer executable instructions stored in the memory to implement the map model generation method as described in the first aspect and any possible implementation of the first aspect.

[0019] According to a fourth aspect of the present application, a computer-readable storage medium is provided, in which instructions are stored. When the instructions are executed on a computer, the computer executes the method for generating a map model described in the first aspect and any possible implementation of the first aspect.

[0020] According to a fifth aspect of the present application, a computer program product is provided, comprising a computer program, which, when executed by a processor, enables the computer to implement the method for generating a map model as described in the first aspect and any possible implementation manner of the first aspect.

[0021] In the above scheme, the technical problems that can be solved and the technical effects achieved by the map model generation device, computer equipment, computer storage medium or computer program product can refer to the technical problems and technical effects solved by the above-mentioned first aspect, and will not be repeated here.

[0022] The technical solution provided by the present application brings at least the following beneficial effects: the generating device can obtain multiple first images and multiple first positions, the multiple first images include the first object to be rescued and the scene object, the first position is the position of collecting the first image, and the multiple first positions correspond to the multiple first images. Afterwards, the generating device can generate a first map model based on the multiple first images and the multiple first positions, the first map model is used to reflect the disaster scene, and the first map model includes the scene object. In addition, the generating device can obtain the position information of the first object to be rescued, and process the first map model through the position information of the first object to be rescued to generate a second map model, the second map model includes the position information of the first object to be rescued, the first object to be rescued and the scene object. In other words, the generating device can mark the position information of the first object to be rescued and the objects in the disaster scene (such as the first object to be rescued and the scene object) on the second map model. In this way, the amount of information in the map model can be increased, and then it can provide valuable reference for rescue personnel to reasonably allocate rescue forces, determine key disaster relief areas, select safe rescue routes, and select sites for post-disaster reconstruction. In addition, compared with the video of the disaster scene, the amount of data of the second map model in the technical solution of the present application is small. In this way, the amount of data transmitted by the generating device can be reduced, saving the time for the generating device to transmit information to the terminal, thereby speeding up the rescue progress and improving the rescue efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 is a schematic diagram of a communication system according to an exemplary embodiment;

[0024] Figure 2 is a flowchart of a method for generating a map model according to an exemplary embodiment;

[0025] Figure 3 is a flowchart of another method for generating a map model according to an exemplary embodiment;

[0026] Figure 4 is a schematic diagram showing an example of determining location information according to an exemplary embodiment;

[0027] Figure 5 is a schematic diagram showing an example of marking position information according to an exemplary embodiment;

[0028] Figure 6is a flowchart of another method for generating a map model according to an exemplary embodiment;

[0029] Figure 7 is a flowchart of another method for generating a map model according to an exemplary embodiment;

[0030] Figure 8 is a schematic diagram showing an example of generating a map model according to an exemplary embodiment;

[0031] Fig. 9 is a flowchart of another method for generating a map model according to an exemplary embodiment;

[0032] Fig.10 is a flowchart of another method for generating a map model according to an exemplary embodiment;

[0033] Fig.11 is a flowchart of another method for generating a map model according to an exemplary embodiment;

[0034] Fig.12 is a structural block diagram of a device for generating a map model according to an exemplary embodiment;

[0035] Fig.13 is a structural schematic diagram of a device for generating a map model according to an exemplary embodiment;

[0036] Fig.14 The present invention is a conceptual partial view of a computer program product according to an exemplary embodiment. DETAILED DESCRIPTION

[0037] In order to enable ordinary persons in the art to better understand the technical solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0038] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the attached claims.

[0039] In addition, the terms "including" and "having" and any variations thereof mentioned in the description of this application are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or modules is not limited to the listed steps or modules, but may optionally include other steps or modules that are not listed, or may optionally include other steps or modules that are inherent to these processes, methods, products or devices.

[0040] In addition, in the embodiments of the present application, words such as "exemplary" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present concepts in a specific way.

[0041] First, the application scenarios of the embodiments of the present application are introduced.

[0042] The method for generating a map model in an embodiment of the present application is applied to a scenario in which a terminal (such as a drone) collects information about a disaster scene. In related technologies, when a drone collects information about a disaster scene, it is necessary to first collect the disaster scene and transmit the collected video back to the terminal. Afterwards, rescuers can understand the situation of the disaster scene and make a rescue plan based on the video content played by the terminal. However, in the current technical solution, drones only collect images of the disaster scene. In this way, without locating the objects in the disaster scene, the rescuers' judgment of the disaster scene is seriously affected, and the rescue time is delayed.

[0043] For example, when the drone detects that it has entered a disaster-stricken village, it can collect videos of the disaster-stricken village and transmit the collected videos of the disaster-stricken village back to the terminal. Afterwards, rescuers can understand the situation of the disaster-stricken village based on the video content played by the terminal and make a rescue plan. If there are person A to be rescued and scene object B in the video played by the terminal, the rescuers cannot determine the specific locations of person A to be rescued and scene object B based on the video content played by the terminal.

[0044] In summary, in the current technical solutions, drones only collect images of the disaster site, and the amount of information collected is relatively small, which seriously affects the rescue personnel's judgment of the disaster scene and delays the rescue time.

[0045] In order to solve the above problems, an embodiment of the present application provides a method for generating a map model, and a drone can obtain an image of a disaster scene and a collection position (i.e., the position where the drone collects the image). Afterwards, the drone can generate a map model of the disaster scene based on the image and collection position of the disaster scene. In addition, the drone can determine the location information of the object to be rescued in the image of the disaster scene. In the case of generating a map model of the disaster scene and determining the location information of the object to be rescued in the image of the disaster scene, the drone can mark the location information of the object to be rescued on the map model of the disaster scene. In this way, rescuers can judge the disaster scene and make a rescue plan based on the marked map model sent back to the terminal by the drone, thereby improving the accuracy of the rescuers' judgment of the disaster scene. In addition, compared with video transmission, the amount of data of the marked map model sent back to the terminal by the drone is small, and the transmission speed is fast, which can speed up the rescue progress.

[0046] The implementation environment of the embodiments of the present application is introduced below.

[0047] Figure 1 A schematic diagram of a communication system provided in an embodiment of the present application, such as Figure 1 As shown, the communication system may include: a terminal 01 and a terminal 02. The terminal 01 may obtain acquisition information (such as an image and an acquisition position (ie, the position where the terminal 01 acquires the image)), and the terminal 01 and the terminal 02 may perform wireless communication.

[0048] For example, terminal 01 may communicate with terminal 02 via satellite communication. For another example, terminal 01 may communicate with terminal 02 via spread spectrum microwave communication. For another example, terminal 01 may communicate with terminal 02 via digital radio communication. For another example, terminal 01 may communicate with terminal 02 via Bluetooth.

[0049] The terminal 01 may process the collected information, such as generating a map model through the image and the collected location. The terminal 01 may send the collected information to the terminal 02. The terminal 02 may display the collected information.

[0050] In some embodiments, the terminal 01 may store data and process the data.

[0051] In other embodiments, the terminal 01 may be connected to a server. The terminal 01 may send the collected information to the server, and the server may store and process the collected information. Afterwards, the server may send the processed collected information to the terminal 01.

[0052] The terminal (such as terminal 01 or terminal 02) can be a mobile phone, tablet computer, desktop, laptop, handheld computer, notebook computer, ultra-mobile personal computer (UMPC), netbook, drone and other devices that can collect images and locate, and the present application does not impose any special restrictions on the specific form of the terminal. It can interact with the user through one or more methods such as keyboard, touchpad, touch screen, remote control, voice interaction or handwriting device.

[0053] After introducing the application scenario and implementation environment of the embodiment of the present application, the method for generating a map model provided by the embodiment of the present application is described in detail below in combination with the above-mentioned implementation environment.

[0054] The following takes terminal 01 as a drone as an example to specifically explain the technical solution provided in the embodiment of the present application.

[0055] Figure 2 FIG. 1 is a flow chart of a method for generating a map model according to an exemplary embodiment. Figure 2 As shown, the method may include S201-S205.

[0056] S201. The drone acquires a plurality of first images.

[0057] The multiple first images include objects to be rescued and scene objects.

[0058] It should be noted that the embodiments of the present application do not limit the first rescue object. For example, the first rescue object may be a person with red hair squatting on the ground. For another example, the first rescue object may be a person with white hair standing against a wall. For another example, the first rescue object may be an animal (such as a cat, dog, pig, etc.). For another example, the first rescue object may be an airplane, a ship, a car, etc. The embodiments of the present application do not limit the scene object. For example, the scene object may be a gray and broken bridge. For another example, the scene object may be a white and complete dome building. For another example, the scene object may be a green and broken tree.

[0059] In a possible implementation, the drone may capture a target video, which is a video of a disaster scene. Afterwards, the drone may obtain each frame of the target video. The plurality of first images may be all or part of the image frames in the target video.

[0060] Exemplarily, the drone obtains video A through acquisition, and video A includes image frame A, image frame B, and image frame C. The drone may use image frame A, image frame B, and image frame C as the first image.

[0061] In another possible implementation, the drone may collect images through a camera to obtain the first image.

[0062] Exemplarily, the drone obtains image A through acquisition and uses image A as the first image B.

[0063] It is understandable that the drone can obtain the first image in different ways, so that the ways for the drone to obtain the first image can be increased, thereby improving operability and flexibility.

[0064] In some embodiments, before the drone acquires the first image, the drone may acquire multiple second images, where the multiple second images are all the acquired images. Afterwards, the drone determines multiple first images from the multiple second images according to the first position corresponding to the target image and the multiple first positions. The distance between the first positions corresponding to the multiple first images and the first position corresponding to the target image is less than a preset distance threshold. The target image is any second image.

[0065] It should be noted that the introduction of multiple first positions can refer to the description in the following embodiments, which will not be repeated here. In addition, the embodiments of the present application do not limit the preset distance threshold. For example, the preset distance threshold may be a height difference of 1 meter and a horizontal distance of 2 meters. For another example, the preset distance threshold may be a height difference of 0.5 meters and a horizontal distance of 1.5 meters. For another example, the preset distance threshold may be a height difference of 2 meters and a horizontal distance of 3 meters.

[0066] For example, if the drone determines the target image A and sets the preset distance thresholds to a height difference of 0.5 meters and a horizontal distance of 1.2 meters. If the height difference between image A and target image A is 0.7 meters and the horizontal distance is 1.3 meters, the position of image A and the position of target image A are greater than the preset distance threshold, and image A is determined not to be the first image. If the height difference between image B and target image A is 0.5 meters and the horizontal distance is 1.2 meters, the position of image B and the position of target image A are equal to the preset distance threshold, and image B is determined not to be the first image. If the height difference between image C and target image A is 0.4 meters and the horizontal distance is 1.1 meters, the position of image C and the position of target image A are less than the preset distance threshold, and image C is determined to be the first image.

[0067] S202: The drone obtains multiple first positions.

[0068] In the embodiment of the present application, the first position is the position where the drone collects the first image, and multiple first positions correspond to multiple first images.

[0069] Exemplarily, the multiple first images include: image A, image B and image C, and the multiple first positions include: position A, position B and position C. Position A is the position where the drone collects image A, position B is the position where the drone collects image B, and position C is the position where the drone collects image C.

[0070] In a possible implementation manner, the drone may obtain the first position by locating the position where the first image is collected.

[0071] It should be noted that the embodiments of the present application do not limit the positioning method. For example, the drone can determine the first position corresponding to the first image through the Global Positioning System (GPS). For another example, the drone can determine the first position corresponding to the first image through the BeiDou Navigation Satellite System (BDS). For another example, the drone can determine the first position corresponding to the first image through the Global Navigation Satellite System (GLONASS).

[0072] In one possible design, the first position may include at least one of the following: latitude and longitude information of the drone, altitude information of the drone, and heading angle of the drone.

[0073] For example, when acquiring image A and image B, the drone determines, based on the BDS, that the location where image A was acquired is 36°N, 110°E, 900m high, and 72° heading, and that the location where image B was acquired is 47°N, 99°E, and 850m high. That is, the first location corresponding to image A is 36°N, 110°E, 900m high, and 72° heading, and the first location corresponding to image B is 47°N, 99°E, and 850m high.

[0074] It should be noted that the embodiment of the present application does not limit the minimum unit of the first position. For example, the longitude and latitude of the first position can be accurate to degrees, such as the first position is 35° north latitude, 130° east longitude, and 720 meters in height. For another example, the longitude and latitude of the first position can be accurate to minutes, such as the first position is 31°22′ north latitude, 121°43′ east longitude, and 900 meters in height. For another example, the height of the first position can be accurate to millimeters, such as the first position is 36° north latitude, 110° east longitude, and 900.005 meters in height.

[0075] S203: The drone generates a first map model according to the multiple first images and the multiple first positions.

[0076] Among them, the first map model is used to reflect the disaster scene.

[0077] In one possible design, the first map model includes a scene object.

[0078] Exemplarily, the first map model includes a scene object A and a scene object B. Scene object A is a black and broken bridge, and scene object B is a red and complete spire building.

[0079] In some embodiments, Figure 3 As shown, in the method for generating the map model, S203 may include S301-S302.

[0080] S301. The drone generates multiple second positions according to multiple first positions and multiple first images.

[0081] In an embodiment of the present application, a reprojection error of the first image acquired by the drone at the second position is smaller than a reprojection error of the first image acquired at the first position. The plurality of second positions corresponds to the plurality of first images.

[0082] It should be noted that the reprojection error refers to the difference between the projection of the real 3D space point on the image plane (that is, the pixel point on the image) and the reprojection (the virtual pixel point obtained). In other words, the difference between the map model generated by the image collected at the second position and the actual scene is smaller than the difference between the map model generated by the image collected at the first position and the actual scene. That is, the map model generated by the image collected at the second position has a higher accuracy.

[0083] Exemplarily, the plurality of first images include: image A, image B, and image C, the plurality of first positions include: first position A, first position B, and first position C, and the plurality of second positions include: second position A, second position B, and second position C. The second position A corresponds to image A, the second position B corresponds to image B, and the second position C corresponds to image C. Moreover, the reprojection error of image A collected by the drone at the second position A is smaller than the reprojection error of image A collected at the first position A, the reprojection error of image B collected by the drone at the second position B is smaller than the reprojection error of image B collected at the first position B, and the reprojection error of image C collected by the drone at the second position C is smaller than the reprojection error of image C collected at the first position C.

[0084] In a possible implementation, the drone may process the plurality of first positions and the plurality of first images by using an aerial triangulation optimization algorithm to generate a plurality of second positions, wherein the aerial triangulation optimization algorithm is used to determine the second positions corresponding to the first images.

[0085] It should be noted that the first position corresponding to the first image and the corresponding second position may be the same position, or two different positions.

[0086] For example, if the first position corresponding to the first image A is 27.2°N, 113.43°E, and 740 meters in altitude, and the first position corresponding to the first image B is 31°N, 105°E, and 330 meters in altitude, then the drone uses the aerial triangulation optimization algorithm to determine that the second position corresponding to the first image A is 27°2′N, 113°43′E, and 740 meters in altitude, and determines that the second position corresponding to the first image B is 31°N, 105°E, and 330 meters in altitude.

[0087] S302: The drone generates a first map model according to the multiple second positions and the multiple first images.

[0088] In a possible implementation, the drone can generate a dense point cloud based on multiple first positions and multiple first images. For an introduction to the dense point cloud, reference can be made to the description in the following embodiments, which will not be repeated here. Afterwards, the drone can process the dense point cloud through multiple second positions to generate a first map model.

[0089] In another possible implementation, the drone may generate a triangulated network model according to the multiple second positions and the multiple first images, and construct the first map model through the triangulated network model.

[0090] It is understandable that the drone can generate multiple second positions based on multiple first positions and multiple first images, the reprojection error of the first image collected at the second position is smaller than the reprojection error of the first image collected at the first position, and the multiple second positions correspond to the multiple first images. In other words, compared with the dense point cloud generated based on the first position and the first image corresponding to the first position, the dense point cloud generated by the second position and the first image corresponding to the second position is more accurate. Afterwards, the drone generates a first map model based on the generated multiple second positions and multiple first images. Since the reprojection error of the first image collected by the drone at the second position is smaller than the reprojection error of the first image collected at the first position. In this way, the accuracy of the first map model can be improved.

[0091] S204: The drone obtains the location information of the first object to be rescued.

[0092] In a possible implementation, the drone can determine the position information of the first object to be rescued based on multiple first positions and multiple target position relationships, where the target position relationship is the relationship between the first position and the position of the first object to be rescued.

[0093] In a possible design, the target position relationship may include at least one of the following: the directional position between the first position and the position of the first object to be rescued, the horizontal distance between the first position and the position of the first object to be rescued, and the height difference between the first position and the position of the first object to be rescued.

[0094] For example, the drone can determine that the first object A to be rescued is located 30° east of the first position A by the south, 50 meters horizontally. Alternatively, the drone can determine that the first object A to be rescued is located 45° west of the first position A by the south, 40 meters horizontally, and 15 meters high. Alternatively, the drone can determine that the first object A to be rescued is located 60° east of the first position A by the north, 60 meters horizontally, and 10 meters low.

[0095] Optionally, the drone can determine the target position relationship in a preset manner. For example, the drone can determine that the first position A is located 30° east of the first object A to be rescued and 50 meters away horizontally through an ultrasonic sensor. For another example, the drone can determine that the first position A is located 45° west of the first object A to be rescued and 40 meters away horizontally through a photoelectric sensor. For another example, the drone can determine that the first position A is located 60° east of the first object A to be rescued and 60 meters away horizontally through a laser rangefinder.

[0096] The following introduces how the drone determines the position information of the first object to be rescued based on the relationship between multiple first positions and multiple target positions.

[0097] For example, Figure 4As shown in the figure, there are three main coordinate systems, namely the image coordinate system with O1 as the origin, the camera coordinate system with O2 as the origin, and the world coordinate system with O3 as the origin. Among them, the image coordinate system is parallel to the camera coordinate system, and O1O2 is perpendicular to the camera coordinate system. The origin of the world coordinate system is the projection point of the camera on the horizontal plane, and the Y axis points to the front of the camera. The height H is used to represent the camera height (i.e., the height of the drone), the angle α is used to represent the pitch angle of the camera, the measurement point Q is used to represent the first object to be rescued, the projection point P is used to represent the projection position of the measurement point Q on the Y axis of the world coordinate system, the pixel point coordinate point Q1 (u, v) is used to represent the position of the measurement point Q in the image coordinate system, the pixel point coordinate point P1 (u0, v) is used to represent the position of the projection point P in the image coordinate system, the coordinate point O3 is used to represent the projection position of the drone in the world coordinate system, the coordinate point O1 (u0, v0) is used to represent the position of the lens midpoint in the image coordinate system, and the coordinate point M is used to represent the position of the center of the image coordinate system (i.e., the origin O1) on the Y axis of the world coordinate system. Among them, dx in the image coordinate system represents the actual length of the unit pixel of the u axis, dy represents the actual length of the unit pixel of the v axis, the actual length dx and dy represented by the unit pixel are in millimeters per pixel (mm / pix), and the focal length f represents the distance between the origin O1 and the origin O2, and the focal length f is in millimeters (mm). In the figure, the relationship between O3M and the pitch angle α can be expressed as H=O3M·tanα. Then the distance O3P from the measurement point Q to the coordinate point O3 in the Y-axis direction can be calculated. The derivation process can refer to Formula 1, Formula 2 and Formula 3.

[0098]

[0099] β=α-γ Formula 2.

[0100]

[0101] The O3P and ΔPO2Q~ΔP1O2Q1 obtained by the above formula (i.e. ), then the distance PQ from the measurement point Q to the coordinate point O3 in the X-axis direction can be calculated. The derivation process can refer to Formula 4, Formula 5 and Formula 6.

[0102]

[0103]

[0104]

[0105] Then the distance from Q to the origin of the world coordinate system is: That is, the distance between the first object to be rescued and the position of the drone on the actual ground. Afterwards, the drone can convert the longitude and latitude information of the first position of the first image into projection coordinates, calculate the coordinates of the first object to be rescued in the projection coordinate system based on the known heading angle of the drone at the first position of the first image and the distance between the drone and the first object to be rescued in the first image, and then convert the coordinates of the first object to be rescued into longitude and latitude coordinates to obtain the position information of the first object to be rescued. The derivation process can refer to Formula 7, Formula 8, Formula 9, Formula 10 and Formula 11.

[0106]

[0107]

[0108]

[0109]

[0110] L=I+L0 Formula 11.

[0111] S205: The drone processes the first map model by using the location information of the first object to be rescued to generate a second map model.

[0112] The second map model includes the location information of the first object to be rescued, the first object to be rescued and the scene object.

[0113] In a possible implementation, the drone may mark the first object to be rescued on the first map model according to the location information of the first object to be rescued, and generate the second map model.

[0114] For example, Figure 5 As shown, if the position information A of the first object to be rescued A corresponds to the position A of the first map model, the position A of the first map model is marked. If the position information B of the first object to be rescued B corresponds to the position B of the first map model, the position B of the first map model is marked. If the position information C of the first object to be rescued C corresponds to the position C of the first map model, the position C of the first map model is marked.

[0115] The technical solution provided by the above embodiment brings at least the following beneficial effects: the generating device can obtain multiple first images and multiple first positions, the multiple first images include the first object to be rescued and the scene object, the first position is the position of collecting the first image, and the multiple first positions correspond to the multiple first images. Afterwards, the generating device can generate a first map model according to the multiple first images and the multiple first positions, the first map model is used to reflect the disaster scene, and the first map model includes the scene object. In addition, the generating device can obtain the position information of the first object to be rescued, and process the first map model according to the position information of the first object to be rescued to generate a second map model, the second map model includes the position information of the first object to be rescued, the first object to be rescued and the scene object. That is to say, the generating device can mark the position information of the first object to be rescued and the objects in the disaster scene (such as the first object to be rescued and the scene object) on the second map model. In this way, the amount of information in the map model can be increased, and then it can provide valuable reference for rescue personnel to reasonably allocate rescue forces, determine key disaster relief areas, select safe rescue routes, and select sites for post-disaster reconstruction. In addition, compared with the video of the disaster scene, in the technical solution of the present application, the amount of data of the second map model is small. In this way, the amount of data transmitted by the generating device can be reduced, saving the time for the generating device to transmit information to the terminal, thereby speeding up the rescue progress and improving the rescue efficiency.

[0116] In some embodiments, Figure 6 As shown, in the method for generating the map model, S302 may include S601-S603.

[0117] S601. The drone processes the plurality of second positions and the plurality of first images through a target preset algorithm to determine a dense point cloud.

[0118] In a possible implementation, a target preset algorithm is stored in the drone, and the target preset algorithm is used to determine a coordinate set corresponding to the pixel points of the multiple first images. The drone can determine the coordinate set corresponding to the pixel points of the multiple first images according to the target preset algorithm. Afterwards, the drone uses the multiple second positions as reference points and determines a dense point cloud according to the coordinate set corresponding to the pixel points of the multiple first images. The dense point cloud is determined by the coordinate set corresponding to the pixel points of the multiple first images.

[0119] It should be noted that the embodiment of the present application does not limit the target preset algorithm.

[0120] Exemplarily, the target preset algorithm includes a stereo pair matching algorithm and a forward intersection algorithm. If the first image A corresponds to the first position A, the first image B corresponds to the first position B, and the first image A is adjacent to the first image B, then the stereo matching image pair algorithm is used to determine the same-name pixel point set A corresponding to the first image A and the first image B. Afterwards, the three-dimensional coordinates of the pixel points in the same-name pixel point set A are determined by the forward intersection algorithm to obtain the three-dimensional coordinate set B. By marking the three-dimensional coordinate set B in the three-dimensional coordinate system, the dense point cloud A is determined.

[0121] In some embodiments, a dense point cloud is obtained by downsampling each image by a preset image downsampling coefficient and specifying multiple point cloud sampling density levels. The point cloud sampling density levels include high density level, medium density level and low density level. Except for the high density level, the full pixel depth map is generated for the image, and the other depth maps are generated by interval pixels, that is, at the medium density level, the depth map is generated for every other pixel in the horizontal and vertical directions of the image; at the low density level, the depth map is generated for every two pixels in the horizontal and vertical directions of the image.

[0122] Exemplarily, in the case of determining a dense point cloud for the first image A, if the preset image downsampling coefficient is 3, and the point cloud sampling density level is specified as a high density level, a medium density level, and a low density level, then in the first image A, a pixel is taken every 3 pixels in each row and column to form the first image B. If the point cloud sampling density level of the first image B is a high density level, a full-pixel depth map is generated for the first image B. If the point cloud sampling density level of the first image B is a medium density level, a depth map is generated for every other pixel in the horizontal and vertical directions of the first image B. If the point cloud sampling density level of the first image B is a low density level, a depth map is generated for every two pixels in the horizontal and vertical directions of the first image B.

[0123] It is understandable that by specifying the downsampling factor and the point cloud sampling density, the number of three-dimensional coordinates of the determined pixel points can be reduced, thereby reducing the time for determining a dense point cloud.

[0124] S602. The UAV processes the dense point cloud to generate a triangulated network model.

[0125] It should be noted that the process of processing dense point clouds by drones to generate triangulated network models can refer to the conventional technology of generating triangulated network models through dense point clouds, which will not be elaborated here.

[0126] Exemplarily, the drone can process the dense point cloud by using the Delaunay tetrahedron partitioning algorithm, generate a Delaunay tetrahedron spatial partitioning, and construct a global optimization graph. The nodes of the global optimization graph are constructed by tetrahedrons in the spatial partitioning, and the edges of the global optimization graph are the triangular faces of adjacent tetrahedrons. Then, the drone can determine the triangular faces in the Delaunay tetrahedron partitioning that intersect the lines connecting each point to the camera it sees, and add a weight value of 1 to the corresponding edges in the global optimization graph to obtain a global optimization graph with visible line constraints. Finally, the drone can segment the global optimization graph by using the minimum cut maximum flow algorithm, determine the internal and external relationship between each tetrahedron and the model surface, extract the shared triangular faces of adjacent tetrahedrons located inside and outside the surface, and form the final triangulated network model.

[0127] S603: The drone renders the triangulated network model using the multiple first images to generate a first map model.

[0128] For example, the drone can select the visible camera closest to the triangular face of each triangulated mesh model as its associated camera. Then, the drone divides the spatially connected triangular faces of the same associated camera into the same group and obtains the image blocks collected on the associated camera. Finally, the drone combines all the image blocks into a texture map according to the packing algorithm, realizes the texture mapping of the triangulated mesh model, and obtains the first map model.

[0129] It is understandable that the drone can process multiple second positions and multiple first images through a target preset algorithm to determine a dense point cloud. The target preset algorithm is used to determine a set of coordinates corresponding to the pixel points of multiple first images. The dense point cloud is determined by a set of coordinates corresponding to the pixel points of multiple first images. And compared with the dense point cloud generated according to the first position and the first image corresponding to the first position, the dense point cloud generated by the second position and the first image corresponding to the second position is more accurate. Afterwards, the drone processes the dense point cloud to generate a triangulated network model, and renders the triangulated network model through multiple first images to generate a first map model. In this way, the accuracy of the first map model can be improved.

[0130] In some embodiments, the drone may determine a characteristic object in each of the plurality of first images. Then, the drone may model the characteristic object based on the first images having the same characteristic object.

[0131] like Figure 7 As shown, after executing S201, the map model generation method may further include S701-S702.

[0132] S701. The drone extracts M feature objects corresponding to a plurality of first images using a first preset algorithm.

[0133] The feature objects include a first object to be rescued and a scene object, and the plurality of first images correspond to M feature objects, where M is a positive integer.

[0134] Exemplarily, 3 (i.e., M is 3) feature objects include: feature object A, feature object B, and feature object C. Feature object A can be the first object to be rescued A (such as a person), feature object B can be scene object A (such as a river), and feature object C can be scene object B (such as a house).

[0135] It should be noted that the number of the plurality of first images may be the same as or different from M, and this is not limited in the present embodiment. Generally, the number of the plurality of first images is greater than M. That is, the feature objects of different first images may be the same.

[0136] Exemplarily, the plurality of first images include: image A, image B, image C, and image D, image A includes feature object A, image B includes feature object B, and image C and image D include the same feature object C.

[0137] Optionally, a first image may include multiple feature objects. For example, the first image A includes: a first object to be rescued A and a scene object A.

[0138] In the embodiment of the present application, the first preset algorithm is used to extract feature objects in multiple first images.

[0139] It should be noted that the first preset algorithm is not limited in the embodiments of the present application. For example, the first preset algorithm may be a Scale Invariant Feature Transform (SIFT) algorithm. For another example, the first preset algorithm may be a Speeded Up Robust Features (SURF) algorithm. For another example, the first preset algorithm may be an Oriented Fast and Rotated Brief (ORB) algorithm.

[0140] It should be noted that the embodiment of the present application does not limit the feature object. For example, the feature object may be a color feature. For another example, the feature object may be a texture feature. For another example, the feature object may be a shape feature.

[0141] S702: The drone processes the multiple first images according to the M feature objects to obtain a set of M images.

[0142] The M image sets correspond to the M feature objects. An image set includes at least one first image, and the feature objects corresponding to the first images in an image set are the same.

[0143] In a possible implementation, the drone may match multiple first images according to feature objects of the first images to obtain a set of images having the same feature objects.

[0144] Optionally, the drone may set a preset pixel size range. If the pixel size of the feature object is within the preset pixel size range, multiple first images with the same feature object are matched. If the pixel size of the feature object is not within the preset pixel size range, the matching of multiple first images with the same feature object is cancelled.

[0145] Exemplarily, the drone sets a preset pixel size range of (2000, 4000). If the pixel size of feature object A is 3000, multiple first images with the same feature object A are matched. If the pixel size of feature object B is 5000, the matching of multiple first images with the same feature object B is cancelled.

[0146] In some embodiments, after the drone matches multiple first images with the same characteristic objects, the drone may model the M characteristic objects according to the M image sets.

[0147] It is understandable that the drone matches multiple first images according to the feature objects extracted from the multiple first images, and can determine multiple first images with the same feature objects. In this way, the same feature objects in the multiple first images can be modeled by using the multiple first images with the same feature objects. In this way, the feature objects can be modeled more accurately. In addition, by setting a preset pixel size range, the feature objects of the first images can be screened. In this way, the matching time of multiple first images with the same feature objects can be greatly reduced.

[0148] The following is an introduction to the method for generating a map model provided by this application with reference to specific examples. Figure 8 As shown in FIG. 1 , taking the establishment of a map model of feature object A as an example, the image set corresponding to feature object A is image set A (i.e., the feature objects of the first image in image set A are all feature object A). The drone can process the second position corresponding to each first image in image set A and image set A through the target preset algorithm to obtain the following: Figure 8 The dense point cloud A of the feature object A shown in (a) in the figure. The drone can generate Figure 8 The triangulated network model A of the feature object A shown in (b) in the figure. The drone renders the triangulated network model A through the image set A to generate Figure 8 Map model A of feature object A shown in (c).

[0149] In some embodiments, Fig. 9 As shown, before S204, the map model generation method may further include S901-S902.

[0150] S901. The UAV obtains the second rescue object.

[0151] In the embodiment of the present application, the plurality of first images include the second object to be rescued, and the second object to be rescued includes the first object to be rescued.

[0152] Exemplarily, the first image A includes a second rescue object A and a second rescue object B. The second rescue object A includes the first rescue object A, the first rescue object B and the first rescue object C, and the second rescue object B includes the first rescue object D and the first rescue object E.

[0153] In one possible design, the drone stores a target detection algorithm, which is used to identify target objects (i.e., second objects to be rescued and scene objects) in multiple first images and extract features from the target objects in the multiple first images. The drone extracts features from the target objects in the multiple first images using the target detection algorithm, and configures a unique identity document (ID) for the target objects with the features.

[0154] It should be noted that the embodiments of the present application do not limit the target detection algorithm. For example, the R-CNN (Region-CNN, region segmentation technology) algorithm based on the candidate region (Region Proposal) (such as R-CNN, Fast R-CNN, Faster R-CNN, etc.). Another example is the one-stage (one stage / step) algorithm (such as Yolo, SSD, etc.).

[0155] For example, if the first image A includes people and buildings, the multi-target tracking algorithm is used to identify the object to be rescued in the first image A and extract features such as hair color and body shape, and to identify the scene object in the first image A and extract features such as color and shape of the scene object. Afterwards, the person with yellow hair and squatting on the ground in the first image A is assigned an ID as the object to be rescued A, the person with white hair and standing against the wall in the first image A is assigned an ID as the object to be rescued B, the white and complete dome building in the first image A is assigned an ID as the scene object A, and the red and broken bridge in the first image A is assigned an ID as the scene object B.

[0156] It is understandable that by identifying the target objects in the multiple first images, the rescuers can formulate a rescue plan based on the rescue needs of the second rescue objects in the disaster scene and the damage of the scene objects. In this way, the rescue progress is accelerated while the safety of the rescuers is guaranteed.

[0157] In a possible implementation, the drone may process multiple first images collected in real time.

[0158] Exemplarily, the drone collects the first image A at time A and processes the first image A. Alternatively, the drone collects the first image B and the first image C at time B and processes the first image B and the first image C. Alternatively, the drone collects the first image D, the first image E, and the first image F at time C and processes the first image D, the first image E, and the first image F.

[0159] In another possible implementation, the drone may process the multiple first images after completing acquisition of the multiple first images of the disaster scene.

[0160] Exemplarily, when the drone completes collecting the disaster scene, it collects the first image A, the first image B, and the first image C. Afterwards, the drone processes the first image A, the first image B, and the first image C.

[0161] It is understandable that the drone can process the first image in different ways, so that the drone can increase the processing methods for the first image, thereby improving operability and flexibility.

[0162] S902: The drone obtains the first object to be rescued by performing deduplication processing on the second object to be rescued.

[0163] In a possible implementation, the drone may remove duplicate objects in the second object to be rescued to obtain the first object to be rescued.

[0164] Exemplarily, the drone first determines multiple motion state definitions of the second object A to be rescued through the Kalman filter algorithm. The multiple motion state definitions change as the second object A moves. The position and speed of the previous frame of the second object A to be rescued will be used to predict the position and speed of the current frame. The two observed normal distribution states are linearly weighted to obtain the predicted state of the current frame. Then, the similarity of the two frames before and after the second object A to be rescued is calculated through the Hungarian algorithm to obtain the similarity matrix of the two frames before and after the second object A to be rescued. The overlap (Intersection over Union, IOU) of the two frames before and after is used to obtain the corresponding similarity matrix. By comparing the similarity matrices of the two frames before and after, the problem of actually matching the targets of the two frames before and after is solved.

[0165] like Fig.10 As shown, if the second object to be rescued with an ID configured includes a second object to be rescued A and a second object to be rescued B, predicting the second object to be rescued includes predicting the second object to be rescued A and predicting the second object to be rescued B. Among them, the second object to be rescued A corresponds to the predicted second object to be rescued A, and the second object to be rescued B corresponds to the predicted second object to be rescued B. Through IOU comparison calculation, the position and speed of the second object to be rescued A are matched with the predicted position and speed of the second object to be rescued A. If the position and speed of the second object to be rescued A successfully match the predicted position and speed of the second object to be rescued A, the position and speed of the second object to be rescued A are updated (that is, the predicted position and speed of the second object to be rescued A is used as the predicted position and speed of the second object to be rescued in the next frame), and enter the IOU comparison calculation. If the position and speed of the second object to be rescued A fail to match the predicted position and speed of the second object to be rescued A, the predicted position and speed of the second object to be rescued B are matched with the predicted position and speed of the second object to be rescued A. If the predicted position and speed of the second object to be rescued B fail to match the predicted position and speed of the second object to be rescued A, a new ID is configured for the predicted second object to be rescued A as the second object to be rescued C, and the position and speed of the second object to be rescued C in the next frame are predicted, and the IOU comparison calculation is entered. If the second object to be rescued A fails to match the predicted second object to be rescued A, and the predicted position and speed of the second object to be rescued B successfully match the predicted position and speed of the second object to be rescued A, the second object to be rescued A is deleted. In this way, the first object to be rescued is finally obtained.

[0166] It is understandable that the drone configures IDs for the target objects in the first image and performs deduplication processing, so that the target objects that need to be located can be reduced, speeding up the rescue progress.

[0167] The following is an introduction to the method for generating a map model provided by this application with reference to specific examples. Fig.11As shown, the system for generating a map model includes a drone camera and an image analysis module. The drone camera can capture original images (i.e., the first image and the first position). The image analysis module can detect and classify the targets (i.e., the first object to be rescued and the scene object) in the captured original image, perform deduplication processing on the classified targets, and locate the deduplication processed targets. Afterwards, the image analysis module displays the deduplication processed and located targets on the map model (i.e., the first map model).

[0168] The above mainly introduces the solution provided by the embodiment of the present application from the perspective of computer equipment. It can be understood that in order to realize the above functions, the computer equipment includes hardware structures and / or software modules corresponding to the execution of each function. Those skilled in the art should easily realize that, in combination with the map model generation method of each example described in the embodiment disclosed in this application, the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0169] The embodiment of the present application also provides a map model generation device. The map model generation device can be a computer device, a CPU in the above-mentioned computer device, a processing module in the above-mentioned computer device for determining the generation of the map model, or a client in the above-mentioned computer device for generating the map model.

[0170] The embodiment of the present application can divide the generation of the map model into functional modules or functional units according to the above method example. For example, each functional module or functional unit can be divided according to each function, or two or more functions can be integrated into one processing module. The above integrated module can be implemented in the form of hardware or in the form of software functional modules or functional units. Among them, the division of modules or units in the embodiment of the present application is schematic, which is only a logical functional division. There may be other division methods in actual implementation.

[0171] like Fig.12 FIG. 1 is a block diagram of a map model generation device provided in an embodiment of the present application. The map model generation device is used to execute Figure 2 The map model generation method shown in FIG. 1200 includes an acquisition module 1201 and a processing module 1202 .

[0172] The acquisition module 1201 is used to acquire multiple first images and multiple first positions, wherein the multiple first images include a first object to be rescued and a scene object, the first position is a position where the first image is collected, and the multiple first positions correspond to the multiple first images. The processing module 1202 is used to generate a first map model based on the multiple first images and the multiple first positions, wherein the first map model is used to reflect the disaster scene, and the first map model includes the scene object. The acquisition module 1201 is also used to acquire the position information of the first object to be rescued. The processing module 1202 is specifically used to process the first map model according to the position information of the first object to be rescued to generate a second map model, wherein the second map model includes the position information of the first object to be rescued, the first object to be rescued, and the scene object.

[0173] Optionally, the acquisition module 1201 is further used to acquire multiple second images, where the multiple second images are all the acquired images. The processing module 1202 is further used to determine multiple first images from the multiple second images according to the first position corresponding to the target image and the multiple first positions, where the target image is any second image, and the distance between the first positions corresponding to the multiple first images and the first position corresponding to the target image is less than a preset distance threshold.

[0174] Optionally, the processing module 1202 is further used to generate a plurality of second positions according to the plurality of first positions and the plurality of first images, the reprojection error of the first image collected at the second position is less than the reprojection error of the first image collected at the first position, and the plurality of second positions correspond to the plurality of first images. The processing module 1202 is further used to generate a first map model according to the plurality of second positions and the plurality of first images.

[0175] Optionally, the processing module 1202 is further used to process the plurality of second positions and the plurality of first images through a target preset algorithm to determine a dense point cloud, the target preset algorithm is used to determine a coordinate set corresponding to the pixel points of the plurality of first images, and the dense point cloud is determined by the coordinate set corresponding to the pixel points of the plurality of first images. The processing module 1202 is further used to process the dense point cloud to generate a triangulated network model. The processing module 1202 is further used to render the triangulated network model through the plurality of first images to generate a first map model.

[0176] Optionally, the plurality of first images include the second object to be rescued, and the second object to be rescued includes the first object to be rescued. The processing module 1202 is further used to perform deduplication processing on the second object to be rescued to obtain the first object to be rescued. The processing module 1202 is further used to determine the position information of the first object to be rescued based on the plurality of first positions and the plurality of target position relationships, where the target position relationship is the relationship between the first position and the position of the first object to be rescued.

[0177] Fig.13Another possible structure of the map model generation device involved in the above-mentioned embodiments is shown. The map model generation device includes: a processor 1301 and a communication interface 1302. The processor 1301 is used to control and manage the actions of the device, for example, to execute the various steps in the method flow shown in the above-mentioned method embodiment, and / or to execute other processes of the technology described in this article. The communication interface 1302 is used to support the communication between the map model generation device and other network entities. The map model generation device may also include a memory 1303 and a bus 1304, and the memory 1303 is used to store the program code and data of the device.

[0178] Among them, the above-mentioned processor 1301 can implement or execute various exemplary logic blocks, units and circuits described in conjunction with the disclosure of the present invention. The processor can be a central processing unit, a general-purpose processor, a digital signal processor, an application-specific integrated circuit, a field programmable gate array or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. It can implement or execute various exemplary logic blocks, units and circuits described in conjunction with the disclosure of the present invention. The processor can also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a digital signal processor (DSP) and a microprocessor, etc.

[0179] The memory 1303 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a read-only memory, a flash memory, a hard disk or a solid-state drive; the memory may also include a combination of the above types of memory.

[0180] The bus 1304 may be an Extended Industry Standard Architecture (EISA) bus, etc. The bus 1304 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Fig.13 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0181] In actual implementation, the acquisition module 1201 can be composed of Fig.13 The communication interface 1302 shown in FIG. 1 is implemented by the processing module 1202. Fig.13 The processor 1301 shown calls the program code in the memory 1303 to implement. The specific execution process can be referred to Figure 2 The description of the generation method of the map model shown will not be repeated here.

[0182] Through the description of the above implementation methods, technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. The specific working process of the system, device and unit described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0183] An embodiment of the present application further provides a computer-readable storage medium, in which instructions are stored. When the instructions are executed on a computer, the computer executes the method for generating a map model in the method flow shown in the above method embodiment.

[0184] Among them, the computer readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer readable storage media (a non-exhaustive list) include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a register, a hard disk, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above, or any other form of computer readable storage medium known in the art. An exemplary storage medium is coupled to a processor so that the processor can read information from the storage medium and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and the storage medium can be located in an application-specific integrated circuit (ASIC). In the embodiments of the present application, a computer-readable storage medium may be any tangible medium that contains or stores a program, which may be used by or in conjunction with an instruction execution system, apparatus, or device.

[0185] Fig.14 A conceptual partial view of a computer program product provided by an embodiment of the present application is schematically shown, where the computer program product includes a computer program for executing a computer process on a computing device.

[0186] In one embodiment, the computer program product is provided using a signal bearing medium 1400. The signal bearing medium 1400 may include one or more program instructions that, when executed by one or more processors, may provide the above-described Figure 2 Thus, for example, reference to Figure 2 In the embodiment shown in , one or more features of S201 to S205 may be undertaken by one or more instructions associated with the signal bearing medium 1400. In addition, Fig.14 The program instructions in also describe example instructions.

[0187] In some examples, the signal bearing medium 1400 may include a computer readable medium 1401, such as, but not limited to, a hard drive, a compact disk (CD), a digital video disk (DVD), a digital tape, a memory, a read-only memory (ROM), or a random access memory (RAM), and the like.

[0188] In some implementations, the signal bearing medium 1400 may include a computer recordable medium 1402 such as, but not limited to, a memory, a read / write (R / W) CD, a R / W DVD, or the like.

[0189] In some implementations, signal bearing medium 1400 may include communication medium 1403 such as, but not limited to, digital and / or analog communication media (eg, fiber optic cables, waveguides, wired communication links, wireless communication links, etc.).

[0190] The signal bearing medium 1400 may be communicated by a wireless form of communication medium 1403. The one or more program instructions may be, for example, computer executable instructions or logic implemented instructions.

[0191] In some examples, such as for Fig.12 The described map model generation apparatus may be configured to provide various operations, functions, or actions in response to one or more program instructions via computer-readable medium 1401 , computer-recordable medium 1402 , and / or communication medium 1403 .

[0192] Since the map model generating device, computer-readable storage medium, and computer program product in the embodiments of the present invention can be applied to the above method, the technical effects that can be obtained can also refer to the above method embodiments, and the embodiments of the present invention will not be repeated here.

[0193] The above are only specific implementations of the present application, but the protection scope of the present application is not limited thereto, and any changes or substitutions within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for generating a map model, characterized in that: The method comprises: Acquire a plurality of first images and a plurality of first positions, wherein the plurality of first images include a first object to be rescued and a scene object, the first position is a position where the first image is collected, and the plurality of first positions correspond to the plurality of first images; Generate a plurality of second positions according to the plurality of first positions and the plurality of first images, wherein a reprojection error of the first image acquired at the second position is smaller than a reprojection error of the first image acquired at the first position, and the plurality of second positions correspond to the plurality of first images; generating a first map model according to the plurality of second positions and the plurality of first images, wherein the first map model is used to reflect the disaster scene, and the first map model includes the scene object; Obtaining the location information of the first object to be rescued; The first map model is processed according to the position information of the first object to be rescued to generate a second map model, wherein the second map model includes the position information of the first object to be rescued, the first object to be rescued and the scene object.

2. The method according to claim 1, characterized in that Before acquiring the plurality of first images, the method further includes: Acquire a plurality of second images, where the plurality of second images are all the acquired images; Acquiring the plurality of first images comprises: According to the first position corresponding to the target image and the multiple first positions, the multiple first images are determined from the multiple second images, the target image is any one of the second images, and the distance between the first position corresponding to the multiple first images and the first position corresponding to the target image is less than a preset distance threshold.

3. The method according to claim 1 or 2, characterized in that: The generating the first map model according to the plurality of second positions and the plurality of first images, the method further comprising: Processing the plurality of second positions and the plurality of first images by a target preset algorithm to determine a dense point cloud, wherein the target preset algorithm is used to determine a coordinate set corresponding to pixel points of the plurality of first images, and the dense point cloud is determined by the coordinate set corresponding to pixel points of the plurality of first images; Processing the dense point cloud to generate a triangulated network model; The triangulated network model is rendered using the multiple first images to generate the first map model.

4. The method according to claim 3, characterized in that The plurality of first images include a second object to be rescued, and the second object to be rescued includes the first object to be rescued; Before obtaining the location information of the first object to be rescued, the method further includes: performing deduplication processing on the second object to be rescued to obtain the first object to be rescued; The obtaining the location information of the first object to be rescued includes: The position information of the first object to be rescued is determined according to the multiple first positions and multiple target position relationships, wherein the target position relationship is the relationship between the first position and the position of the first object to be rescued.

5. A map model generation device, characterized in that: The device comprises: An acquisition module, configured to acquire a plurality of first images and a plurality of first positions, wherein the plurality of first images include a first object to be rescued and a scene object, the first position is a position where the first image is collected, and the plurality of first positions correspond to the plurality of first images; a processing module, configured to generate a plurality of second positions according to the plurality of first positions and the plurality of first images, wherein a reprojection error of the first image acquired at the second position is smaller than a reprojection error of the first image acquired at the first position, and the plurality of second positions correspond to the plurality of first images; The processing module is further used to generate a first map model according to the plurality of second positions and the plurality of first images, wherein the first map model is used to reflect the disaster scene, and the first map model includes the scene object; The acquisition module is further used to acquire the position information of the first object to be rescued; The processing module is further used to process the first map model according to the position information of the first object to be rescued to generate a second map model, wherein the second map model includes the position information of the first object to be rescued, the first object to be rescued and the scene object.

6. The device according to claim 5, characterized in that The acquisition module is further used to acquire a plurality of second images, wherein the plurality of second images are all the acquired images; The processing module is also used to determine the multiple first images from the multiple second images based on the first position corresponding to the target image and the multiple first positions, the target image being any of the second images, and the distance between the first position corresponding to the multiple first images and the first position corresponding to the target image is less than a preset distance threshold.

7. The device according to claim 5 or 6, characterized in that The processing module is further used to process the multiple second positions and the multiple first images through a target preset algorithm to determine a dense point cloud, wherein the target preset algorithm is used to determine a coordinate set corresponding to the pixel points of the multiple first images, and the dense point cloud is determined by the coordinate set corresponding to the pixel points of the multiple first images; The processing module is further used to process the dense point cloud to generate a triangulated network model; The processing module is further used to render the triangulated network model using the multiple first images to generate the first map model.

8. The device according to claim 7, characterized in that The plurality of first images include a second object to be rescued, and the second object to be rescued includes the first object to be rescued; The processing module is further used to perform a deduplication process on the second object to be rescued to obtain the first object to be rescued; The processing module is further used to determine the position information of the first object to be rescued according to the multiple first positions and multiple target position relationships, and the target position relationship is the relationship between the first position and the position of the first object to be rescued.

9. A map model generation device, characterized in that: include: Processor and memory; The processor is coupled to the memory; The memory is used to store one or more programs, which include computer executable instructions. When the map model generation device is running, the processor executes the computer executable instructions stored in the memory to enable the map model generation device to perform the map model generation method described in any one of claims 1 to 4.

10. A computer-readable storage medium, wherein instructions are stored in the computer-readable storage medium, characterized in that: When a computer executes the instruction, the computer executes the method for generating a map model as described in any one of claims 1 to 4.

11. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for generating a map model described in any one of claims 1 to 4 is implemented.

Images