Method and device for quickly generating fire rescue map, electronic equipment and medium

Abstract

The invention provides a fire rescue map rapid generation method and device, electronic equipment and a medium, and belongs to the field of map construction. The method comprises the following steps: acquiring scanning point cloud data of a rescue area, and performing Tai Chi algorithm disc sampling on the scanning point cloud data to obtain a point cloud sampling image; performing image graying and Gaussian blur processing on the point cloud sampling image to obtain a filtering point cloud image, performing image gradient calculation on the filtering point cloud image to obtain an initial edge image, and performing dual-threshold boundary selection on the initial edge image to obtain a rescue area map; and fusing the rescue area map and a preset global map to obtain a fire rescue map. According to the method, the processing efficiency of point cloud data is effectively improved through Tai Chi algorithm disc sampling, redundant data is reduced through image graying, data denoising is performed through Gaussian blur processing, the calculation efficiency is effectively improved, and rapid generation of a fire rescue map is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of map construction, and in particular to a method, device, electronic equipment and medium for quickly generating a fire rescue map. Background Art

[0002] In fire emergency rescue, in order to ensure the safety of personnel, chemical detection vehicles, drones, rescue robots and other equipment will be used to collect pollutants and detect and conduct rescue investigations in contaminated areas. However, during fire rescue, the existing topography is prone to change and damage, and secondary changes and damage to the topography may occur at any time during the on-site detection process. Existing emergency rescue usually relies on the images sent back by the robot camera combined with the images sent back by the drone for manual operation.

[0003] In order to enable fire emergency rescue, various vehicles and robots to perform automatic path planning, it is first necessary to reconstruct the map of the changed terrain in real time. However, in the existing map construction methods, for such destructive geographical environments, a large amount of image data calculation is often required, making it difficult to quickly generate new maps.

[0004] Therefore, the existing map construction method in the prior art has the technical problem that it is difficult to quickly generate a new map due to the large amount of image data calculation, and needs to be improved. Summary of the invention

[0005] In view of this, it is necessary to provide a method, device, electronic device and medium for quickly generating fire rescue maps to solve the technical problems existing in the prior art that the amount of image data calculation is large and it is difficult to quickly generate new maps.

[0006] In order to solve the above problems, on the one hand, the present invention provides a method for quickly generating a fire rescue map, comprising: Obtain the scanned point cloud data of the rescue area, and perform Tai Chi algorithm disk sampling on the scanned point cloud data to obtain a point cloud sampling image; The point cloud sampling image is grayed and Gaussian blurred to obtain a filtered point cloud image, the filtered point cloud image is image gradient calculated to obtain an initial edge image, and the initial edge image is double-thresholded to select the boundary to obtain a rescue area map; The rescue area map and the preset global map are merged to obtain the fire rescue map.

[0007] In a possible implementation, the scanning point cloud data is subjected to Tai Chi algorithm disk sampling to obtain a point cloud sampling image, including: Divide the point cloud space of the scanned point cloud data into a number of point cloud space grids; Selecting an initial sampling point, and determining a grid to be scanned at the initial sampling point according to the initial sampling point; Scan the point cloud data in the grid to be scanned to obtain a candidate point cloud, and determine the point distance between the candidate point cloud and the initial sampling point; Eliminate candidate point clouds whose point distance is less than or equal to the preset sampling radius, and use candidate point clouds whose point distance is greater than the preset sampling radius as candidate sampling points; The distances between the candidate sampling points and the existing sampling points of the grid to be scanned are judged in turn. If they are all greater than the preset sampling radius, the candidate sampling points are used as the new round of sampling points, and all the new round of sampling points are obtained by traversing the candidate sampling points; The sampling process is iterated until the scanning point cloud data is sampled and a point cloud sampling image is obtained.

[0008] In a possible implementation, graying and Gaussian blurring are performed on the point cloud sampling image to obtain a filtered point cloud image, including: Based on the preset grayscale weight, the point cloud sampling image is reduced in dimension and grayscaled to obtain a grayscale sampling image; The grayscale sampling image is denoised by Gaussian filtering to obtain a filtered point cloud image.

[0009] In a possible implementation, performing image gradient calculation on the filtered point cloud image to obtain an initial edge image includes: Calculate the X-direction gradient and Y-direction gradient of each pixel in the filtered point cloud image in the X-axis direction and the Y-axis direction respectively; According to the X-direction gradient and the Y-direction gradient, the gradient amplitude and the gradient direction of each pixel in the filtered point cloud image are determined to obtain an initial gradient image; The initial edge image is obtained by performing non-maximum suppression on the initial gradient image.

[0010] In a possible implementation, a double threshold boundary selection is performed on the initial edge image to obtain a rescue area map, including: Setting a first gradient threshold and a second gradient threshold, wherein the first gradient threshold is smaller than the second gradient threshold; Traversing the gradient values ​​of each gradient edge in the initial edge image, taking the gradient edge whose gradient value is greater than the second gradient threshold as a reliable edge, and taking the gradient edge whose gradient value is less than or equal to the second gradient threshold and greater than the first gradient threshold as a candidate edge; Determine the connectivity between the candidate edge and the reliable edge, and if the connectivity is greater than a preset connectivity threshold, merge the candidate edge into the reliable edge; A rescue area map is constructed based on reliable edges.

[0011] In a possible implementation, the rescue area map and the preset global map are merged to obtain a fire rescue map, including: Identify map elements on the preset global map and rescue area map to obtain map element attribute parameters; According to the map feature attribute parameters, the preset global map and the rescue area map are associated and matched, and then the map is fused to obtain the fire rescue map.

[0012] In a possible implementation, the method for quickly generating a fire rescue map further includes: Obtaining the road condition collection information of the robot to be positioned, generating a road information model based on the road condition collection information, generating a point cloud information model based on the scanned point cloud data, and matching the road information model with the point cloud information model to obtain a semantic positioning result; Generate a local positioning map based on the collected road condition information, and obtain a map positioning result by matching the local positioning map with the fire rescue map; The semantic localization results and map localization results are combined to obtain the robot localization results.

[0013] On the other hand, the present invention also provides a device for quickly generating a fire rescue map, comprising: A point cloud sampling unit is used to obtain scanned point cloud data of the rescue area and perform Tai Chi algorithm disk sampling on the scanned point cloud data to obtain a point cloud sampling image; A map generation unit is used to grayscale and Gaussian blur the point cloud sampling image to obtain a filtered point cloud image, perform image gradient calculation on the filtered point cloud image to obtain an initial edge image, and perform double threshold boundary selection on the initial edge image to obtain a rescue area map; The map fusion unit is used to fuse the rescue area map with the preset global map to obtain a fire rescue map.

[0014] On the other hand, the present invention also provides an electronic device, including a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the above-mentioned method for rapidly generating a fire rescue map is implemented.

[0015] On the other hand, the present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned method for quickly generating a fire rescue map is implemented.

[0016] The beneficial effects of the present invention are as follows: in the method for quickly generating a fire rescue map provided by the present invention, firstly, the scanned point cloud data of the rescue area is obtained, and the scanned point cloud data is subjected to Tai Chi algorithm disk sampling to obtain a point cloud sampling image; then, the point cloud sampling image is subjected to image graying and Gaussian blur processing to obtain a filtered point cloud image, and the filtered point cloud image is subjected to image gradient calculation to obtain an initial edge image, and the initial edge image is subjected to double threshold boundary selection to obtain a rescue area map; finally, the rescue area map and the preset global map are merged to obtain a fire rescue map. The present invention effectively speeds up the processing efficiency of point cloud data through Tai Chi algorithm disk sampling, reduces redundant data through image graying, and performs data denoising through Gaussian blur processing, thereby effectively improving the calculation efficiency and realizing the rapid generation of fire rescue maps. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0018] Figure 1 A schematic diagram of a flow chart of an embodiment of a method for quickly generating a fire rescue map provided by the present invention; Figure 2 A schematic diagram of the process of Tai Chi algorithm disk sampling according to an embodiment of the present invention; Figure 3 A schematic diagram of the grayscale and Gaussian blur process of an embodiment of the present invention; Figure 4 A schematic diagram of a flow chart of image gradient calculation according to an embodiment of the present invention; Figure 5 A schematic diagram of a process for selecting a dual threshold boundary according to an embodiment of the present invention; Figure 6 A schematic diagram of a process of map fusion according to an embodiment of the present invention; Figure 7 A schematic diagram of a process flow of robot positioning and matching according to an embodiment of the present invention; Figure 8 A schematic diagram of the structure of an embodiment of a device for quickly generating a fire rescue map provided by the present invention; Fig. 9 A schematic structural diagram of an embodiment of an electronic device provided by the present invention. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0020] In the description of the embodiments of the present invention, unless otherwise specified, "multiple" means two or more than two. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" may mean: A exists alone, A and B exist at the same time, and B exists alone.

[0021] The descriptions of "first" and "second" in the embodiments of the present invention are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the technical features defined as "first" and "second" may explicitly or implicitly include at least one of the features.

[0022] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0023] The present invention provides a method, device, electronic equipment and medium for quickly generating a fire rescue map, which are respectively described below.

[0024] Figure 1 A flowchart of an embodiment of the method for quickly generating a fire rescue map provided by the present invention is shown in FIG. Figure 1 As shown, a method for quickly generating a fire rescue map is provided. It includes: S101, obtaining scanning point cloud data of the rescue area, and performing Tai Chi algorithm disk sampling on the scanning point cloud data to obtain a point cloud sampling image; S102, performing image graying and Gaussian blur processing on the point cloud sampling image to obtain a filtered point cloud image, performing image gradient calculation on the filtered point cloud image to obtain an initial edge image, and performing double threshold boundary selection on the initial edge image to obtain a rescue area map; S103: The rescue area map and the preset global map are integrated to obtain a fire rescue map.

[0025] Compared with the prior art, in the method for quickly generating a fire rescue map provided by an embodiment of the present invention, firstly, the scanned point cloud data of the rescue area is obtained, and the scanned point cloud data is subjected to Tai Chi algorithm disk sampling to obtain a point cloud sampling image; then, the point cloud sampling image is subjected to image graying and Gaussian blur processing to obtain a filtered point cloud image, and the filtered point cloud image is subjected to image gradient calculation to obtain an initial edge image, and the initial edge image is subjected to double threshold boundary selection to obtain a rescue area map; finally, the rescue area map and the preset global map are merged to obtain a fire rescue map. The present invention effectively speeds up the processing efficiency of point cloud data through Tai Chi algorithm disk sampling, reduces redundant data through image graying, and performs data denoising through Gaussian blur processing, thereby effectively improving the computing efficiency and realizing the rapid generation of fire rescue maps.

[0026] In some embodiments of the present invention, Figure 2 FIG. 4 is a flow chart of the Tai Chi algorithm disk sampling process according to an embodiment of the present invention. Figure 2 As shown, the scanning point cloud data is sampled by Tai Chi algorithm disk to obtain a point cloud sampling image, including: S201, dividing the point cloud space of the scanned point cloud data into a plurality of point cloud space grids; S203, selecting an initial sampling point, and determining a grid to be scanned at the initial sampling point according to the initial sampling point; S203, scanning the point cloud data in the grid to be scanned to obtain a candidate point cloud, and determining the point distance between the candidate point cloud and the initial sampling point; S204, eliminating candidate point clouds whose point distances are less than or equal to a preset sampling radius, and taking candidate point clouds whose point distances are greater than the preset sampling radius as candidate sampling points; S205, determining the distances between the candidate sampling points and the existing sampling points of the to-be-scanned grid in turn, and if they are all greater than a preset sampling radius, taking the candidate sampling points as new round sampling points, and traversing the candidate sampling points to obtain all new round sampling points; S206 , iterate the sampling process until the scanning point cloud data is sampled to obtain a point cloud sampling image.

[0027] Specifically, in the embodiment, the scanning point cloud data is obtained by the laser radar, high-precision inertial navigation equipment, mapping camera, infrared telemetry detection camera and the camera of the robot, etc. The collected data is transmitted and processed in the form of point cloud.

[0028] Since the original scanned point cloud data is large, the scanned point cloud data needs to be sampled in order to reduce the data volume. During the sampling process, in order to ensure uniform distribution of the data, the embodiment performs Poisson disk sampling, which can ensure that the distance between any two points is not too close.

[0029] At the same time, in order to improve the efficiency of data processing during the sampling process, the embodiment samples the Tai Chi algorithm combined with Poisson disk sampling to speed up the processing speed. In order to avoid the need to check the distance between the sampling point and all existing points every time in Poisson disk sampling, the Tai Chi algorithm uses a gridding technique.

[0030] First, in the embodiment, the point cloud space of the point cloud data is divided into a number of square point cloud space grids, and then an initial point p is selected, and the square where it is located is marked as D. Then, any point whose distance from this point p is less than or equal to the preset distance threshold of Poisson disk sampling must be located in a square area composed of squares with a certain range of content around the square D as the center. Then, when checking the distance of the sampling point, it is only necessary to calculate for this square area.

[0031] During the sampling process, the embodiment first selects an initial sampling point, determines a to-be-scanned grid of a square area corresponding to the initial sampling point, and scans a plurality of point cloud data in the grid as candidate point clouds.

[0032] Then, the embodiment calculates the distance between each candidate point cloud and the initial sampling point, removes the points whose distance is less than the preset distance threshold, and uses the candidate point cloud whose distance is greater than the preset distance threshold as the candidate sampling point. Then, for each candidate sampling point, it is necessary to determine the point distance between the candidate sampling point and the existing sampling point in turn. When the point distance between a candidate sampling point and the existing sampling point in the grid to be scanned is greater than the preset distance threshold, it is used as a new round of sampling points, and all candidate sampling points are traversed to obtain all new round sampling points.

[0033] It should be noted that in the process of traversing candidate sampling points, when a candidate sampling point is used as a new round of sampling points, it is immediately added to the existing sampling points. At this time, the next candidate sampling point also needs to be judged with the newly added sampling point when judging the distance. At the same time, since the Tai Chi algorithm disk sampling process is carried out from the inside to the outside, it is similar to the growth process of cabbage: starting from a seed point, new points are added layer by layer. Each time we add a new point, it is located around the outermost point and covers the outermost layer as much as possible. Therefore, each judgment also only needs to be judged with the existing sampling points in the grid to be scanned.

[0034] It has been verified that when running this solution on a 400×400 grid area, the algorithm attempts to obtain 100k evenly distributed points, and actually generates about 53.7k. On a single CPU thread, excluding the compilation time, the time consumed by the present invention is only a little more than 0.7s, while the conventional method takes about 70s. The Tai Chi algorithm disk sampling of the present invention is about 100 times faster than the conventional method, and the efficiency is greatly improved.

[0035] In some embodiments of the present invention, Figure 3 FIG. 1 is a flow chart of grayscale and Gaussian blur processing according to an embodiment of the present invention. Figure 3 As shown, the point cloud sampling image is grayed and Gaussian blurred to obtain a filtered point cloud image, including: S301, reducing the dimension of the point cloud sampling image into grayscale based on a preset grayscale weight to obtain a grayscale sampling image; S302, performing Gaussian filtering and denoising on the grayscale sampling image to obtain a filtered point cloud image.

[0036] Specifically, for a picture, when only its boundary is concerned, a single-channel picture is sufficient to detect boundary information. Therefore, the embodiment further grayscales the RGB three-channel picture or the higher-dimensional hyperspectral remote sensing image. Grayscale is actually a dimensionality reduction operation, which can reduce redundant data and reduce the amount of calculation. The grayscale operation can select different preset grayscale weights to perform grayscale operations according to the actual scene requirements.

[0037] The grayscale process can be expressed as:

[0038] in, is the grayscale value of the sampling point, , and is the preset grayscale weight, , and is the value of each color channel at the sampling point.

[0039] After graying, the embodiment also uses Gaussian blur processing to denoise the grayed image. In the embodiment, Gaussian blur processing is performed by convolving the image with a normal distribution, discretizing the Gaussian function, and substituting the corresponding horizontal and vertical coordinate indexes in the filter into the Gaussian function to obtain the corresponding values. For filters of different sizes, the obtained values ​​are different. The embodiment uses a two-dimensional Gaussian function formula to express it as:

[0040]

[0041] Among them, the Gaussian filter size is , .

[0042] In some embodiments of the present invention, Figure 4 FIG. 1 is a flow chart of image gradient calculation according to an embodiment of the present invention. Figure 4 As shown, the image gradient calculation is performed on the filtered point cloud image to obtain the initial edge image, including: S401, calculating the X-direction gradient and the Y-direction gradient of each pixel in the filtered point cloud image in the X-axis direction and the Y-axis direction respectively; S402, determining the gradient amplitude and gradient direction of each pixel in the filtered point cloud image according to the X-direction gradient and the Y-direction gradient to obtain an initial gradient image; S403, performing non-maximum suppression on the initial gradient image to obtain an initial edge image.

[0043] Specifically, by calculating the difference between the current pixel and the adjacent pixel in the image, we can determine whether the pixel is inside the object or at the boundary. This pixel difference is called the image gradient. The calculation of the gradient includes calculating the gradient amplitude and gradient direction. The first-order derivative formula for calculating the gradient is expressed as:

[0044] For the above formula, the actual operation in the embodiment is to subtract the current pixel value from the next pixel value. .

[0045] For the convenience of calculation, the embodiment first calculates the X-direction gradient and the Y-direction gradient in the X-axis and Y-axis directions respectively. They are two vectors, and the gradient amplitude is the vector sum of the two vectors. The formula is expressed as:

[0046] in, represents the gradient amplitude, represents the gradient in the Y direction, Represents the gradient in the X direction.

[0047] The direction of the gradient amplitude is also determined according to the X-direction gradient and the Y-direction gradient, and the formula is expressed as:

[0048] In addition, ideally, the edge finally obtained should be very thin. Therefore, it is necessary to make the edge thinner by non-maximum suppression. In this process, the embodiment first traverses all points on the gradient matrix and retains pixels with maximum values ​​in the edge direction.

[0049] In some embodiments of the present invention, Figure 5 FIG. 1 is a schematic diagram of a process for selecting a dual threshold boundary according to an embodiment of the present invention. Figure 5 As shown, the initial edge image is subjected to double threshold boundary selection to obtain a rescue area map, including: S501, setting a first gradient threshold and a second gradient threshold, wherein the first gradient threshold is smaller than the second gradient threshold; S502, traversing the gradient values ​​of each gradient edge in the initial edge image, taking the gradient edges whose gradient values ​​are greater than the second gradient threshold as reliable edges, and taking the gradient edges whose gradient values ​​are less than or equal to the second gradient threshold and greater than the first gradient threshold as candidate edges; S503, determining the connectivity between the candidate edge and the reliable edge, and if the connectivity is greater than a preset connectivity threshold, merging the candidate edge into the reliable edge; S504: Obtain a rescue area map based on reliable edge construction.

[0050] Specifically, in order to accurately select edges during map construction, the embodiment sets two thresholds: a first gradient threshold and the second gradient threshold .

[0051] Then the embodiment first traverses the gradient values ​​of all gradient edges in the initial gradient image and compares them with the two thresholds. , then the edge is considered a reliable edge, and the gradient value is less than or equal to the first gradient threshold , then the edge is definitely a non-edge and is therefore discarded.

[0052] For edges whose gradient values ​​are between these two thresholds, the embodiment calculates the connectivity between these edges and the reliable edges. If the connectivity between a certain edge and the reliable edge is calculated to be greater than the preset connectivity threshold, then the edge is considered to be connected to the reliable edge pixel and is considered to be part of the reliable edge, and is incorporated into the corresponding reliable edge. If the connectivity between a certain edge and each reliable edge is less than the preset connectivity threshold, then the edge will also be discarded. Finally, the embodiment constructs a rescue area map based on all the reliable edges obtained.

[0053] In some embodiments of the present invention, Figure 6 FIG. 1 is a flow chart of map fusion according to an embodiment of the present invention. Figure 6 As shown, the rescue area map and the preset global map are merged to obtain a fire rescue map, including: S601, identifying map elements on a preset global map and a rescue area map to obtain map element attribute parameters; S602: According to the map element attribute parameters, the preset global map and the rescue area map are associated and matched, and then the maps are fused to obtain a fire rescue map.

[0054] Specifically, after obtaining the rescue area map, it is also necessary to merge it with the preset global map. In this process, the embodiment uses the latest map downloaded from the server as the global map, and associates and matches it with the rescue area map. The association matching process embodiment uses the attribute parameters of the vector map elements as variables, establishes a unified objective function based on the similarity constraints of the attributes, and optimizes the solution to obtain the map fusion result. This fusion process can be carried out in full and on a regular basis, or it can be based on the results of map updates. During the rescue equipment engineering process, efficient fusion is performed after event triggering to provide updated and more accurate map information. Accurate and fast positioning services can be quickly provided for various types of rescue equipment.

[0055] In some embodiments of the present invention, Figure 7 FIG. 1 is a flow chart of robot positioning and matching according to an embodiment of the present invention. Figure 7 As shown, the method for quickly generating a fire rescue map also includes: S701, obtaining road condition collection information of the robot to be positioned, generating a road information model according to the road condition collection information, generating a point cloud information model according to the scanned point cloud data, and matching the road information model with the point cloud information model to obtain a semantic positioning result; S702, generating a local positioning map according to the collected road condition information, and obtaining a map positioning result by matching the local positioning map with the fire rescue map; S703: Combine the semantic positioning result and the map positioning result to obtain the robot positioning result.

[0056] Specifically, in the process of positioning a rescue robot or some other rescue equipment using the fire rescue map obtained in the embodiment of the present invention, the embodiment provides two methods: semantic positioning and map positioning. Among them, semantic positioning generates a road information model through the robot camera and the road condition collection information collected by the laser radar, high-precision inertial navigation, mapping camera and infrared telemetry detection camera carried by the drone, and matches it with the point cloud information model established by the scanned point cloud data to determine the current position. Map positioning is to generate a local positioning map again with the road condition collection information through the above-mentioned method of the present invention, and match it with the existing map to obtain a map positioning result. The embodiment can better determine the approximate position of the rescue robot through map positioning, and can make detailed corrections to the specific position through semantic positioning. Finally, the semantic positioning is combined with the map positioning results to obtain a more accurate robot positioning result.

[0057] In addition, to further speed up the calculation speed of building maps and positioning, the embodiment adopts multi-access edge computing and further extends it to other multi-source wireless access networks. Cloud computing capabilities and corresponding service environments are sunk to the edge of the network (including cellular networks and various wireless network interfaces) to meet the requirements of low latency and high bandwidth.

[0058] The hardware is equipped with the PR series intelligent edge server, which has super computing power and supports multiple Atlas single-slot half-height and half-length PCle standard cards. The whole machine provides 384 channels of 1080P 30FPS video real-time analysis capabilities, providing powerful data processing capabilities for intelligent edge applications.

[0059] The intelligent edge server is equipped with an efficient Kunpeng processor, which has the characteristics of high performance and low power consumption, providing stable and reliable computing power for various edge computing applications. With ultra-high energy efficiency, the power consumption of the Atlas single-slot high half-length PCIe standard card is only 72W, which accelerates the computing power of AI servers and provides a better energy efficiency ratio, saving energy costs while improving computing power.

[0060] The multi-core design and low power consumption characteristics of the Kunpeng architecture enable the PR intelligent edge server to maintain low power consumption while maintaining high performance, building a high-performance, low-power AI computing platform for inference scenarios.

[0061] In summary, in order to quickly generate new maps in fire emergency rescue and provide maps and positioning services for rescue equipment, the present invention first obtains the scanned point cloud data of the rescue area, performs Tai Chi algorithm disk sampling on the scanned point cloud data to obtain a point cloud sampling image; then performs image graying and Gaussian blur processing on the point cloud sampling image to obtain a filtered point cloud image, performs image gradient calculation on the filtered point cloud image to obtain an initial edge image, performs double threshold boundary selection on the initial edge image to obtain a rescue area map; finally, the rescue area map and the preset global map are merged to obtain a fire rescue map. The present invention effectively speeds up the processing efficiency of point cloud data through Tai Chi algorithm disk sampling, reduces redundant data through image graying, and performs data denoising through Gaussian blur processing, effectively improves the calculation efficiency, and realizes the rapid generation of fire rescue maps.

[0062] In order to better implement the method for quickly generating a fire rescue map in the embodiment of the present invention, based on the method for quickly generating a fire rescue map, correspondingly, Figure 8 As shown, the present invention also provides a fire rescue map rapid generation device, the fire rescue map rapid generation device 800 comprises: The point cloud sampling unit 801 is used to obtain the scanned point cloud data of the rescue area, and perform Tai Chi algorithm disk sampling on the scanned point cloud data to obtain a point cloud sampling image; A map generation unit 802 is used to perform image graying and Gaussian blur processing on the point cloud sampling image to obtain a filtered point cloud image, perform image gradient calculation on the filtered point cloud image to obtain an initial edge image, and perform double threshold boundary selection on the initial edge image to obtain a rescue area map; The map fusion unit 803 is used to fuse the rescue area map with the preset global map to obtain a fire rescue map.

[0063] The fire rescue map rapid generation device 800 provided in the above embodiment can implement the technical solution described in the above fire rescue map rapid generation method embodiment. The specific implementation principles of the above modules or units can refer to the corresponding contents in the above fire rescue map rapid generation method embodiment, which will not be repeated here.

[0064] like Fig. 9 As shown, the present invention also provides an electronic device 900. The electronic device 900 includes a processor 901, a memory 902 and a display 903. Fig. 9 Only some components of the electronic device 900 are shown, but it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0065] In some embodiments, the processor 901 may be a central processing unit (CPU), a microprocessor or other data processing chip, used to run program codes or process data stored in the memory 902, such as the method for quickly generating a fire rescue map in the present invention.

[0066] In some embodiments, the processor 901 may be a single server or a server group. The server group may be centralized or distributed. In some embodiments, the processor 901 may be local or remote. In some embodiments, the processor 901 may be implemented in a cloud platform. In one embodiment, the cloud platform may include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, an internal cloud, a multi-cloud, etc., or any combination thereof.

[0067] In some embodiments, the memory 902 may be an internal storage unit of the electronic device 900, such as a hard disk or memory of the electronic device 900. In other embodiments, the memory 902 may also be an external storage device of the electronic device 900, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc., equipped on the electronic device 900.

[0068] Furthermore, the memory 902 may include both an internal storage unit of the electronic device 900 and an external storage device. The memory 902 is used to store application software installed in the electronic device 900 and various data.

[0069] In some embodiments, the display 903 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, etc. The display 903 is used to display information of the electronic device 900 and to display a visual user interface. The components 901-903 of the electronic device 900 communicate with each other through a system bus.

[0070] In one embodiment, when the processor 901 executes the fire rescue map rapid generation program in the memory 902, the following steps may be implemented: Obtain the scanned point cloud data of the rescue area, and perform Tai Chi algorithm disk sampling on the scanned point cloud data to obtain a point cloud sampling image; The point cloud sampling image is grayed and Gaussian blurred to obtain a filtered point cloud image, the filtered point cloud image is image gradient calculated to obtain an initial edge image, and the initial edge image is double-thresholded to select the boundary to obtain a rescue area map; The rescue area map and the preset global map are merged to obtain the fire rescue map.

[0071] It should be understood that: when the processor 901 executes the fire rescue map rapid generation program in the memory 902, in addition to the above functions, other functions can also be implemented. For details, please refer to the description of the corresponding method embodiment above.

[0072] Accordingly, an embodiment of the present application also provides a computer-readable storage medium, which is used to store computer-readable programs or instructions. When the program or instructions are executed by a processor, it can implement the steps or functions of the fire rescue map rapid generation method provided in the above-mentioned method embodiments.

[0073] Those skilled in the art will appreciate that all or part of the processes of the above-mentioned embodiments can be implemented by instructing related hardware (such as a processor, a controller, etc.) through a computer program, and the computer program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0074] The method, device, electronic device and storage medium for quickly generating fire rescue maps provided by the present invention are introduced in detail above. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A method for quickly generating a fire rescue map, characterized in that: include: Obtain scanned point cloud data of the rescue area, and perform Tai Chi algorithm disk sampling on the scanned point cloud data to obtain a point cloud sampling image; Performing image graying and Gaussian blur processing on the point cloud sampling image to obtain a filtered point cloud image, performing image gradient calculation on the filtered point cloud image to obtain an initial edge image, and performing double threshold boundary selection on the initial edge image to obtain a rescue area map; The rescue area map and the preset global map are merged to obtain a fire rescue map.

2. The method for quickly generating a fire rescue map according to claim 1, characterized in that: The step of performing Tai Chi algorithm disk sampling on the scanned point cloud data to obtain a point cloud sampling image includes: Dividing the point cloud space of the scanned point cloud data into a plurality of point cloud space grids; Selecting an initial sampling point, and determining a grid to be scanned at the initial sampling point according to the initial sampling point; Scanning point cloud data in the to-be-scanned grid to obtain a candidate point cloud, and determining a point distance between the candidate point cloud and the initial sampling point; Eliminate the candidate point clouds whose point distance is less than or equal to the preset sampling radius, and use the candidate point clouds whose point distance is greater than the preset sampling radius as candidate sampling points; Determine the distances between the candidate sampling points and the existing sampling points of the to-be-scanned grid in turn, and if they are all greater than the preset sampling radius, use the candidate sampling points as new round sampling points, and traverse the candidate sampling points to obtain all new round sampling points; The sampling process is iterated until the scanning point cloud data is sampled to obtain a point cloud sampling image.

3. The method for quickly generating a fire rescue map according to claim 1, characterized in that: The step of performing image graying and Gaussian blur processing on the point cloud sampling image to obtain a filtered point cloud image includes: Performing dimensionality reduction and grayscale conversion on the point cloud sampling image based on a preset grayscale weight to obtain a grayscale sampling image; The grayscale sampling image is subjected to Gaussian filtering and denoising to obtain a filtered point cloud image.

4. The method for quickly generating a fire rescue map according to claim 1, characterized in that: The performing image gradient calculation on the filtered point cloud image to obtain an initial edge image comprises: Calculating the X-direction gradient and the Y-direction gradient of each pixel in the filtered point cloud image in the X-axis direction and the Y-axis direction respectively; Determine the gradient amplitude and gradient direction of each pixel in the filtered point cloud image according to the X-direction gradient and the Y-direction gradient to obtain an initial gradient image; Non-maximum suppression is performed on the initial gradient image to obtain an initial edge image.

5. The method for quickly generating a fire rescue map according to claim 1, characterized in that: The step of performing double threshold boundary selection on the initial edge image to obtain a rescue area map includes: Setting a first gradient threshold and a second gradient threshold, wherein the first gradient threshold is smaller than the second gradient threshold; Traversing the gradient values ​​of each gradient edge in the initial edge image, taking the gradient edge whose gradient value is greater than the second gradient threshold as a reliable edge, and taking the gradient edge whose gradient value is less than or equal to the second gradient threshold and greater than the first gradient threshold as a candidate edge; Determine the connectivity between the candidate edge and the reliable edge, and if the connectivity is greater than a preset connectivity threshold, merge the candidate edge into the reliable edge; A rescue area map is constructed based on the reliable edges.

6. The method for quickly generating a fire rescue map according to claim 1, characterized in that: The step of fusing the rescue area map with a preset global map to obtain a fire rescue map includes: Performing map element recognition on a preset global map and the rescue area map to obtain map element attribute parameters; The preset global map and the rescue area map are associated and matched according to the map element attribute parameters, and then map fusion is performed to obtain a fire rescue map.

7. The method for quickly generating a fire rescue map according to claim 1, characterized in that: The method further comprises: Acquire road condition collection information of the robot to be positioned, generate a road information model according to the road condition collection information, generate a point cloud information model according to the scanned point cloud data, and match the road information model with the point cloud information model to obtain a semantic positioning result; Generate a local positioning map according to the collected road condition information, and obtain a map positioning result by matching the local positioning map with the fire rescue map; The semantic positioning result and the map positioning result are combined to obtain a robot positioning result.

8. A device for quickly generating a fire rescue map, characterized in that: include: A point cloud sampling unit is used to obtain scanned point cloud data of the rescue area, and perform Tai Chi algorithm disk sampling on the scanned point cloud data to obtain a point cloud sampling image; A map generation unit, configured to perform image graying and Gaussian blur processing on the point cloud sampling image to obtain a filtered point cloud image, perform image gradient calculation on the filtered point cloud image to obtain an initial edge image, and perform double threshold boundary selection on the initial edge image to obtain a rescue area map; The map fusion unit is used to fuse the rescue area map with a preset global map to obtain a fire rescue map.

9. An electronic device comprising a processor, a memory, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method for quickly generating a fire rescue map according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for quickly generating a fire rescue map according to any one of claims 1 to 7 is implemented.

Images