Semantic map construction method based on semantic information and readable storage medium

Abstract

The embodiment of the invention provides a semantic map construction method based on semantic information and a readable storage medium, and belongs to the technical field of fire-fighting maps. The semantic map construction method comprises the following steps: acquiring historical semantic information data of a substation fire-fighting map; constructing a substation fire-fighting map semantic model; according to the method, the historical semantic information data of the transformer substation fire-fighting map is acquired, the semantic model of the transformer substation fire-fighting map is constructed, the historical semantic information data is adopted to train the semantic model of the transformer substation fire-fighting map, and the real-time radar coordinate information and the real-time semantic information of the fire-fighting robot are acquired in the inspection process of the fire-fighting robot; according to the method, the real-time radar coordinate information is obtained, the real-time semantic information is input into the substation fire-fighting map semantic model, the semantic content is obtained, the real-time fire-fighting map is obtained in combination with the real-time radar coordinate information and the semantic content, and the map drawing precision and reliability can be effectively improved through the mode.

Description

Technical Field

[0001] The present invention relates to the technical field of fire map technology, and in particular to a semantic map construction method based on semantic information and a readable storage medium. Background Art

[0002] The fire-fighting robot in the substation is mainly used for fire inspection inside the substation. When conducting inspections, the fire-fighting robot needs to build a trajectory map to assist in movement.

[0003] Currently, the construction of substation fire maps generally involves installing radar on firefighting robots and using the radar to identify and draw the substation fire maps. However, this method only considers the dimensional parameters of radar recognition and ignores the influence of semantic information, resulting in low accuracy in fire map drawing.

[0004] The inventors of this application discovered during the process of realizing the present invention that the above-mentioned solution in the prior art has the defect of low accuracy in drawing fire maps. Summary of the Invention

[0005] The purpose of the embodiments of the present invention is to provide a semantic map construction method based on semantic information and a readable storage medium, which have the function of high drawing accuracy.

[0006] To achieve the above objectives, an embodiment of the present invention provides a method for constructing a semantic map based on semantic information, comprising:

[0007] Obtain historical semantic information data of substation fire map;

[0008] Construct a semantic model of substation fire map;

[0009] Training the substation fire map semantic model according to the historical semantic information data;

[0010] Obtain real-time radar coordinate information and real-time semantic information of the firefighting robot;

[0011] Inputting the real-time semantic information into the substation fire map semantic model to obtain semantic content;

[0012] A real-time firefighting map of the firefighting robot is obtained according to the real-time radar coordinate information and semantic content.

[0013] Optionally, obtaining historical semantic information data of the substation fire map includes:

[0014] Obtain the historical semantic text of the substation fire map;

[0015] adding a label to each of the historical semantic texts;

[0016] A training set is constructed according to the historical semantic text and the labels.

[0017] Optionally, training the substation fire map semantic model according to the historical semantic information data includes:

[0018] Acquire multiple keywords based on the historical semantic text;

[0019] Obtaining a semantic feature vector for each of the keywords;

[0020] The semantic feature vector of the historical semantic text is calculated according to formula (1):

[0021] W=ε1T1+ε2T2+…ε n T n , (1)

[0022] Where W is the semantic feature vector of the historical semantic text, ε n is the semantic weight of the nth keyword, T n is the semantic feature vector of the nth keyword, and n is the total number of keywords.

[0023] Optionally, training the substation fire map semantic model according to the historical semantic information data further includes:

[0024] Obtaining correlation coefficients between the plurality of keywords;

[0025] According to formula (2), the relevant feature vector of the historical semantic text is calculated.

[0026] V=ρ 12 T1T2+ρ 13 T1T3+…ρ (n-1)n T n-1 T n , (2)

[0027] Where V is the relevant feature vector of the historical semantic text, ρ (n-1)n is the correlation coefficient between the n-1th keyword and the nth keyword.

[0028] Optionally, training the substation fire map semantic model according to the historical semantic information data further includes:

[0029] The feature vector of the historical semantic text is calculated according to formula (3):

[0030] X=W+V, (3)

[0031] Wherein, X is the feature vector of the historical semantic text;

[0032] The softmax function is used to classify the feature vectors of the historical semantic text and link them with the corresponding labels.

[0033] Optionally, obtaining a real-time firefighting map of the firefighting robot according to the real-time radar coordinate information and semantic content includes:

[0034] According to formula (4), the yaw angle of the object pixel scanned by the radar is obtained.

[0035]

[0036] Wherein, yaw is the yaw angle, and (x, y, z) are the pixel coordinates of the object scanned by the radar;

[0037] According to formula (5), the pitch angle of the object pixel scanned by the radar is obtained.

[0038]

[0039] Wherein, pitch is the pitch angle.

[0040] Optionally, obtaining a real-time firefighting map of the firefighting robot according to the real-time radar coordinate information and semantic content further includes:

[0041] According to formula (6) and formula (7), the yaw angle and pitch angle of the object scanned by the radar are transformed.

[0042] (yaw)′=yaw+π, (6)

[0043] (pitch)′=fov up -pitch, (7)

[0044] Among them, (yaw)′ is the new yaw angle of the object scanned by the radar, (pitch)′ is the new pitch angle of the object scanned by the radar, and fov up is the vertical downward field of view of the radar sensor.

[0045] Optionally, obtaining a real-time firefighting map of the firefighting robot according to the real-time radar coordinate information and semantic content further includes:

[0046] According to formula (8), the image coordinates of the object scanned by the radar are obtained.

[0047]

[0048] Among them, (u, v) is the image coordinate, col is the width of the image, row is the height of the image, fov down is the vertical upward field of view angle of the radar sensor.

[0049] Optionally, obtaining a real-time firefighting map of the firefighting robot according to the real-time radar coordinate information and semantic content further includes:

[0050] Acquire the real-time position parameters of the firefighting robot according to the image coordinates of the object scanned by the radar;

[0051] The real-time position parameters of the fire-fighting robot are corrected according to the semantic content.

[0052] On the other hand, the present invention further provides a computer-readable storage medium storing instructions, wherein the instructions are used to be read by a machine so as to enable the machine to execute any of the semantic map construction methods described above.

[0053] Through the above technical solution, the semantic map construction method based on semantic information and the readable storage medium provided by the present invention obtain the historical semantic information data of the substation fire map, then construct the substation fire map semantic model, and use the historical semantic information data to train the substation fire map semantic model. During the inspection process of the fire robot, the real-time radar coordinate information and real-time semantic information of the fire robot are collected, and the real-time semantic information is input into the substation fire map semantic model to obtain semantic content. The real-time radar coordinate information and semantic content are combined to obtain a real-time fire map. This method can effectively improve the accuracy and reliability of map drawing.

[0054] Other features and advantages of the embodiments of the present invention will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following detailed description, they are used to explain the embodiments of the present invention, but do not constitute a limitation of the embodiments of the present invention. In the accompanying drawings:

[0056] Figure 1 is a flowchart of a method for constructing a semantic map based on semantic information according to one embodiment of the present invention;

[0057] Figure 2 is a flowchart of constructing a training set in a method for constructing a semantic map based on semantic information according to an embodiment of the present invention;

[0058] Figure 3 is a flowchart of classifying feature vectors in a semantic map construction method based on semantic information according to an embodiment of the present invention;

[0059] Figure 4This is a flowchart of correcting real-time location parameters in a semantic map construction method based on semantic information according to one embodiment of the present invention. DETAILED DESCRIPTION

[0060] The following describes the specific implementation of the embodiment of the present invention in detail with reference to the accompanying drawings. It should be understood that the specific implementation described herein is only used to illustrate and explain the embodiment of the present invention and is not used to limit the embodiment of the present invention.

[0061] Figure 1 Flowchart of a method for constructing a semantic map based on semantic information according to an embodiment of the present invention. Figure 1 In the method for constructing a semantic map, the method may include:

[0062] In step S10, historical semantic information data of the substation fire map is obtained.

[0063] In step S11, a substation fire map semantic model is constructed.

[0064] In step S12, the substation fire map semantic model is trained based on the historical semantic information data.

[0065] In step S13, the real-time radar coordinate information and real-time semantic information of the firefighting robot are obtained.

[0066] In step S14, the real-time semantic information is input into the substation fire map semantic model to obtain semantic content.

[0067] In step S15, a real-time firefighting map of the firefighting robot is obtained according to the real-time radar coordinate information and semantic content.

[0068] In steps S10 to S15, when it is necessary to construct a real-time fire map for the fire-fighting robot, first obtain the historical semantic information data of the substation fire map, then construct a substation fire map semantic model, and use the historical semantic information data to train the substation fire map semantic model. During the inspection process of the fire-fighting robot, collect the real-time radar coordinate information and real-time semantic information of the fire-fighting robot, and input the real-time semantic information into the substation fire map semantic model to obtain semantic content. Combine the real-time radar coordinate information and semantic content to obtain a real-time fire map.

[0069] In this embodiment of the present invention, the steps for acquiring historical semantic information data of the substation fire map can be as follows: Figure 2 Specifically, Figure 2 In the method for constructing a semantic map, the method may further include:

[0070] In step S20, the historical semantic text of the substation fire map is obtained.

[0071] In step S21 , a label is added to each historical semantic text.

[0072] In step S22, a training set is constructed based on the historical semantic text and labels.

[0073] In this embodiment of the present invention, after obtaining the historical semantic information data, it is necessary to train the substation fire map semantic model. The specific training steps can be as follows: Figure 3 Specifically, Figure 3 In the method for constructing a semantic map, the method may further include:

[0074] In step S30, a plurality of keywords are obtained according to the historical semantic text.

[0075] In step S31, the semantic feature vector of each keyword is obtained.

[0076] In step S32, the semantic feature vector of the historical semantic text is calculated according to formula (1):

[0077] W=ε1T1+ε2T2+…ε n T n , (1)

[0078] Among them, W is the semantic feature vector of historical semantic text, ε n is the semantic weight of the nth keyword, T n is the semantic feature vector of the nth keyword, and n is the total number of keywords.

[0079] In step S33, the correlation coefficients between the multiple keywords are obtained.

[0080] In step S34, the relevant feature vector of the historical semantic text is calculated according to formula (2):

[0081] V=ρ 12 T1T2+ρ 13 T1T3+…ρ (n-1)n T n-1 T n , (2)

[0082] Among them, V is the relevant feature vector of historical semantic text, ρ (b-1)n is the correlation coefficient between the n-1th keyword and the nth keyword.

[0083] In step S35, the feature vector of the historical semantic text is calculated according to formula (3):

[0084] X=W+V, (3)

[0085] Among them, X is the feature vector of historical semantic text

[0086] In step S36, the softmax function is used to classify the feature vectors of the historical semantic text and link them with the corresponding labels.

[0087] In steps S30 to S36, the historical semantic text is first segmented to obtain multiple keywords and semantic feature vectors. Semantic feature vectors and related feature vectors are then calculated for the historical semantic text to derive a feature vector. This feature vector is then classified and linked to the pre-labeled data to train the semantic model for the substation fire map.

[0088] In this embodiment of the present invention, after obtaining the real-time radar coordinate information and semantic content, it is necessary to draw a real-time map of the fire-fighting robot. The specific steps can be as follows: Figure 4 Specifically, Figure 4 In the method for constructing a semantic map, the method may further include:

[0089] In step S40, the yaw angle of the object pixel scanned by the radar is obtained according to formula (4):

[0090]

[0091] Among them, yaw is the yaw angle, and (x, y, z) is the pixel coordinates of the object scanned by the radar.

[0092] In step S41, the pitch angle of the object pixel scanned by the radar is obtained according to formula (5):

[0093]

[0094] Among them, pitch is the pitch angle.

[0095] In step S42, the yaw angle and pitch angle of the object scanned by the radar are transformed according to formula (6) and formula (7).

[0096] (yaw)′=yaw+π, (6)

[0097] (pitch)′=fov up -pitch, (7)

[0098] Among them, (yaw)′ is the new yaw angle of the object scanned by the radar, (pitch)′ is the new pitch angle of the object scanned by the radar, and fov up is the vertical downward field of view of the radar sensor.

[0099] In step S43, the image coordinates of the object scanned by the radar are obtained according to formula (8):

[0100]

[0101] Among them, (u, v) is the image coordinate, col is the width of the image, row is the height of the image, fov down It is the vertical upward field of view angle of the radar sensor.

[0102] In step S44, the real-time position parameters of the firefighting robot are obtained according to the image coordinates of the object scanned by the radar.

[0103] In step S45, the real-time position parameters of the firefighting robot are corrected according to the semantic content.

[0104] In steps S40 through S45, the real-time radar coordinate information is processed. Specifically, the yaw and pitch angles of the object pixels scanned by the radar are calculated and transformed to obtain the image coordinates of the scanned object. The firefighting robot's real-time position is determined based on the object's image coordinates and corrected based on the semantic content to obtain an accurate firefighting map for the robot.

[0105] On the other hand, the present invention further provides a computer-readable storage medium storing instructions, wherein the instructions are used to be read by a machine so as to enable the machine to execute any of the above semantic map construction methods.

[0106] Through the above technical solution, the semantic map construction method based on semantic information and the readable storage medium provided by the present invention obtain the historical semantic information data of the substation fire map, then construct the substation fire map semantic model, and use the historical semantic information data to train the substation fire map semantic model. During the inspection process of the fire robot, the real-time radar coordinate information and real-time semantic information of the fire robot are collected, and the real-time semantic information is input into the substation fire map semantic model to obtain semantic content. The real-time radar coordinate information and semantic content are combined to obtain a real-time fire map. This method can effectively improve the accuracy and reliability of map drawing.

[0107] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0108] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0109] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0110] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0111] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0112] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0113] Computer-readable media includes permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0114] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0115] The above are merely embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

Claims

1. A method for constructing a semantic map based on semantic information, characterized in that: include: Obtain historical semantic information data of substation fire map; Construct a semantic model of substation fire map; Training the substation fire map semantic model according to the historical semantic information data; Obtain the real-time radar coordinate information and real-time semantic information of the fire-fighting robot; Inputting the real-time semantic information into the substation fire map semantic model to obtain semantic content; A real-time firefighting map of the firefighting robot is obtained according to the real-time radar coordinate information and semantic content.

2. The method for constructing a semantic map according to claim 1, characterized in that: The historical semantic information data of the substation fire map includes: Get the historical semantic text of the substation fire map; Adding a label to each of the historical semantic texts; A training set is constructed according to the historical semantic text and the labels.

3. The method for constructing a semantic map according to claim 2, characterized in that: Training the substation fire map semantic model according to the historical semantic information data includes: Acquire multiple keywords according to the historical semantic text; Obtaining a semantic feature vector of each of the keywords; The semantic feature vector of the historical semantic text is calculated according to formula (1): W=ε1T1+ε2T2+…ε n T n , (1) Where w is the semantic feature vector of the historical semantic text, ε n is the semantic weight of the nth keyword, T n is the semantic feature vector of the nth keyword, and n is the total number of the keywords.

4. The method for constructing a semantic map according to claim 3, characterized in that: Training the substation fire map semantic model according to the historical semantic information data further includes: Obtaining correlation coefficients between the multiple keywords; According to formula (2), the relevant feature vector of the historical semantic text is calculated. V=ρ 12 T1T2+p 13 T1T3+…p (n-1)n T n-1 T n , (2) Where V is the relevant feature vector of the historical semantic text, ρ (n-1)n is the correlation coefficient between the n-1th keyword and the nth keyword.

5. The method for constructing a semantic map according to claim 4, characterized in that: Training the substation fire map semantic model according to the historical semantic information data further includes: The feature vector of the historical semantic text is calculated according to formula (3): X=W+V, (3) Wherein, X is the feature vector of the historical semantic text; The softmax function is used to classify the feature vectors of the historical semantic text and link them with the corresponding labels.

6. The method for constructing a semantic map according to claim 1, characterized in that: Acquiring the real-time firefighting map of the firefighting robot according to the real-time radar coordinate information and semantic content includes: According to formula (4), the yaw angle of the object pixel scanned by the radar is obtained: Wherein, yaw is the yaw angle, and (x, y, z) is the pixel coordinates of the object scanned by the radar; According to formula (5), the pitch angle of the object pixel scanned by the radar is obtained: Wherein, pitch is the pitch angle.

7. The method for constructing a semantic map according to claim 6, characterized in that: Acquiring the real-time firefighting map of the firefighting robot according to the real-time radar coordinate information and semantic content also includes: According to formula (6) and formula (7), the yaw angle and pitch angle of the object scanned by the radar are transformed. (yaw)′=yaw+π, (6) (pitch)′=fov up -pitch, (7) Among them, (yaw)′ is the new yaw angle of the object scanned by the radar, (pitch)′ is the new pitch angle of the object scanned by the radar, and fov up is the vertical downward field of view of the radar sensor.

8. The method for constructing a semantic map according to claim 7, characterized in that: Acquiring the real-time firefighting map of the firefighting robot according to the real-time radar coordinate information and semantic content also includes: According to formula (8), the image coordinates of the object scanned by the radar are obtained: Among them, (u, v) is the image coordinate, col is the width of the image, row is the height of the image, fov down It is the vertical upward field of view angle of the radar sensor.

9. The method for constructing a semantic map according to claim 8, characterized in that: Acquiring the real-time firefighting map of the firefighting robot according to the real-time radar coordinate information and semantic content also includes: Acquire the real-time position parameters of the fire-fighting robot according to the image coordinates of the object scanned by the radar; The real-time position parameters of the fire-fighting robot are corrected according to the semantic content.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and the instructions are used to be read by a machine so that the machine executes the semantic map construction method according to any one of claims 1 to 9.

Images