Cognitive positioning method based on ai vision

Abstract

The application provides an AI vision-based cognitive positioning method, comprising the following steps: S1, target recognition is performed on an image captured by an airborne camera, and output information is converted to a normalized coordinate system with an image center as an origin; S2, corresponding coordinates of a prediction box of the normalized coordinate system are calculated, and the coordinates are enlarged according to an original size of the image to obtain corner point coordinates in an image coordinate system; S3, a coordinate dimension is increased, and the image coordinate point is converted to a camera coordinate system; S4, the camera coordinate system is converted to a carrier body coordinate system again; and S5, the carrier body coordinate system is converted to a geographic coordinate system, and positioning of a target is completed. The method identifies a target in an image through an artificial intelligence algorithm, has low cost and high identification accuracy.

Description

Technical Field

[0001] This invention relates to the field of visual positioning and target detection technology, specifically to a cognitive positioning method, electronic device, and readable storage medium based on AI vision. Background Technology

[0002] Target localization has always been a key research focus in the field of electronic reconnaissance. Currently, mature localization methods mainly include radar localization and satellite localization, but these two methods are susceptible to electromagnetic interference and are relatively expensive. Passive localization relies primarily on the analysis and processing of electromagnetic wave signals radiated by the target, but requires coordinated data processing from multiple observation stations.

[0003] Therefore, it is necessary to provide a new target identification and localization method to reduce the cost of target localization. Summary of the Invention

[0004] Technical problems to be solved

[0005] To address the aforementioned shortcomings of existing technologies, this invention provides a cognitive localization method based on AI vision. This method identifies targets in images using artificial intelligence algorithms, resulting in low cost and high accuracy.

[0006] Technical solution

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] This invention provides a cognitive localization method based on AI vision, comprising the following steps:

[0009] S1. Perform target recognition on images captured by the airborne camera and convert the output information to a normalized coordinate system with the image center as the origin.

[0010] S2. Calculate the corresponding coordinates of the normalized coordinate system prediction box, and enlarge the coordinates according to the original size of the image to obtain the corner coordinates in the image coordinate system;

[0011] S3. Add coordinate dimensions to transform image coordinate points to the camera coordinate system;

[0012] S4. Transform the camera coordinate system to the aircraft body coordinate system;

[0013] S5. Transform the aircraft body coordinate system to the geographic coordinate system to complete the target location.

[0014] Furthermore, the AI ​​vision-based cognitive localization method uses the YOLO v4 target detection algorithm to perform target recognition on images captured by an airborne camera.

[0015] Furthermore, the output information of the YOLO v4 object detection algorithm is transformed into a normalized coordinate system with the image center as the origin using the following formula:

[0016]

[0017] Where, x i ,y i These are the center coordinates of the YOLO network's predicted bounding box, w i ,h i These are the width and height values ​​of the prediction box, respectively.

[0018] Further, the corresponding coordinates of the predicted bounding box in the normalized coordinate system are calculated according to the following formula:

[0019]

[0020] Furthermore, by enlarging the corresponding coordinates of the normalized coordinate system prediction box according to the original size of the image, the corner coordinates in the image coordinate system are obtained as follows:

[0021]

[0022] Then the image coordinates are transformed to the camera coordinate system [x] s ,y s ,z s ].

[0023] Furthermore, the camera coordinate system [x] s ,y s ,z s Transform to the aircraft body coordinate system using the following formula:

[0024] (x a ,y a ,z a ) T =Q2Q1(x s ,y s ,z s ) T

[0025]

[0026] Where α and β are the camera's azimuth and elevation angles, respectively.

[0027] Furthermore, the aircraft body coordinate system is transformed to the geographic coordinate system according to the following formula:

[0028] (x g ,y g ,z g ) T =Q5Q4Q3(x a ,ya ,z a ) T

[0029]

[0030] Where, ψ,θ, These are the drone's yaw angle, pitch angle, and roll angle, respectively.

[0031] Based on the same inventive concept, the present invention also provides an electronic device, including a processor and a memory, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, it implements the method described in any of the above-mentioned embodiments.

[0032] Based on the same inventive concept, the present invention also provides a readable storage medium storing a computer program, which, when executed by a processor, implements the method described in any of the above-mentioned embodiments.

[0033] Beneficial effects

[0034] This invention designs a cognitive positioning method based on AI vision, which combines AI vision with positioning technology. It uses an unmanned platform (drone) to acquire visual information of the target, identifies the target in the image through artificial intelligence algorithms, and performs coordinate transformation between the image position and the actual spatial position of the target according to the principle of visual imaging, thereby calculating the geographical location information of the target. A relatively accurate target position is obtained through matrix transformation, which is low cost. Attached Figure Description

[0035] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0036] Figure 1 This is a schematic diagram illustrating the steps of a cognitive localization method based on AI vision according to an embodiment of the present invention.

[0037] Figure 2 This is a schematic diagram of a cognitive localization method based on AI vision provided in an embodiment of the present invention. Detailed Implementation

[0038] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0039] See Figure 1 An embodiment of the present invention provides a cognitive localization method based on AI vision, comprising the following steps:

[0040] S1. Perform target recognition on images captured by the airborne camera and convert the output information to a normalized coordinate system with the image center as the origin.

[0041] S2. Calculate the corresponding coordinates of the normalized coordinate system prediction box, and enlarge the coordinates according to the original size of the image to obtain the corner coordinates in the image coordinate system;

[0042] S3. Add coordinate dimensions to transform image coordinate points to the camera coordinate system;

[0043] S4. Transform the camera coordinate system to the aircraft body coordinate system;

[0044] S5. Transform the aircraft body coordinate system to the geographic coordinate system to complete the target location.

[0045] In practice, the process begins with accurate target identification within the captured image. Then, based on the position of the image point in the image's coordinate system, its coordinates in the camera coordinate system are calculated. Next, based on the relationship between the camera coordinate system and the aircraft's body coordinate system, the transformation matrix between them is calculated to determine the image point's coordinates in the aircraft's body coordinate system. Then, based on the transformation relationship between the aircraft's body coordinate system and the aircraft's geographic coordinate system, the image point's coordinates in the aircraft's geographic coordinate system are determined. Finally, based on the geometric relationship between the image point and the target point, the target's coordinates in the ground coordinate system are calculated, thus completing the target localization. This invention combines AI vision with localization technology, utilizing matrix-based coordinate transformation relationships to effectively extract target location information.

[0046] In this embodiment, refer to Figure 2 First, based on the YOLO v4 target detection algorithm, image targets are identified. The training process was optimized to meet the characteristics of remote sensing images and the requirements of localization algorithms, and the output parameters were modified to achieve automatic target detection and subsequent localization. The YOLO output information is converted to a normalized coordinate system with the image center as the origin using the following formula:

[0047]

[0048] Where, x i ,y i These are the center coordinates of the YOLO network's predicted bounding box, w i ,h i These are the width and height values ​​of the prediction box, respectively.

[0049] Then calculate the corresponding coordinates of the predicted bounding box in that coordinate system.

[0050]

[0051] Finally, the coordinates are enlarged according to the original size of the image to obtain the corner coordinates in the image coordinate system.

[0052]

[0053] Transform the image coordinates to the camera coordinate system. Since the image coordinate system and the camera coordinate system are parallel, the coordinates of the target center in the camera coordinate system only need to be increased by adding the camera's z-axis coordinate, i.e., [x...]. s ,y s ,z s ].

[0054] Then transform the camera coordinate system to the carrier body coordinate system.

[0055] (x a ,y a ,z a ) T =Q2Q1(x s ,y s ,z s ) T

[0056]

[0057] Where α and β are the camera's azimuth and elevation angles, respectively.

[0058] Finally, the aircraft's coordinate system is transformed to the geographic coordinate system:

[0059] (x g ,y g ,z g ) T =Q5Q4Q3(x a ,y a ,z a ) T

[0060]

[0061] Where, ψ,θ, These are the drone's yaw angle, pitch angle, and roll angle, respectively.

[0062] Specifically, this invention proposes a cognitive positioning method based on AI vision. It acquires visual information of a target using an unmanned platform (UAV), identifies the target in the image using an artificial intelligence algorithm, and performs coordinate transformation between the target's image position and its actual spatial position based on visual imaging principles, thereby calculating the target's geographical location information. The specific process of the algorithm proposed in this invention is as follows: First, the target within the captured image is accurately identified; then, based on the position of the image point in the image's coordinate system, the coordinates of the image point in the camera coordinate system are calculated; next, based on the relationship between the camera coordinate system and the aircraft's body coordinate system, the transformation matrix between them is calculated to solve for the image point's coordinates in the aircraft's body coordinate system; then, based on the transformation relationship between the aircraft's body coordinate system and the aircraft's geographic coordinate system, the coordinates of the image point in the aircraft's geographic coordinate system are solved; finally, based on the geometric relationship between the image point and the target point, the coordinates of the target in the ground coordinate system are solved, completing the target positioning. The essence of target positioning is solving the matrix transformation problem between different coordinate systems.

[0063] In this embodiment, the YOLO-v4 algorithm, based on the original YOLO object detection architecture, employs the best optimization strategies in the CNN field in recent years. It features optimizations to varying degrees in data processing, backbone network, network training, activation functions, and loss functions. While lacking theoretical innovation, it is welcomed by many engineers who experiment with various optimization algorithms. This article serves as a review of object detection tricks, achieving a balance between FPS and precision in object detection.

[0064] Based on the same inventive concept, the present invention also provides an electronic device, including a processor and a memory, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the AI ​​vision-based cognitive positioning method is implemented.

[0065] In some embodiments, the processor may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor (e.g., a GPU (Graphics Processing Unit)), or other data processing chip. The processor is typically used to control the overall operation of the electronic device. In this embodiment, the processor is used to run program code stored in the memory or process data, for example, to run the program code of the AI ​​vision-based cognitive localization method.

[0066] The memory includes at least one type of readable storage medium, including flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory may be an internal storage unit of the electronic device, such as the hard disk or memory of the electronic device. In other embodiments, the memory may also be an external storage device of the electronic device, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc. Of course, the memory may include both internal storage units and external storage devices of the electronic device. In this embodiment, the memory is typically used to store operating methods and various application software installed on the electronic device, such as the program code of the AI ​​vision-based cognitive positioning method. Furthermore, the memory can also be used to temporarily store various types of data that have been output or will be output.

[0067] Based on the same inventive concept, the present invention also provides a readable storage medium storing a computer program, which, when executed by a processor, implements the AI ​​vision-based cognitive positioning method.

[0068] The advantage of this invention lies in the design of a cognitive positioning method based on AI vision, which combines AI vision with positioning technology. It uses an unmanned platform (drone) to acquire the visual information of the target, identifies the target in the image through artificial intelligence algorithms, and performs coordinate transformation between the image position and the actual spatial position of the target according to the principle of visual imaging, thereby calculating the geographical location information of the target. A relatively accurate target position is obtained through matrix transformation, and the cost is low.

[0069] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the protection scope of the technical solutions of the embodiments of the present invention.

Claims

1. A cognitive localization method based on AI vision, characterized in that, Includes the following steps: S1. Perform target recognition on images captured by the airborne camera and convert the output information to a normalized coordinate system with the image center as the origin. S2. Calculate the corresponding coordinates of the normalized coordinate system prediction box, and enlarge the coordinates according to the original size of the image to obtain the corner coordinates in the image coordinate system; S3. Add coordinate dimensions to transform image coordinate points to the camera coordinate system; S4. Transform the camera coordinate system to the aircraft body coordinate system; S5. Transform the aircraft body coordinate system to the geographic coordinate system to complete the target location; The AI ​​vision-based cognitive localization method is based on the YOLO v4 target detection algorithm and performs target recognition on images captured by an airborne camera. The output information of the YOLO v4 object detection algorithm is converted to a normalized coordinate system with the image center as the origin using the following formula: , in, These are the center coordinates of the YOLO network's predicted bounding box. These are the width and height values ​​of the prediction box, respectively. The corresponding coordinates of the predicted bounding box in the normalized coordinate system are calculated according to the following formula: ； The coordinates of the predicted bounding box in the normalized coordinate system are enlarged based on the original size of the image to obtain the corner coordinates in the image coordinate system as follows: , Then the image coordinates are transformed to the camera coordinate system. .

2. The AI ​​vision-based cognitive localization method according to claim 1, characterized in that, The camera coordinate system Transform to the aircraft body coordinate system using the following formula: , , in, For the camera's azimuth and elevation angles.

3. The AI ​​vision-based cognitive localization method according to claim 2, characterized in that, The aircraft body coordinate system is transformed to the geographic coordinate system using the following formula: , , in, These are the drone's yaw angle, pitch angle, and roll angle, respectively.

4. An electronic device, characterized in that, It includes a processor and a memory, wherein the memory stores a computer program, which, when executed by the processor, implements the method as described in claims 1-3.

5. A readable storage medium, characterized in that, The readable storage medium stores a computer program, which, when executed by a processor, implements the method as described in claims 1-3.

Images