Relative position-based manual feature representation method, device, equipment and medium

Abstract

The application relates to a relative position-based manual feature representation method, device, equipment and medium. After a detection image containing a to-be-detected target is acquired, a target relative coordinate system of a space position relative to an image coordinate system of the detection image is generated on the detection image with the to-be-detected target as an origin. Then, according to a conversion relationship between the image coordinate system and the target relative coordinate system, the azimuth of a neighbor target of the to-be-detected target in the target relative coordinate system is calculated. Then, after the coordinate position and the azimuth of the neighbor target in the target relative coordinate system are normalized respectively, the manual feature representation of the to-be-detected target is generated. The relative spatial relationship between the target and other targets in the environment is fully utilized for feature representation, so that the targets are more distinguishable and are not affected by the appearance of the targets, and the feature representation reliability of the aircraft target is greatly improved.

Description

Technical Field

[0001] This invention belongs to the field of target detection data processing technology, and relates to a method, apparatus, device and medium for manual feature representation based on relative position. Background Technology

[0002] Target feature representation plays a crucial role in target identification, tracking, and association in the field of aircraft detection, contributing to improved target detection performance and accuracy. Currently, commonly used target feature representations mainly include various hand-crafted feature representations and target appearance depth feature representations extracted through convolutional neural networks (CNNs). Hand-crafted features primarily include target location features, attribute features, appearance features (such as color features and HOG (Histogram of Oriented Gradients) features), and topological features. However, precise target location information in the geodetic coordinate system is often difficult to obtain, and due to aircraft payload limitations, aircraft without a target positioning system cannot locate targets, making location features unavailable. Furthermore, when multiple targets share the same identity attributes, it is impossible to distinguish different targets based on attribute features to determine their identity ID. Therefore, traditional hand-crafted feature representation techniques suffer from insufficient reliability. Summary of the Invention

[0003] To address the problems existing in the above-mentioned traditional methods, this invention proposes a manual feature representation method based on relative position, a manual feature representation device based on relative position, a computer device, and a computer-readable storage medium, which can significantly improve the reliability of feature representation of aircraft targets.

[0004] To achieve the above objectives, the embodiments of the present invention adopt the following technical solutions:

[0005] On the one hand, a method for handcrafted feature representation based on relative position is provided, including the following steps:

[0006] Acquire a detection image containing the target to be tested;

[0007] Using the target to be measured as the origin, a target relative coordinate system is generated on the detection image; the horizontal axis of the target relative coordinate system points to the first geographic direction and is different from the horizontal axis of the image coordinate system of the detection image, and the vertical axis and horizontal axis of the target relative coordinate system are perpendicular to each other.

[0008] Based on the transformation relationship between the image coordinate system and the target relative coordinate system, calculate the azimuth angle of the neighboring targets of the target under test in the target relative coordinate system;

[0009] Based on the normalized coordinates and azimuth of neighboring targets in the target's relative coordinate system, a manual feature representation of the target to be tested is generated.

[0010] In one embodiment, the first geographical direction is due east.

[0011] In one embodiment, the transformation relationship between the image coordinate system and the target relative coordinate system is as follows:

[0012]

[0013] in, This represents the neighboring target A of the target relative to the origin A0 of the coordinate system. i In the target's relative coordinate system, θ0 represents the horizontal axis x of the image coordinate system. I The x-axis of the coordinate system relative to the target T The included angle, [t x ,t y ] T This represents the component of the target relative to the origin A0 of the coordinate system in the image coordinate system. Indicates the origin A0 and its neighboring target A i The angle between the line connecting the two coordinate systems and the horizontal axis of the image coordinate system.

[0014] On the other hand, a handcrafted feature representation device based on relative position is also provided, comprising:

[0015] The image acquisition module is used to acquire detection images containing the target to be tested;

[0016] The relative coordinate module is used to generate a target relative coordinate system on the detection image with the target to be measured as the origin. The horizontal axis of the target relative coordinate system points to the first geographic direction and is different from the horizontal axis of the image coordinate system of the detection image. The vertical axis and horizontal axis of the target relative coordinate system are perpendicular to each other.

[0017] The azimuth calculation module is used to calculate the azimuth angle of the neighboring targets of the target under test in the target relative coordinate system according to the transformation relationship between the image coordinate system and the target relative coordinate system;

[0018] The feature representation module is used to generate a manual feature representation of the target under test based on the normalized coordinate position and azimuth of neighboring targets in the target's relative coordinate system.

[0019] In one embodiment, the first geographical direction is due east.

[0020] In one embodiment, the transformation relationship between the image coordinate system and the target relative coordinate system is as follows:

[0021]

[0022] in, This represents the neighboring target A of the target relative to the origin A0 of the coordinate system. i In the target's relative coordinate system, θ0 represents the horizontal axis x of the image coordinate system. I The x-axis of the coordinate system relative to the target T The included angle, [t x ,t y ] T This represents the component of the target relative to the origin A0 of the coordinate system in the image coordinate system. Indicates the origin A0 and its neighboring target A i The angle between the line connecting the two coordinate systems and the horizontal axis of the image coordinate system.

[0023] In another aspect, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described manual feature representation method based on relative position.

[0024] Furthermore, a computer-readable storage medium is also provided, on which a computer program is stored, which, when executed by a processor, implements the steps of the aforementioned manual feature representation method based on relative position.

[0025] One of the above technical solutions has the following advantages and beneficial effects:

[0026] The aforementioned manual feature representation method, apparatus, device, and medium based on relative position, after acquiring a detection image containing the target, generates a target relative coordinate system on the detection image with the target as the origin, the target's spatial position relative to the image coordinate system of the detection image itself. Then, based on the transformation relationship between the image coordinate system and the target relative coordinate system, the azimuth angles of the target's neighboring targets in the target relative coordinate system are calculated. The coordinate positions and azimuth angles of the neighboring targets in the target relative coordinate system are then normalized to generate a manual feature representation of the target. Compared to traditional manual feature representation techniques, the above scheme fully utilizes the relative spatial relationship between the target and other targets in the environment for feature representation, resulting in stronger distinguishability between targets and independence from the target's appearance. It can overcome the influence of differences in viewing angle and aircraft altitude, significantly improving the reliability of aircraft target feature representation. Attached Figure Description

[0027] To more clearly illustrate the technical solutions in the embodiments of this application or the conventional technology, the drawings used in the description of the embodiments or the conventional technology will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0028] Figure 1 This is a flowchart illustrating a manual feature representation method based on relative position in one embodiment;

[0029] Figure 2 This is a schematic diagram of the target relative coordinate system and the image coordinate system in one embodiment;

[0030] Figure 3 This is another schematic diagram of the target relative coordinate system and the image coordinate system in one embodiment;

[0031] Figure 4 This is a schematic diagram of the module structure of a manual feature representation device based on relative position in one embodiment. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0033] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the application.

[0034] It should be noted that, in this document, the reference to "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The presentation of this phrase in various locations throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments.

[0035] Those skilled in the art will understand that the embodiments described herein can be combined with other embodiments. The term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items, and all possible combinations thereof.

[0036] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0037] Please see Figure 1 In one embodiment, a manual feature representation method based on relative position is provided, which may include the following processing steps S12 to S18:

[0038] S12, acquire a detection image containing the target to be tested.

[0039] As can be understood, the target to be detected refers to the target aircraft that needs to be detected in the current detection scenario. A detection image containing the target to be detected refers to an image or photograph obtained by the carrier aircraft or ground platform performing the detection mission, capturing images of the detection airspace and including the target aircraft. The detection image may show two or more aircraft targets, typically manned aircraft, drones, helicopters, or other aircraft that need to be detected. These images can be captured or obtained by various image acquisition devices (such as cameras, satellites, and drones) deployed on carrier aircraft or ground platforms, and are used to monitor, identify, track, or detect aircraft targets.

[0040] By rationally applying computer vision and image processing technologies to analyze and process detection images containing aircraft targets, it is possible to automatically identify, track, and analyze aircraft targets. These technologies can help achieve more efficient, accurate, and safe aircraft target monitoring and management, and have significant practical application value. Target feature representation is a crucial foundational step in realizing these application values.

[0041] S14: Using the target as the origin, generate a target relative coordinate system on the probe image. The horizontal axis of the target relative coordinate system points to the first geographic direction and is different from the horizontal axis of the probe image coordinate system. The vertical axis and horizontal axis of the target relative coordinate system are perpendicular to each other.

[0042] It is understandable that the first geographical direction can be any of the geographical directions relative to a reference point on the Earth's surface or a compass, such as due east, due north, due south, or due west. The choice depends on the ease of calculation required. After acquiring the probe image, the carrier aircraft can automatically generate the image coordinate system for that image. Generally, such as... Figure 2 As shown, dark nodes in the image coordinate system represent the specified target to be measured, and the origin of the image coordinate system is the upper left corner of the probe image. The horizontal axis of the target relative coordinate system points to the first geographic direction and is different from the horizontal axis of the probe image's coordinate system, ensuring that the relative spatial relationship between the probe target and its surrounding neighboring targets can be effectively utilized. The vertical axis and horizontal axis of the target relative coordinate system are perpendicular to each other; that is, the target relative coordinate system preferably also adopts a two-dimensional rectangular coordinate system to simplify subsequent calculations. Those skilled in the art will understand that, depending on the needs of different detection scenarios, such as target feature representation under three-dimensional probe images, a vertical axis perpendicular to the horizontal and vertical axes can be added similarly, and components in the corresponding direction can be added in subsequent calculations.

[0043] For ease of description, the image coordinate system can be represented as Ix. I y I The target's relative coordinate system is represented as Tx T y TLet the target to be tested be denoted as A0, and its surrounding neighboring targets be numbered as A0. i Then the neighboring target A of A0 is A i In Tx T y T The coordinates in can be represented as i = 1, 2, ..., n, where n is the number of neighboring targets. This is achieved by using the neighboring targets A0 of the target A0 in the detected image under the image coordinate system. i The relative spatial relationships can be used to establish the image coordinate system Ix based on the mathematical principles of coordinate system transformation. I y I Relative coordinate system Tx to the target T y T The transformation relationship is used for calculation and processing in subsequent steps.

[0044] In some implementations, further, the image coordinate system Ix I y I Relative coordinate system Tx to the target T y T The conversion relationship is as follows:

[0045]

[0046] in, This represents the neighboring target A of the target relative to the origin A0 of the coordinate system. i In the target's relative coordinate system, θ0 represents the horizontal axis x of the image coordinate system. I The x-axis of the coordinate system relative to the target T The included angle, [t x ,t y ] T This represents the component of the target relative to the origin A0 of the coordinate system in the image coordinate system. Indicates the origin A0 and its neighboring target A i The angle between the line connecting the coordinates and the horizontal axis of the image coordinate system can be understood as a relatively concise expression of the relationship in this embodiment, which can improve the computational efficiency of subsequent calculations and save computational resources. Those skilled in the art can also use other expressions besides the specific transformation relationship described above, such as, but not limited to, simplified expressions used in coordinate system transformation reasoning, or optimized expressions with adaptive corrections added according to other specific application scenarios.

[0047] S16. Based on the transformation relationship between the image coordinate system and the target relative coordinate system, calculate the azimuth angle of the neighboring targets of the target under test in the target relative coordinate system.

[0048] It is understandable that the neighboring target A0 of the target to be tested is calculated. iIn the target relative coordinate system Tx T y T The azimuth angle below can be expressed as Specifically

[0049] S18. Based on the normalized coordinates and azimuth of neighboring targets in the target's relative coordinate system, generate a manual feature representation of the target to be tested.

[0050] It is understandable that, based on the aforementioned calculated data, the handcrafted feature representation of the target A0 can ultimately be represented as the matrix shown below:

[0051]

[0052] in, They represent the normalized values ​​respectively.

[0053] The aforementioned manual feature representation method based on relative position, after acquiring a detection image containing the target, generates a target-relative coordinate system on the detection image, with the target as the origin, whose spatial position is relative to the image coordinate system of the detection image itself. Then, based on the transformation relationship between the image coordinate system and the target-relative coordinate system, the azimuth angles of the target's neighboring targets in the target-relative coordinate system are calculated. Finally, the coordinate positions and azimuth angles of the neighboring targets in the target-relative coordinate system are normalized to generate a manual feature representation of the target. Compared to traditional manual feature representation techniques, this scheme fully utilizes the relative spatial relationships formed between the target and other targets in the environment for feature representation, resulting in stronger distinguishability between targets and independence from the target's appearance. It can overcome the influence of differences in viewing angle and aircraft altitude, significantly improving the reliability of aircraft target feature representation.

[0054] In one embodiment, further, such as Figure 3 As shown, the first geographical direction is due east.

[0055] It is understood that, in this embodiment, the due east direction is used as the target's relative coordinate system Tx. T y T x-axis T Pointing, target relative to coordinate system Tx T y T The vertical axis y T It can point to due south or due north, depending on the specific application. By uniformly using due east as the horizontal axis (x), it can be optimized. T Pointing can avoid the target's relative coordinate system Tx under different detection images. T y TThis further simplifies the computation process, improves the efficiency of feature representation output, and saves computational resources by addressing complex computational situations such as disordered coordinate axes and non-zero directional components.

[0056] It should be understood that, although the above process Figure 1 The steps in the diagram are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order in which these steps are executed; they can be performed in other orders. Furthermore, the above process... Figure 1 At least some of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0057] In one embodiment, such as Figure 4 As shown, a manual feature representation device 100 based on relative position is provided, including an image acquisition module 11, a relative coordinate module 13, an azimuth calculation module 15, and a feature representation module 17. The image acquisition module 11 acquires a detection image containing the target to be measured. The relative coordinate module 13 generates a target relative coordinate system on the detection image with the target to be measured as the origin; the horizontal axis of the target relative coordinate system points to a first geographic direction and points differently from the horizontal axis of the image coordinate system of the detection image, while the vertical axis and horizontal axis of the target relative coordinate system are perpendicular to each other. The azimuth calculation module 15 calculates the azimuth angles of neighboring targets in the target relative coordinate system according to the transformation relationship between the image coordinate system and the target relative coordinate system. The feature representation module 17 generates a manual feature representation of the target to be measured based on the normalized coordinate positions and azimuth angles of neighboring targets in the target relative coordinate system.

[0058] It is understood that the explanation of each feature in this embodiment can be understood by referring to the explanation of the corresponding features in the manual feature representation method based on relative position, and will not be repeated here.

[0059] The aforementioned manual feature representation device 100 based on relative position, through the collaboration of its modules, after acquiring a detection image containing the target, generates a target relative coordinate system on the detection image with the target as the origin, the target's spatial position relative to the image coordinate system of the detection image itself. Then, based on the transformation relationship between the image coordinate system and the target relative coordinate system, it calculates the azimuth angles of the target's neighboring targets in the target relative coordinate system. After normalizing the coordinate positions and azimuth angles of the neighboring targets in the target relative coordinate system, it generates a manual feature representation of the target. Compared to traditional manual feature representation techniques, this scheme fully utilizes the relative spatial relationships formed between the target and other targets in the environment for feature representation, resulting in stronger distinguishability between targets and independence from the target's appearance. It can overcome the influence of differences in viewing angle and aircraft altitude, significantly improving the reliability of aircraft target feature representation.

[0060] In one embodiment, the first geographical direction is due east.

[0061] In one embodiment, the transformation relationship between the image coordinate system and the target relative coordinate system is further as follows:

[0062]

[0063] in, This represents the neighboring target A of the target relative to the origin A0 of the coordinate system. i In the target's relative coordinate system, θ0 represents the horizontal axis x of the image coordinate system. I The x-axis of the coordinate system relative to the target T The included angle, [t x ,t y ] T This represents the component of the target relative to the origin A0 of the coordinate system in the image coordinate system. Indicates the origin A0 and its neighboring target A i The angle between the line connecting the two coordinate systems and the horizontal axis of the image coordinate system.

[0064] For specific limitations regarding the manual feature representation device 100 based on relative position, please refer to the corresponding limitations of the manual feature representation method based on relative position described above, which will not be repeated here. Each module in the aforementioned manual feature representation device 100 based on relative position can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in hardware or independently of a device capable of processing aircraft detection data, or stored in software in the memory of the aforementioned device, so that the processor can call and execute the operations corresponding to each module. The aforementioned device can be, but is not limited to, various types of image data processing devices already existing in the art.

[0065] In one embodiment, a computer device is also provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to perform the following processing steps: acquiring a probe image containing a target to be measured; generating a target relative coordinate system on the probe image with the target to be measured as the origin; the horizontal axis of the target relative coordinate system points to a first geographic direction and is different from the horizontal axis of the image coordinate system of the probe image, and the vertical axis and horizontal axis of the target relative coordinate system are perpendicular to each other; calculating the azimuth angles of neighboring targets of the target to be measured in the target relative coordinate system according to the transformation relationship between the image coordinate system and the target relative coordinate system; and generating a manual feature representation of the target to be measured based on the normalized coordinate positions and azimuth angles of the neighboring targets in the target relative coordinate system.

[0066] It is understood that, in addition to the memory and processor mentioned above, the computer equipment described above also includes other hardware and software components not listed in this specification. The specific components can be determined according to the model of the computer equipment in different application scenarios, and will not be listed and described in detail in this specification.

[0067] In one embodiment, when the processor executes the computer program, it can also implement the steps or sub-steps added in the various embodiments of the manual feature representation method based on relative position described above.

[0068] In one embodiment, a computer-readable storage medium is also provided, on which a computer program is stored. When executed by a processor, the computer program performs the following processing steps: acquiring a probe image containing a target to be measured; generating a target relative coordinate system on the probe image with the target to be measured as the origin; the horizontal axis of the target relative coordinate system points to a first geographic direction and is different from the horizontal axis of the image coordinate system of the probe image, and the vertical axis and horizontal axis of the target relative coordinate system are perpendicular to each other; calculating the azimuth angles of the neighboring targets of the target to be measured in the target relative coordinate system according to the transformation relationship between the image coordinate system and the target relative coordinate system; and generating a hand-crafted feature representation of the target to be measured based on the normalized coordinate positions and azimuth angles of the neighboring targets in the target relative coordinate system.

[0069] In one embodiment, when the computer program is executed by a processor, it may also implement the steps or sub-steps added to the various embodiments of the manual feature representation method based on relative position described above.

[0070] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), memory bus DRAM (RDRAM), and interface DRAM (DRDRAM), etc.

[0071] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0072] The above embodiments merely illustrate several implementation methods of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, all of which fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A manual feature representation method based on relative position, characterized in that, Including the following steps: Acquire a detection image containing the target to be tested; Using the target to be measured as the origin, a target relative coordinate system is generated on the detection image; the horizontal axis of the target relative coordinate system points to a first geographic direction and is different from the horizontal axis of the image coordinate system of the detection image, and the vertical axis and horizontal axis of the target relative coordinate system are perpendicular to each other; Based on the transformation relationship between the image coordinate system and the target relative coordinate system, calculate the azimuth angle of the neighboring targets of the target under test in the target relative coordinate system; Based on the normalized coordinate positions and azimuth angles of the neighboring targets in the target's relative coordinate system, a manual feature representation of the target to be tested is generated; The transformation relationship between the image coordinate system and the target relative coordinate system is as follows: in, The origin of the target relative to the coordinate system is indicated. The neighbor target The coordinates in the target relative coordinate system The horizontal axis representing the image coordinate system The horizontal axis of the coordinate system relative to the target The included angle, The origin of the target relative to the coordinate system is indicated. Components in the image coordinate system, Represents the origin With the neighboring target The angle between the line connecting the two coordinate systems and the horizontal axis of the image coordinate system.

2. The manual feature representation method based on relative position according to claim 1, characterized in that, The first geographical direction is due east.

3. A manual feature representation device based on relative position, characterized in that, include: The image acquisition module is used to acquire detection images containing the target to be tested; The relative coordinate module is used to generate a target relative coordinate system on the detection image with the target under test as the origin; The horizontal axis of the target relative coordinate system points to the first geographic direction and is different from the horizontal axis of the image coordinate system of the detected image. The vertical axis and horizontal axis of the target relative coordinate system are perpendicular to each other. The orientation calculation module is used to calculate the azimuth angle of the neighboring targets of the target under test in the target relative coordinate system according to the transformation relationship between the image coordinate system and the target relative coordinate system; The feature representation module is used to generate a manual feature representation of the target under test based on the normalized coordinate position and azimuth of the neighboring targets in the target relative coordinate system. The transformation relationship between the image coordinate system and the target relative coordinate system is as follows: in, The origin of the target relative to the coordinate system is indicated. The neighbor target The coordinates in the target relative coordinate system The horizontal axis representing the image coordinate system The horizontal axis of the coordinate system relative to the target The included angle, The origin of the target relative to the coordinate system is indicated. Components in the image coordinate system, Represents the origin With the neighboring target The angle between the line connecting the two coordinate systems and the horizontal axis of the image coordinate system.

4. The manual feature representation device based on relative position according to claim 3, characterized in that, The first geographical direction is due east.

5. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the manual feature representation method based on relative position as described in claim 1 or 2.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the manual feature representation method based on relative position as described in claim 1 or 2.

Images