Super-view-field target capturing method based on image matching

Abstract

The invention discloses a super-view-field target capturing method based on image matching, and the method comprises the steps: S1, enabling a projectile body to fly to a designated target region according to navigation information, and enabling a seeker to start working; s2, receiving a real-time infrared / television image by a seeker, and selecting a target template in a module library to carry out traversal matching on the real-time image in combination with current bullet body information and pre-bound target information; s3, if the target is successfully captured, entering S7; otherwise, entering S4; s4, if the target is not in the field of view, sequentially selecting the reference point templates in the template library to carry out traversal matching on the real-time image until the matching is successful; s5, according to the successfully matched reference point position and the relative position information of the reference point relative to the target point stored in the template library, calculating the angular deviation of the target relative to the reference point; s6, the seeker outputs a projectile body angle control signal to control the posture of the projectile body to turn to a designated position, and the step S2 is executed; s7, entering a target tracking mode until the target is hit; the target not in the field of view can be captured.

Description

Technical Field

[0001] The present invention belongs to the technical field of computer vision, and in particular relates to a super-viewing field target capturing method based on image matching. Background Art

[0002] As early as the 1970s, image matching was applied to military navigation and terminal guidance, and has now been widely used in many military and civilian fields. The commonly used terminal guidance image matching process usually uses a specified reconnaissance method to obtain the target image and store it as a template image in the missile-borne computer. When the seeker performs search and capture, the system matches the pre-prepared target template image with the real-time image to obtain the position of the target in the real-time image and realize automatic capture of the target. With the low-cost and miniaturized design of the missile body, more and more seekers adopt the strapdown mode, which cannot actively search for multi-angle search, and in order to ensure the sufficient working distance of the seeker, it is impossible to use a super-field-of-view camera. However, there are often errors in the navigation accuracy of the missile body, so it cannot be guaranteed that the target will appear in the field of view of the seeker when the seeker starts working. In view of this situation, the existing target template matching method will not be able to find the target by matching the target template in the field of view, which is not enough to meet the use requirements. It is necessary to provide an ultra-field-of-view target capture method based on image matching. Summary of the invention

[0003] In view of this, the present invention provides a super-field of view target capture method based on image matching, which can solve the technical problem that the existing target template matching method cannot capture targets that are not in the field of view.

[0004] The technical solution for implementing the present invention is as follows:

[0005] A super-field target capture method based on image matching comprises the following steps:

[0006] S1: Before launch, the seeker receives reference images and image parameter information of the target area and surrounding environment from the electro-optical pod or high-definition satellite map;

[0007] S2: Mark the target area in the reference image as the target point, and all the areas with obvious features around the target as reference points, and set a number for each reference point;

[0008] S3: Perform affine transformation and key feature extraction on the target point and all reference point area images, and store the target template, reference point template, image parameters and other information in the template library;

[0009] S4: After launch, the missile body flies to the designated target area according to the predetermined navigation information, and the seeker starts working;

[0010] S5: The seeker receives the real-time infrared / TV image, combines the current missile body information and the pre-bound target information, selects the target template in the module library and performs traversal matching in the real-time image;

[0011] S6: If the target is captured successfully, go to S10; otherwise, go to S07;

[0012] S7: If the target is not in the field of view, select the reference point templates in the template library in turn to perform traversal matching in the real-time image until the match is successful;

[0013] S8: Calculate the angle deviation of the target relative to the reference point according to the position of the successfully matched reference point and the relative position information of the reference point relative to the target point stored in the template library;

[0014] S9: The seeker outputs a missile body angle control signal to control the missile body to turn to the specified position, and enters S5;

[0015] S10: Enter the target tracking mode until the target is hit.

[0016] Furthermore, in step S5, the current projectile information is the real-time attitude angle information of the projectile in the target system.

[0017] Furthermore, in step S6, when the target template cannot be successfully matched in the real-time image, it indicates that there is no target in the current field of view, and the process proceeds to step S7 for reference point template matching. When the target template is successfully matched in the real-time image, it indicates that the target is successfully captured, and the process proceeds to step S10 for target tracking mode.

[0018] Furthermore, in step S7, the reference point templates are selected in order from small to large according to the reference point numbers, and the reference point template matching method is the same as the target template matching method in step S5. All reference point templates in the template library are traversed until a match is successful.

[0019] Furthermore, in step S8, according to the position of the reference point corresponding to the successfully matched reference point template in the target system, the camera parameters are combined to transform from the target system to the imaging coordinate system, and the pixel deviation of the reference point relative to the target point in the imaging coordinate system is calculated. According to the field of view angle and the real-time image resolution, the angular deviation information is solved and output.

[0020] The present invention has at least the following beneficial effects:

[0021] 1. The present invention uses a super-field template matching method to improve the search range of the strapdown seeker without changing the optical system, thereby reducing the navigation accuracy requirements for the missile body and reducing costs.

[0022] 2. The present invention uses a super-field of view template matching method, and the strapdown seeker can select an optical system with a larger focal length to increase the effective distance without changing the navigation accuracy of the missile body.

[0023] 3. The present invention solves the problem of being unable to capture the target when the target is not in the field of view through the super-field template matching method, thereby improving the capture probability. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0025] The present invention is described in detail below with reference to the accompanying drawings and embodiments.

[0026] See also Figure 1 , which is an embodiment of the present invention, the technical solution of the present invention is a step flow chart of a super-field target capture method based on image matching, comprising the following steps:

[0027] Step S1: Before launching, the seeker receives a reference image and image parameter information of the target area and surrounding environment taken by an optoelectronic pod or a high-definition satellite map;

[0028] There are two types of image parameter information. When the reference image is taken by a pod, the parameter information includes the camera resolution, focal length, pixel size, and the angle information of the shooting position relative to the target position. When the reference image is a high-definition satellite image, the parameter information is the satellite image resolution, which refers to the size of a pixel in the satellite image corresponding to the actual geographical area. For example, a resolution of 0.5 meters means that a pixel in the image refers to 0.5 meters in the actual geographical area.

[0029] Step S2: Mark the target area in the reference image as the target point, and all the areas with obvious features around the target as reference points, and set a number for each reference point;

[0030] Among them, the marking method is to intercept the reference image with the target point as the center and the search range of the seeker as the radius. The search range is determined by the ballistic navigation accuracy. The higher the accuracy, the smaller the search range. In the reference image, find other building targets with obvious features other than the target and set them as reference points, and number the reference points in sequence based on the distance from the reference point to the target point. The sparseness of the reference point setting is determined by the instantaneous field of view after the seeker reaches the effective distance. When the instantaneous field of view is 100 meters, the distance between the two reference points shall not exceed 100 meters.

[0031] Step S3: Perform affine transformation and key feature extraction on the target point and all reference point area images, and store information such as the target template, reference point template, and image parameters in a template library;

[0032] Among them, affine transformation is to set multiple continuous transformation angles in azimuth, pitch and roll directions, and perform affine transformation on the target point and all reference point area images in sequence according to the transformation angles to adapt the template to multi-angle image matching. A fixed minimum interval is set between each of the transformation angles, and the maximum transformation angle is the maximum matching angle that the current template can adapt to. The minimum interval is the maximum angle that the matching algorithm can successfully match without changing the template.

[0033] Among them, the target template includes the image features of the target area, the transformation angle and the image features of the target area after the transformation; the reference point template includes the image features of the reference point area, the transformation angle and the image features of the reference point area after the transformation, the reference point relative to the target point position information and the reference point number, the reference point relative to the target point position information is the reference point coordinates in the target system, and the target system is defined as: the target point is the origin, the X-axis is the projection of the X-axis of the missile system on the ground at the time of launch, forward is positive; the Y-axis is celestial, upward is positive; the Z-axis is the right-hand rule; the image parameter information is the same as step S1.

[0034] Among them, image features include gradient features, edge features, grayscale features, and specifically color features for visible light imaging.

[0035] Step S4: After launching, the missile body flies to the designated target area according to the predetermined navigation information, and the seeker starts to work.

[0036] Step S5: The seeker receives the real-time infrared / TV image, combines the current missile body information and the pre-bound target information, selects the target template in the module library and performs traversal matching in the real-time image;

[0037] Among them, the current missile body information is the real-time attitude angle information of the missile body in the target system, and the azimuth angle θ′, pitch angle α′ and roll angle γ′ of the missile body relative to the target in the target system when the target is in the field of view are obtained, and the calculation formula is as follows. The attitude angle conversion sequence in the target system is defined as: the conversion sequence from the target system to the missile system is first yaw, then pitch, and then roll. According to the azimuth angle θ′, pitch angle α′ and roll angle γ′, the target template corresponding to the attitude angle with the smallest difference from the calculated angle is selected from the template library, and the target is captured from the real-time image through the image feature matching algorithm.

[0038]

[0039] Among them, θ is the yaw angle from the target system to the missile system, α is the pitch angle from the target system to the missile system, and γ is the roll angle from the target system to the missile system.

[0040] Step S6: If the target is successfully captured, go to step S10; otherwise, go to step S07. When the target template cannot be successfully matched in the real-time image, it indicates that there is no target in the current field of view, and go to step S7 for reference point template matching. When the target template is successfully matched in the real-time image, it indicates that the target is successfully captured, and go to step S10 target tracking mode.

[0041] Step S7: If the target is not in the field of view, select the reference point templates in the template library in turn and perform traversal matching in the real-time image until the match is successful. The order of selecting the reference point templates is selected in order from small to large according to the reference point number. The reference point template matching method is the same as the target template matching method in step S05. Traverse all reference point templates in the template library until the match is successful.

[0042] Step S8: Calculate the angle deviation of the target relative to the reference point based on the position of the successfully matched reference point and the relative position information of the reference point relative to the target point stored in the template library. According to the position of the reference point corresponding to the successfully matched reference point template in the target system, the pixel deviation of the reference point relative to the target point in the imaging coordinate system is calculated in combination with the camera parameters for conversion from the target system to the imaging coordinate system, and the angle deviation information is solved and output based on the field of view angle and the real-time image resolution.

[0043] Step S9: the seeker outputs a projectile angle control signal to control the projectile to turn to a specified position, and then proceeds to step S5;

[0044] Step S10: Enter the target tracking mode until the target is hit.

[0045] In summary, the above are only preferred embodiments of the present invention and are not intended to limit the protection scope of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A super-field target capture method based on image matching, before launching, the seeker receives a reference image and image parameter information containing the target area and the surrounding environment; the target area position is marked as the target point in the reference image, and the positions of all regions with obvious features around the target are reference points, and each reference point is numbered; the target point and all reference point area images are affine transformed and key features are extracted, and the target template, reference point template, image parameters and other information are stored in a template library; the characteristics are as follows: The following steps are involved: S1: After launch, the missile body flies to the designated target area according to the predetermined navigation information, and the seeker starts working; S2: The seeker receives real-time infrared / TV images, combines the current missile body information and pre-bound target information, and selects the target template in the module library to perform traversal matching in the real-time image; S3: If the target is captured successfully, go to S7; otherwise, go to S4; S4: If the target is not in the field of view, select the reference point templates in the template library in turn to perform traversal matching in the real-time image until the match is successful; S5: Calculate the angle deviation of the target relative to the reference point according to the position of the successfully matched reference point and the relative position information of the reference point relative to the target point stored in the template library; S6: The seeker outputs a missile body angle control signal to control the missile body to turn to the specified position and enter S2; S7: Enter target tracking mode until the target is hit.

2. The method for capturing an ultra-field target according to claim 1, wherein: In the step S2, the current projectile information is the real-time attitude angle information of the projectile under the target system.

3. The method for capturing an object beyond the visual field as claimed in claim 1, characterized in that: In step S3, when the target template cannot be successfully matched in the real-time image, it indicates that there is no target in the current field of view, and the process proceeds to step S4 for reference point template matching; when the target template is successfully matched in the real-time image, it indicates that the target is successfully captured, and the process proceeds to step S7 for target tracking mode.

4. The method for capturing an ultra-field target according to claim 3, wherein: In step S4, the reference point templates are selected in order from small to large according to the reference point numbers, and the reference point template matching method is the same as the target template matching method in step S2; all reference point templates in the template library are traversed until the match is successful.

5. The method for capturing an ultra-field target according to claim 3 or 4, characterized in that: In step S5, according to the position of the reference point corresponding to the successfully matched reference point template in the target system, the target system is converted to the imaging coordinate system in combination with the camera parameters, and the pixel deviation of the reference point relative to the target point in the imaging coordinate system is calculated. According to the field of view angle and the real-time image resolution, the angular deviation information is solved and output.

Images