Inertial navigation information correction method, device and equipment based on multi-sub-region prior information

Abstract

This application relates to an inertial navigation information correction method, apparatus, and device based on multi-sub-region prior information. The method includes: acquiring a real-time image dataset consisting of multiple frames of real-time images of ground targets acquired from high altitude, and a reference image containing accurate information about the ground target's location; performing geometric correction on each frame of the real-time image to obtain a corrected real-time image; selecting multiple regions of interest (ROIs) in the reference image; locating these ROIs in the corrected real-time image using a matching method to obtain corresponding ROIs; calculating an estimated height error based on the distances between the centers of each ROI in the reference image and the distances between the centers of the corresponding ROIs in the corrected real-time image; and correcting the height information in the inertial navigation information based on the estimated height error, thereby improving the environmental adaptability and positioning accuracy of the target recognition system.

Description

Technical Field

[0001] This application relates to the field of ground target recognition technology, and in particular to an inertial navigation information correction method, apparatus and equipment based on multi-sub-region prior information. Background Technology

[0002] The ground multi-mode imaging guidance system adopts advanced detection and information processing technologies to achieve precise strikes on targets. Among them, target recognition technology under complex ground backgrounds is the core and difficult point. At present, target recognition technology is mainly divided into the following two schemes: (1) The scheme of identifying targets by extracting information from images, namely Automatic Target Recognition (ATR); This scheme: First, a "target knowledge base" is constructed based on an expert system; Then, the target is identified by calculating the projection of the image onto the space spanned by the "target knowledge base". Here, the "target knowledge base" can be the feature information of the image or the statistical information of the image. (2) The scheme of capturing targets by scene matching (image matching), namely Automatic Target Acquisition (ATA); This scheme: First, the primitives to be matched in the reference image and the real-time image are extracted; Then, the descriptors of the primitives to be matched are constructed and matched; Finally, the image registration parameters are calculated by using the matching space relationship between the primitives, thereby capturing the target.

[0003] Automatic Target Recognition (ATR) schemes heavily rely on the construction of a "target knowledge base," which computer vision typically instantiates using image features. Therefore, ATR schemes either employ feature engineering methods such as sparse representation and bag-of-words models, or traditional machine learning methods based on SVM. However, the former cannot effectively address the robustness of the "target knowledge base" under various perspective distortion conditions, while the latter can only identify the target categories contained in the image, not the precise location of the target within the image. Thus, traditional ATR schemes are difficult to apply to ground imaging guidance missions in high-speed missile-borne environments. Automatic Target Acquisition (ATA), on the other hand, inherently possesses the ability to precisely locate the coordinates of the strike point. This scheme finds the corresponding position of any point in the reference image in the real-time image through registration between a reference image and a real-time image. Therefore, the ATA method remains the mainstream research direction for missile-borne imaging guidance schemes.

[0004] However, existing automatic target acquisition schemes require the use of inertial navigation parameters to correct real-time images. However, the errors in inertial navigation parameters gradually accumulate with the flight distance, resulting in scale differences between the corrected real-time image and the reference image, making it difficult to achieve correct matching. Summary of the Invention

[0005] Therefore, it is necessary to provide an inertial navigation information correction method, apparatus, and device based on multi-sub-region prior information that can perform rapid correction to address the above-mentioned technical problems.

[0006] A method for inertial navigation information correction based on multi-sub-region prior information, the method comprising:

[0007] Acquire a real-time image dataset and a reference image, wherein the real-time image dataset contains multiple frames of real-time images of ground targets acquired from high altitude, and the reference image contains accurate information about the location of the ground targets;

[0008] Geometric correction is performed on each frame of the real-time image to obtain a corrected real-time image. Multiple regions of interest are selected in the reference image, and the multiple regions of interest are located in the corrected real-time image based on a matching method to obtain the corresponding multiple regions of interest.

[0009] The first estimated height error is calculated based on the distance between the centers of each region of interest in the reference image and the distance between the centers of the corresponding regions of interest in the corrected real-time image.

[0010] The first estimated height error is multiplied by the corresponding confidence level to obtain the second estimated height error corresponding to the real-time image;

[0011] The average value of the second estimated altitude error for every 6 frames of real-time images is calculated, and this average value is used as the altitude error. The altitude information in the inertial navigation information is gradually corrected based on the altitude error until the altitude error converges.

[0012] In one embodiment, the region of interest includes two or more than three regions.

[0013] In one embodiment, when there are three or more regions of interest, multiple first estimated height errors can be calculated for each real-time image frame, and their average value is calculated as the first estimated height error corresponding to each real-time image frame.

[0014] In one embodiment, before calculating the average of multiple first estimated altitude errors, each first estimated altitude error is judged based on the inherent error of the inertial navigation system, and if the first estimated altitude error is greater than the inherent error, it is discarded.

[0015] In one embodiment, when there are two regions of interest, the first estimated height error corresponding to each frame of real-time image is judged based on the inherent error of the inertial navigation system. If the first estimated height error is greater than the inherent error, it is discarded.

[0016] In one embodiment, the confidence level is 0.65 when there are two regions of interest, and 0.75 when there are three or more regions of interest.

[0017] In one embodiment, the first estimated height error is calculated based on the distance between the centers of each region of interest in the reference image and the distance between the corresponding centers of the regions of interest in the corrected real-time image using the following formula:

[0018]

[0019] In the above formula, L AB L represents the distance between the centers of region of interest A and region of interest B in the reference image. A′B′ H' represents the distance between the centers of regions of interest A and B' corresponding to regions of interest A and B in the corrected real-time image, and H' represents the altitude information obtained from the inertial navigation system.

[0020] An inertial navigation information correction device based on multi-sub-region prior information, the device comprising:

[0021] The image data acquisition module is used to acquire a real-time image dataset and a reference image. The real-time image dataset contains multiple frames of real-time images acquired from high altitudes of ground targets, and the reference image contains accurate information about the location of the ground targets.

[0022] The region of interest matching module is used to perform geometric correction on each frame of the real-time image to obtain a corrected real-time image, select multiple regions of interest in the reference image, and locate the multiple regions of interest in the corrected real-time image based on the matching method to obtain the corresponding multiple regions of interest.

[0023] The first estimated height error calculation module is used to calculate the first estimated height error based on the distance between the centers of each region of interest in the reference image and the distance between the centers of the corresponding regions of interest in the corrected real-time image.

[0024] The second estimated height error calculation module is used to multiply the first estimated height error according to the corresponding confidence level to obtain the second estimated height error corresponding to the real-time image.

[0025] The inertial navigation information correction module is used to calculate the average value of the second estimated altitude error for every 6 frames of real-time images, and use this average value as the altitude error. The altitude information in the inertial navigation information is gradually corrected based on the altitude error until the altitude error converges.

[0026] A computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program performing the following steps:

[0027] Acquire a real-time image dataset and a reference image, wherein the real-time image dataset contains multiple frames of real-time images of ground targets acquired from high altitude, and the reference image contains accurate information about the location of the ground targets;

[0028] Geometric correction is performed on each frame of the real-time image to obtain a corrected real-time image. Multiple regions of interest are selected in the reference image, and the multiple regions of interest are located in the corrected real-time image based on a matching method to obtain the corresponding multiple regions of interest.

[0029] The first estimated height error is calculated based on the distance between the centers of each region of interest in the reference image and the distance between the centers of the corresponding regions of interest in the corrected real-time image.

[0030] The first estimated height error is multiplied by the corresponding confidence level to obtain the second estimated height error corresponding to the real-time image;

[0031] The average value of the second estimated altitude error for every 6 frames of real-time images is calculated, and this average value is used as the altitude error. The altitude information in the inertial navigation information is gradually corrected based on the altitude error until the altitude error converges.

[0032] A computer-readable storage medium having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0033] Acquire a real-time image dataset and a reference image, wherein the real-time image dataset contains multiple frames of real-time images of ground targets acquired from high altitude, and the reference image contains accurate information about the location of the ground targets;

[0034] Geometric correction is performed on each frame of the real-time image to obtain a corrected real-time image. Multiple regions of interest are selected in the reference image, and the multiple regions of interest are located in the corrected real-time image based on a matching method to obtain the corresponding multiple regions of interest.

[0035] The first estimated height error is calculated based on the distance between the centers of each region of interest in the reference image and the distance between the centers of the corresponding regions of interest in the corrected real-time image.

[0036] The first estimated height error is multiplied by the corresponding confidence level to obtain the second estimated height error corresponding to the real-time image;

[0037] The average value of the second estimated altitude error for every 6 frames of real-time images is calculated, and this average value is used as the altitude error. The altitude information in the inertial navigation information is gradually corrected based on the altitude error until the altitude error converges.

[0038] The aforementioned inertial navigation information correction method, apparatus, and device based on multi-sub-region prior information acquire a real-time image dataset consisting of multiple frames of real-time images of ground targets acquired from high altitudes, and a reference image containing accurate information about the location of the ground targets. Geometric correction is performed on each frame of real-time image to obtain a corrected real-time image. Multiple regions of interest (ROIs) are selected in the reference image, and these ROIs are located in the corrected real-time image using a matching method to obtain corresponding ROIs. An estimated height error is calculated based on the distance between the centers of each ROI in the reference image and the distance between the centers of the corresponding ROIs in the corrected real-time image. The height information in the inertial navigation information is then corrected based on the estimated height error, thereby improving the environmental adaptability and positioning accuracy of the target recognition system. Attached Figure Description

[0039] Figure 1 This is a flowchart illustrating an inertial navigation information correction method based on multi-sub-region prior information in one embodiment.

[0040] Figure 2 This is a real-time image after geometric correction in one embodiment;

[0041] Figure 3 This is a reference image in one embodiment;

[0042] Figure 4 In one embodiment, three regions of interest are selected from a reference image, wherein... Figure 4 (a) represents region of interest 1. Figure 4 (b) represents region of interest 2. Figure 4 (c) represents region of interest 3;

[0043] Figure 5 This is a schematic diagram showing the positions of three regions of interest in a reference image in one embodiment;

[0044] Figure 6 This is a schematic diagram showing the corresponding positions of three regions of interest in the corrected real-time image in one embodiment;

[0045] Figure 7 This is a schematic diagram of height estimation in one embodiment;

[0046] Figure 8 This is a schematic diagram of the -300m height error estimation results in an experimental verification.

[0047] Figure 9 This is a schematic diagram of the +300m height error estimation results in an experimental verification.

[0048] Figure 10 This is a real-time image of two sets of inputs used in an experiment to verify the data.

[0049] Figure 11 This is a schematic diagram of the recognition results when the height error between two sets of real-time images is -300m in an experimental verification.

[0050] Figure 12 This is a schematic diagram of the recognition results when the height error between two sets of real-time images is +300m in an experimental verification.

[0051] Figure 13 This is a schematic diagram of the recognition results after height error correction for two sets of real-time images in an experimental verification.

[0052] Figure 14 This is a structural block diagram of an inertial navigation information correction device based on multi-sub-region prior information in one embodiment;

[0053] Figure 15 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0054] 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.

[0055] Existing automatic target acquisition schemes require inertial navigation parameters to correct real-time images. However, these parameters accumulate errors over time, leading to scale differences between the corrected real-time image and the reference image, making accurate matching difficult. Figure 1 As shown, an inertial navigation information correction method based on multi-sub-region prior information is provided, including the following steps:

[0056] Step S100: Obtain a real-time image dataset and a reference image, wherein the real-time image dataset contains multiple frames of real-time images of ground targets acquired from high altitude, and the reference image contains accurate information about the location of the ground targets.

[0057] Step S110: Perform geometric correction on each frame of real-time image to obtain the corrected real-time image. Select multiple regions of interest in the reference image and locate the multiple regions of interest in the corrected real-time image based on the matching method to obtain the corresponding multiple regions of interest.

[0058] Step S120: Calculate the first estimated height error based on the distance between the centers of each region of interest in the reference image and the distance between the corresponding centers of the regions of interest in the corrected real-time image.

[0059] Step S130: Multiply the first estimated height error according to the corresponding confidence level to obtain the second estimated height error corresponding to the real-time image;

[0060] Step S140: Calculate the average value of the second estimated height error for every 6 frames of real-time images, and use this average value as the height error. Then, gradually correct the height information in the inertial navigation information based on the height error until the height error converges.

[0061] For the localization of static ground targets, effective information surrounding the target can be fully utilized for target identification. This involves comprehensively locating the target by matching multiple regions of interest (ROIs) around it. Simultaneously, the relative positions of these ROIs can be used to correct inertial navigation parameters, thereby improving the environmental adaptability and positioning accuracy of the target identification system.

[0062] In this embodiment, firstly, multiple regions of interest (ROIs) with obvious features are selected in the reference image. Then, the position of each ROI in the real-time image is located by matching, thereby obtaining multiple potential positions of the target. Finally, the error value of the inertial navigation parameters can be calculated by using the relative positional relationship between each ROI in the real-time image, thereby correcting the error of the inertial navigation parameters.

[0063] Specifically, in step S100, each frame of the real-time image dataset is acquired from high altitude by an infrared airborne device, and each frame of the real-time image is an infrared image. The reference image is an optical image acquired by a satellite, which includes precise information about the location of the ground target.

[0064] In step S110, geometric correction processing is performed on each frame of the real-time image in the real-time image dataset. The corrected real-time image is as follows: Figure 2 As shown. It should be noted that this method does not operate on the complete real-time image dataset that has already been acquired, but rather operates on the currently acquired real-time image while acquiring the real-time image, and corrects the inertial navigation parameters at the same time as acquiring the real-time image.

[0065] like Figure 3-5 As shown, three regions of interest can be extracted from the reference image. After matching the reference image with the real-time image, the positions of the three regions of interest in the reference image in the corrected real-time image are as follows: Figure 6 As shown.

[0066] In this embodiment, the method for matching the reference image and the real-time image is not limited. Various methods can be used to match the two images in order to locate the region of interest of the corresponding reference image in the real-time image.

[0067] In this embodiment, the regions of interest extracted from the reference image can be two or more than three. This paper uses three regions of interest as an example.

[0068] Since the distance between each region of interest (ROI) in the real-time image looking down differs from the distance between each ROI in the reference image after the image is corrected using the height value with error, the specific value of the height error can be calculated using the difference between them.

[0069] In step S120, the distances between each Region of Interest (ROI) in the down-view real-time image are first calculated. Then, the distances between each ROI in the reference image are calculated. Next, the ratio of the distances between corresponding two ROIs is calculated; this difference contains height error information, and this information is linearly related to the height error. Therefore, using this error information, the height error can be estimated, such as... Figure 7 As shown.

[0070] exist Figure 7 In the diagram, A and B are any two Regions of Interest (ROIs) in the baseline image, and A′ and B′ are the matched positions of ROI-A and ROI-B in the corrected real-time image. If the altitude information has no error, i.e., H′ = H, then A and A′ coincide, and B and B′ coincide, where H′ represents the altitude information obtained from the inertial navigation system. However, when there is an error in the altitude information, L... AB With L A′B′ The ratio is equal to the ratio of H′ to H:

[0071]

[0072] In formula (1), L AB L A′B′ Since H′ is a known value, the altitude error (i.e., the first estimated altitude error) dH = HH′ can be calculated using the following formula:

[0073]

[0074] Specifically, when there are three or more regions of interest, multiple first estimated height errors can be calculated for each real-time image frame, and their average value is used as the first estimated height error corresponding to each real-time image frame.

[0075] Furthermore, the estimated first altitude error must be limited to a range that does not exceed the inherent error of the entire inertial navigation system. Therefore, when there are three or more regions of interest, before calculating the average of multiple first altitude errors, it is necessary to judge each first altitude error based on the inherent error of the inertial navigation system, and discard any first altitude errors that exceed the inherent error, excluding them from the calculation of the average.

[0076] Similarly, when there are two regions of interest, only one estimated height error can be calculated for each real-time image frame. Therefore, the first estimated height error corresponding to each real-time image frame is judged based on the inherent error of the inertial navigation system, and the first estimated height error that is greater than the inherent error is eliminated.

[0077] In step S130, in specific applications, the height estimation process requires strict constraints, therefore a confidence level for the height estimation needs to be set. When there are two regions of interest, the confidence level is 0.65; when there are three or more regions of interest, the confidence level is 0.75. The first estimated height error calculated for each frame of real-time image is multiplied by the corresponding confidence level to obtain the second estimated height error.

[0078] In step S140, after calculating the average altitude error of the current frame (i.e., the second estimated altitude error), it cannot be directly used as the altitude estimation error of the entire system. It needs to be placed into a circular queue of length 16. Then, every 6 altitude errors are counted, the average of 6 consecutive frames is calculated again, and this average is used as the altitude estimation error of the entire system to correct the altitude information in the inertial navigation system. This process is repeated to complete the altitude correction throughout the entire process.

[0079] In this embodiment, when correcting the altitude information in the inertial navigation system based on the altitude error, the correction is performed gradually, not all at once. Figure 8-9 .

[0080] In this paper, experimental verification was also conducted based on the proposed method. By adding ±10% altitude error to the input inertial navigation altitude information, that is, adding +300m and -300m errors when the altitude is 3000m, the altitude error was corrected using the present invention. The results are shown in Figures 8-9. As can be seen from the figures, the proposed method can accurately estimate the ±10% altitude error within 10 frames.

[0081] like Figure 10-13 As shown, the experimental results of this method are further demonstrated. Figure 10 Given two sets of real-time input images, Figure 11 This is a schematic diagram of the recognition results when the height error between two sets of real-time images is -300m. Figure 12 This is a schematic diagram of the recognition results when the height error between two sets of real-time images is +300m. Figure 13 This is a schematic diagram of the recognition results after height error correction for two sets of real-time images.

[0082] In the aforementioned inertial navigation information correction method based on multi-sub-region prior information, a real-time image dataset consisting of multiple frames of real-time images of ground targets acquired from high altitudes, and a reference image containing accurate information about the location of ground targets, is obtained. Geometric correction is performed on each frame of the real-time image to obtain a corrected real-time image. Multiple regions of interest (ROIs) are selected in the reference image, and these ROIs are located in the corrected real-time image using a matching method to obtain corresponding ROIs. An estimated height error is calculated based on the distance between the centers of each ROI in the reference image and the distance between the centers of the corresponding ROIs in the corrected real-time image. The height information in the inertial navigation information is then corrected based on this estimated height error, thereby improving the environmental adaptability and positioning accuracy of the target recognition system.

[0083] It should be understood that, although Figure 1 The steps in the flowchart 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 herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 1 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed 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.

[0084] In one embodiment, such as Figure 14 As shown, an inertial navigation information correction device based on multi-sub-region prior information is provided, including: an image data acquisition module 200, a region of interest matching module 210, a first estimated height error calculation module 220, a second estimated height error calculation module 230, and an inertial navigation information correction module 240, wherein:

[0085] The image data acquisition module 200 is used to acquire a real-time image dataset and a reference image. The real-time image dataset contains multiple frames of real-time images acquired from high altitudes of ground targets, and the reference image contains accurate information about the location of the ground targets.

[0086] The region of interest matching module 210 is used to perform geometric correction on each frame of the real-time image to obtain a corrected real-time image, select multiple regions of interest in the reference image, and locate the multiple regions of interest in the corrected real-time image based on the matching method to obtain the corresponding multiple regions of interest.

[0087] The first estimated height error calculation module 220 is used to calculate the first estimated height error based on the distance between the centers of each region of interest in the reference image and the distance between the centers of the corresponding regions of interest in the corrected real-time image.

[0088] The second estimated height error calculation module 230 is used to multiply the first estimated height error according to the corresponding confidence level to obtain the second estimated height error corresponding to the real-time image.

[0089] The inertial navigation information correction module 240 is used to calculate the average value of the second estimated altitude error for every 6 frames of real-time images, and use the average value as the altitude error. Based on the altitude error, the altitude information in the inertial navigation information is gradually corrected until the altitude error converges.

[0090] Specific limitations regarding the inertial navigation information correction device based on multi-sub-region prior information can be found in the limitations of the inertial navigation information correction method based on multi-sub-region prior information mentioned above, and will not be repeated here. Each module in the aforementioned inertial navigation information correction device based on multi-sub-region prior information can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in hardware or independently of the processor in a computer device, or stored in software in the memory of a computer device, so that the processor can call and execute the operations corresponding to each module.

[0091] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 15 As shown, the computer device includes a processor, memory, network interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used to communicate with external terminals via a network connection. When executed by the processor, the computer program implements an inertial navigation information correction method based on multi-sub-region prior information. The display screen can be a liquid crystal display (LCD) or an e-ink display. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device casing, or an external keyboard, touchpad, or mouse.

[0092] Those skilled in the art will understand that Figure 15The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0093] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0094] Acquire a real-time image dataset and a reference image, wherein the real-time image dataset contains multiple frames of real-time images of ground targets acquired from high altitude, and the reference image contains accurate information about the location of the ground targets;

[0095] Geometric correction is performed on each frame of the real-time image to obtain a corrected real-time image. Multiple regions of interest are selected in the reference image, and the multiple regions of interest are located in the corrected real-time image based on a matching method to obtain the corresponding multiple regions of interest.

[0096] The first estimated height error is calculated based on the distance between the centers of each region of interest in the reference image and the distance between the centers of the corresponding regions of interest in the corrected real-time image.

[0097] The first estimated height error is multiplied by the corresponding confidence level to obtain the second estimated height error corresponding to the real-time image;

[0098] The average value of the second estimated altitude error for every 6 frames of real-time images is calculated, and this average value is used as the altitude error. The altitude information in the inertial navigation information is gradually corrected based on the altitude error until the altitude error converges.

[0099] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0100] Acquire a real-time image dataset and a reference image, wherein the real-time image dataset contains multiple frames of real-time images of ground targets acquired from high altitude, and the reference image contains accurate information about the location of the ground targets;

[0101] Geometric correction is performed on each frame of the real-time image to obtain a corrected real-time image. Multiple regions of interest are selected in the reference image, and the multiple regions of interest are located in the corrected real-time image based on a matching method to obtain the corresponding multiple regions of interest.

[0102] The first estimated height error is calculated based on the distance between the centers of each region of interest in the reference image and the distance between the centers of the corresponding regions of interest in the corrected real-time image.

[0103] The first estimated height error is multiplied by the corresponding confidence level to obtain the second estimated height error corresponding to the real-time image;

[0104] The average value of the second estimated altitude error for every 6 frames of real-time images is calculated, and this average value is used as the altitude error. The altitude information in the inertial navigation information is gradually corrected based on the altitude error until the altitude error converges.

[0105] 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, and when executed, it 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), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0106] 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.

[0107] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, 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, and these all 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 method for correcting inertial navigation information based on multi-subregion prior information, characterized in that, The method comprises: acquiring a real-time image data set and a reference image, the real-time image data set comprising a plurality of real-time images acquired by an aerial vehicle from a ground target, and the reference image comprising accurate information of a position of the ground target; performing geometric correction on each of the real-time images to obtain corrected real-time images, selecting a plurality of regions of interest in the reference image, and locating the plurality of regions of interest in the corrected real-time images based on a matching method to obtain corresponding regions of interest; calculating a first estimated height error based on distances between centers of the regions of interest in the reference image and distances between centers of the corresponding regions of interest in the corrected real-time images; multiplying the first estimated height error by a corresponding confidence to obtain a second estimated height error corresponding to the real-time images; calculating an average of the second estimated height errors of every six real-time images, and taking the average as a height error, and gradually correcting height information in inertial navigation information based on the height error until the height error converges.

2. The inertial navigation information correction method according to claim 1, characterized by, The regions of interest comprise two or more than or equal to three.

3. The inertial navigation information correction method according to claim 2, characterized by, When the regions of interest are more than or equal to three, a plurality of first estimated height errors can be calculated for each real-time image, and an average of the first estimated height errors is taken as a first estimated height error corresponding to each real-time image.

4. The inertial navigation information correction method according to claim 3, characterized by, Before calculating the average of the first estimated height errors, each of the first estimated height errors is further judged based on an inherent error of an inertial navigation system, and if the first estimated height error is greater than the inherent error, the first estimated height error is removed.

5. The inertial navigation information correction method according to claim 2, characterized by, When the regions of interest are two, each of the first estimated height errors corresponding to each real-time image is judged based on the inherent error of the inertial navigation system, and if the first estimated height error is greater than the inherent error, the first estimated height error is removed.

6. The inertial navigation information correction method according to claim 2, characterized by, When the regions of interest are two, the confidence is 0.65, and when the regions of interest are more than or equal to three, the confidence is 0.

75.

7. The inertial navigation information correction method according to any one of claims 1 to 6, characterized in that, The first estimated height error is calculated based on distances between centers of the regions of interest in the reference image and distances between centers of the corresponding regions of interest in the corrected real-time images using the following formula: In the above formula, L AB represents the distance between the centers of the regions of interest A and B in the reference image, L A′B′ represents the distance between the centers of the regions of interest A' and B' corresponding to the regions of interest A and B in the corrected real-time image, and H' represents the height information obtained by the inertial navigation system.

8. The inertial navigation information correction device based on multi-subregion prior information, characterized in that, The device comprises: an image data acquisition module configured to acquire a real-time image data set and a reference image, the real-time image data set comprising a plurality of real-time images acquired by an aerial vehicle from a ground target, and the reference image comprising accurate information of a position of the ground target; a region of interest matching module configured to perform geometric correction on each of the real-time images to obtain corrected real-time images, select a plurality of regions of interest in the reference image, and locate the plurality of regions of interest in the corrected real-time images based on a matching method to obtain corresponding regions of interest; a first estimated height error calculation module configured to calculate a first estimated height error based on distances between centers of the regions of interest in the reference image and distances between centers of the corresponding regions of interest in the corrected real-time images; and a second estimated height error calculation module configured to multiply the first estimated height error by a corresponding confidence to obtain a second estimated height error corresponding to the real-time images. a second estimated height error calculation module, configured to multiply the first estimated height error according to the corresponding confidence to obtain a second estimated height error corresponding to the real-time image; an inertial navigation information correction module, configured to calculate an average value of the second estimated height error of every 6 real-time images, and take the average value as a height error, and gradually correct height information in inertial navigation information according to the height error until the height error converges. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-8 when the computer program is executed by the processor. The processor implements the steps of the method in claim 7 when executing the computer program.

Images