Visual navigation methods, devices, equipment and storage media

Abstract

This invention discloses a visual navigation method, apparatus, device, and storage medium, belonging to the field of image processing technology. During visual navigation, this invention acquires detected feature point data; performs symmetry processing on the feature point data to obtain processed feature point data; performs distortion correction processing on the processed feature point data to obtain target feature data; and performs visual navigation based on the target feature data. By performing symmetry processing on the feature point data, this invention achieves multi-scene coverage of visual odometry, solving the problem of visual odometry failure caused by the limited range of visual sensors. This enables the visual odometry to output stably in different scenarios, improving its applicability and enhancing the visual navigation effect.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and in particular to a visual navigation method, apparatus, device, and storage medium. Background Technology

[0002] In common visual odometry applications, a symmetrical design is typically used (all visual sensors are consistent). That is, regardless of whether it's monocular, binocular, or multi-view, the sensors and lenses have the same technical parameters. This is done to: 1. facilitate manufacturing; and 2. facilitate left-right image synchronization and output quality control (exposure, noise, grayscale consistency, etc.). Asymmetrical sensor combinations are typically found in autonomous driving perception applications (e.g., short focal length with a large field of view to cover a wider area, combined with a long focal length with a small field of view to perceive sufficiently long distances).

[0003] While symmetrical and consistent sensor architecture design is simple and convenient, it can also suffer from insufficient scene coverage. For example, because the focal length and field of view are unique, it cannot simultaneously cover targets that are too close or too far away (targets that are too close will become out of focus and blur, causing feature detection to fail). Furthermore, if a zoom structure is used, it is difficult to meet airborne specifications for structural stability. Summary of the Invention

[0004] The main objective of this invention is to provide a visual navigation method, apparatus, device, and storage medium, aiming to solve the technical problem of low applicability of existing visual odometry.

[0005] To achieve the above objectives, the present invention provides a visual navigation method, the method comprising the following steps:

[0006] During visual navigation, acquire the detected feature point data;

[0007] The feature point data is symmetrically processed to obtain the processed feature point data;

[0008] The processed feature point data is then subjected to distortion correction to obtain the target feature data.

[0009] Visual navigation is performed based on the target feature data.

[0010] Optionally, the step of symmetricizing the feature point data to obtain processed feature point data includes:

[0011] The object state is determined based on the feature point data;

[0012] Set empirical coefficients based on the state of the object;

[0013] The feature point data is symmetrically processed using the empirical coefficients to obtain the processed feature point data.

[0014] Optionally, setting the empirical coefficient based on the object's state includes:

[0015] The number of left and right co-view feature point pairs is obtained based on the feature point data.

[0016] When the object is in motion, a first constant value and a first ground height are obtained;

[0017] An empirical coefficient is calculated based on the first constant value, the number of left and right co-view feature point pairs, and the first ground height.

[0018] Optionally, setting the empirical coefficient based on the object's state includes:

[0019] When the object is in a stationary state, the second ground height is obtained;

[0020] Compare the second ground height with a preset height threshold;

[0021] When the second ground height is less than the preset height threshold, the empirical coefficient is set to a second constant value.

[0022] Optionally, after comparing the second ground height with a preset height threshold, the method further includes:

[0023] When the second ground height is greater than or equal to the preset height threshold, a third constant value is obtained;

[0024] The empirical coefficient is calculated based on the third constant value, the number of left and right co-view feature point pairs, and the second ground height.

[0025] Optionally, the step of symmetricizing the feature point data using the empirical coefficients to obtain processed feature point data includes:

[0026] Obtain the dimensions of the captured image;

[0027] The region of interest is calculated using the empirical coefficients and the image size;

[0028] Set an image mask according to the region of interest;

[0029] The feature point data is symmetrically processed based on the image mask to obtain the processed feature point data.

[0030] Optionally, acquiring the detected feature point data during visual navigation includes:

[0031] During visual navigation, acquire image pairs data from captured images;

[0032] Feature point detection is performed on the image data to obtain feature point data.

[0033] Furthermore, to achieve the above objectives, the present invention also proposes a visual navigation device, the visual navigation device comprising:

[0034] The acquisition module is used to acquire detected feature point data during visual navigation.

[0035] The processing module is used to perform symmetry processing on the feature point data to obtain processed feature point data;

[0036] The processing module is also used to perform distortion correction processing on the processed feature point data to obtain target feature data;

[0037] The calculation module is used for visual navigation based on the target feature data.

[0038] Furthermore, to achieve the above objectives, the present invention also proposes a visual navigation device, the visual navigation device comprising: a memory, a processor, and a visual navigation program stored in the memory and executable on the processor, the visual navigation program being configured to implement the steps of the visual navigation method as described above.

[0039] In addition, to achieve the above objectives, the present invention also proposes a storage medium storing a visual navigation program, which, when executed by a processor, implements the steps of the visual navigation method as described above.

[0040] This invention acquires detected feature point data during visual navigation; performs symmetry processing on the feature point data to obtain processed feature point data; performs distortion correction processing on the processed feature point data to obtain target feature data; and performs visual navigation based on the target feature data. By performing symmetry processing on the feature point data, it achieves multi-scene coverage of visual odometry and solves the problem of visual odometry failure caused by the limited range of visual sensors. This enables the visual odometry to output stably in different scenarios, improves the applicability of visual odometry, and enhances the visual navigation effect. Attached Figure Description

[0041] Figure 1 This is a schematic diagram of the structure of a visual navigation device in the hardware operating environment involved in the embodiments of the present invention;

[0042] Figure 2 This is a flowchart illustrating the first embodiment of the visual navigation method of the present invention;

[0043] Figure 3 This is a schematic diagram showing the out-of-focus and blurred central region of an ultra-wide-angle camera in one embodiment of the visual navigation method of the present invention;

[0044] Figure 4 This is a schematic diagram of conventional camera shooting in one embodiment of the visual navigation method of the present invention;

[0045] Figure 5 This is a schematic diagram of the field of view after combining a conventional static lens and a wide-angle lens in one embodiment of the visual navigation method of the present invention;

[0046] Figure 6 This is a schematic diagram of two feature points detected in the original image in one embodiment of the visual navigation method of the present invention;

[0047] Figure 7 This is a schematic diagram of two feature points detected after distortion removal processing in one embodiment of the visual navigation method of the present invention;

[0048] Figure 8 This is a flowchart illustrating the second embodiment of the visual navigation method of the present invention;

[0049] Figure 9 This is a flowchart illustrating the third embodiment of the visual navigation method of the present invention;

[0050] Figure 10 This is a flowchart illustrating the fourth embodiment of the visual navigation method of the present invention;

[0051] Figure 11 This is a schematic diagram of the front-end processing flow of visual navigation in one embodiment of the visual navigation method of the present invention;

[0052] Figure 12 This is a structural block diagram of the first embodiment of the visual navigation device of the present invention.

[0053] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0054] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0055] Reference Figure 1 , Figure 1 This is a schematic diagram of the structure of a visual navigation device in the hardware operating environment involved in the embodiments of the present invention.

[0056] like Figure 1As shown, the visual navigation device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen or an input unit such as a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wireless-Fidelity (Wi-Fi) interface). The memory 1005 may be high-speed random access memory (RAM) or stable non-volatile memory (NVM), such as a disk storage device. Optionally, the memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0057] Those skilled in the art will understand that Figure 1 The structure shown does not constitute a limitation on the visual navigation device and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0058] like Figure 1 As shown, the memory 1005, which serves as a storage medium, may include an operating system, a network communication module, a user interface module, and a visual navigation program.

[0059] exist Figure 1 In the visual navigation device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the visual navigation device of the present invention can be set in the visual navigation device, and the visual navigation device calls the visual navigation program stored in the memory 1005 through the processor 1001 and executes the visual navigation method provided in the embodiment of the present invention.

[0060] This invention provides a visual navigation method, referring to... Figure 2 , Figure 2 This is a flowchart illustrating the first embodiment of the visual navigation method of the present invention.

[0061] In this embodiment, the visual navigation method includes the following steps:

[0062] Step S10: During visual navigation, acquire the detected feature point data.

[0063] It should be noted that the executing entity in this embodiment can be a visual navigation device, or other devices that can achieve the same or similar functions. This embodiment does not limit this; this embodiment uses a visual navigation device as an example for explanation.

[0064] In practical implementation, addressing the issue of insufficient scene coverage by existing symmetrical and consistent sensors, this embodiment proposes a non-consistent sensor architecture for visual odometry. By utilizing combinations of different focal lengths, field of view, and resolutions, it achieves multi-scene coverage for visual odometry while ensuring sufficient shared viewing area among different cameras. Furthermore, special adaptations have been made for asymmetrical sensor combinations, ensuring the entire system maintains stability and robustness while meeting scene coverage requirements.

[0065] It should be noted that, as Figure 3 As shown, Figure 3 This diagram illustrates the out-of-focus blurring in the central area (within the circle) of an ultra-wide-angle camera. Taking a stereo camera as an example (camera facing downwards towards the ground), for an ultra-wide-angle lens, the area near the center is too close to the target (the ground) (approximately 10cm above the ground), causing it to lose focus. This results in blurring in that area, leading to the failure of feature point detection. This is because the surrounding field of view captures visual information from further distances, and this information falls within the camera's focal length. Figure 3 As shown, Figure 4 This is a schematic diagram taken with a conventional camera. In this scenario, the conventional lens is completely out of focus, and the entire image is blurred, making it impossible to extract feature points.

[0066] While wide-angle lenses can cover a wider field of view, with the sensor resolution remaining the same, the large FOV means the image cannot display more detail (a single pixel represents a larger physical area). In other words, compared to a regular lens, the viewing distance is shorter, and the ranging capability is reduced. So, if a regular lens and a wide-angle lens are combined, such as... Figure 5 As shown, Figure 5 This diagram illustrates the field of view of a combination of a standard static lens and a wide-angle lens, enabling coverage of "too close" scenes while maintaining the viewing distance capability of a standard lens. Furthermore, even with two identical standard lenses and a standard baseline (the distance between the optical centers of the two cameras, typically less than 20cm), the ranging accuracy decreases as the viewing distance (the distance of the object from the camera) increases, and at high altitudes, it degenerates directly into monocular mode.

[0067] In practical implementation, during visual navigation, feature point data detected by visual sensors can be acquired. Visual sensors can be combinations of cameras with different fields of view, thus forming a heterogeneous visual odometry. For example, visual sensors can be binocular or multi-camera combinations, and the parameters of the cameras can be different, such as camera sensor model, sensor resolution, lens FOV, lens focal length, and lens aperture.

[0068] Furthermore, the steps for acquiring detected feature point data during visual navigation specifically include: acquiring captured image pairs during visual navigation; and performing feature point detection on the image pairs to obtain feature point data.

[0069] In practical implementation, when performing visual navigation, multiple visual sensors can be used to capture image data to obtain image pairs. Feature point data can then be obtained by detecting and tracking feature points in the image pairs.

[0070] Step S20: Perform symmetry processing on the feature point data to obtain processed feature point data.

[0071] Understandably, since the parameters and distortion of lenses with different fields of view are different, if the conventional symmetrical binocular processing method is used, the overall stability of visual odometry will be problematic and the accuracy will not be guaranteed. Therefore, the obtained feature point data can be further symmetrically processed to solve the problem of left-right asymmetry in feature point fluctuations.

[0072] Understandably, the fluctuations in feature points are caused by distortion from the ultra-wide-angle lens. The greater the distortion, the more pronounced the accuracy fluctuations. When the original image is severely distorted, after distortion correction, the pixels closer to the camera are significantly stretched. The error fluctuations of these feature points are also relatively large. For example... Figure 6 As shown, Figure 6 This is a schematic diagram of the two feature points detected in the original image. Figure 7 This image shows two feature points detected after distortion correction. After distortion correction, they become A' and B' (the pixels between them are obtained through interpolation). This makes these points extremely sensitive to the accuracy of feature point detection. A feature point deviation of 0.1 pixels in the original image may become a deviation of 5-10 pixels after distortion correction.

[0073] Therefore, for wide-field lenses, different processing methods should be applied according to the different conditions of the carrier to achieve symmetry between the left and right cameras, thereby improving the effect of subsequent distortion correction.

[0074] It should be noted that the feature point data is symmetrically processed, for example, by symmetric processing based on the number of consensus points in the feature point data, thus obtaining the processed feature point data.

[0075] Step S30: Perform distortion correction on the processed feature point data to obtain the target feature data.

[0076] In practice, after obtaining the processed feature point data, distortion correction can be performed on the processed feature point data to obtain the target feature data.

[0077] Step S40: Perform visual navigation based on the target feature data.

[0078] It is understandable that after obtaining the target feature data, visual navigation can be performed using the target feature data, thereby solving the problem of visual odometry failure caused by blurry images in small fields of view and improving the effectiveness of visual navigation.

[0079] This embodiment acquires detected feature point data during visual navigation; performs symmetry processing on the feature point data to obtain processed feature point data; performs distortion correction processing on the processed feature point data to obtain target feature data; and performs visual navigation based on the target feature data. By performing symmetry processing on the feature point data, it achieves multi-scene coverage of visual odometry, solving the problem of visual odometry failure caused by the limited range of visual sensors. This enables the visual odometry to output stably in different scenarios, improving its applicability and enhancing the visual navigation effect.

[0080] refer to Figure 8 , Figure 8 This is a flowchart illustrating the second embodiment of the visual navigation method of the present invention.

[0081] Based on the first embodiment described above, step S20 of the visual navigation method in this embodiment includes:

[0082] Step S201: Determine the object state based on the feature point data.

[0083] It should be noted that when performing symmetry processing on feature point data, different processing methods can be applied depending on the state of the object. Therefore, the state of the object can be determined based on the feature point data.

[0084] It is understandable that an object can be in a static state or a dynamic state.

[0085] Step S202: Set empirical coefficients based on the state of the object.

[0086] In practice, different empirical coefficients can be set for different object states, thereby performing corresponding symmetry processing on the feature point data.

[0087] Step S203: The feature point data is symmetrically processed using the empirical coefficients to obtain the processed feature point data.

[0088] It should be noted that feature point data can be symmetrically processed by setting empirical coefficients to obtain processed feature point data.

[0089] By symmetrically processing the feature point data, the additional information brought by the wide-field-of-view camera can be aligned with that of a conventional field-of-view lens. Furthermore, the information input can be dynamically adjusted according to the scene to reduce the impact of distortion.

[0090] Further, the step of symmetricizing the feature point data using the empirical coefficients to obtain processed feature point data includes: acquiring the captured image size; calculating the region of interest using the empirical coefficients and the image size; setting an image mask based on the region of interest; and symmetricizing the feature point data using the image mask to obtain processed feature point data.

[0091] It should be noted that after obtaining the empirical coefficients, the size of the captured image is obtained, and then the region of interest is calculated using the image size and the empirical coefficients. The calculation process of the region of interest is as follows: Equation 1:

[0092] ROI = alpha * img (Equation 1)

[0093] In Equation 1 above, ROI is the region of interest, alpha is the empirical coefficient, and img is the image size.

[0094] In practice, an image mask can be set according to the calculated region of interest. The selection of the actual feature point detection region can be achieved by relying on the image mask, thereby symmetric processing of the feature point data to obtain the processed feature point data.

[0095] This embodiment determines the object state based on the feature point data; sets empirical coefficients based on the object state; and performs symmetry processing on the feature point data using the empirical coefficients to obtain processed feature point data. Corresponding empirical coefficients can be set according to different object states, thereby performing symmetry processing on the feature point data using empirical coefficients, dynamically adjusting information input according to the scene, and reducing the impact of distortion.

[0096] refer to Figure 9 , Figure 9 This is a flowchart illustrating the third embodiment of the visual navigation method of the present invention.

[0097] Based on the second embodiment described above, step S202 of the visual navigation method in this embodiment includes:

[0098] Step S2021: Obtain the number of left and right co-view feature point pairs based on the feature point data.

[0099] It should be noted that the number of left and right co-view feature point pairs can be obtained from the feature point data.

[0100] In specific implementation, the number of left and right co-view feature point pairs can be set to around 100, such as 98 or 102. This embodiment does not impose any restrictions on this.

[0101] Step S2022: When the object is in motion, obtain the first constant value and the first ground height.

[0102] It should be noted that it can detect whether the object is in a stationary state. If the object is in a non-stationary state, i.e. in motion, the first constant value and the first ground height can be obtained directly.

[0103] The first ground height is the height of the object above the ground when it is in motion. The first constant value can be set according to requirements, for example, to 500.

[0104] Step S2023: Calculate the empirical coefficient based on the first constant value, the number of left and right co-view feature point pairs, and the first ground height.

[0105] It should be noted that the empirical coefficient can be calculated based on the first constant value, the number of left and right co-view feature point pairs, and the first ground height. The empirical coefficient alpha = 500 / [k*sqrt(h1)], where k is the number of left and right co-view feature point pairs, h1 is the first ground height, and 500 is the first constant value.

[0106] This embodiment obtains the number of left and right co-view feature point pairs based on the feature point data; when the object is in motion, it obtains a first constant value and a first ground height; based on the first constant value, the number of left and right co-view feature point pairs, and the first ground height, it calculates an empirical coefficient, which can quickly determine the first constant value and the first ground height according to the object's motion state, thereby flexibly calculating the empirical coefficient.

[0107] refer to Figure 10 , Figure 10 This is a flowchart illustrating the fourth embodiment of the visual navigation method of the present invention.

[0108] Based on the second embodiment described above, step S202 of the visual navigation method in this embodiment includes:

[0109] Step S2021': When the object is in a stationary state, obtain the second ground height.

[0110] It should be noted that when the object is in a stationary state, the second ground height h2 when the object is stationary can be obtained.

[0111] Step S2022': Compare the second ground height with a preset height threshold.

[0112] In practice, the preset height threshold can be set to 1m. By comparing the second ground height with the preset height threshold, a corresponding empirical coefficient can be set.

[0113] Step S2023': When the second ground height is less than the preset height threshold, set the empirical coefficient to the second constant value.

[0114] It should be noted that when the second ground height is less than the preset height threshold of 1m, it can be rounded down directly. For example, the empirical coefficient can be set as the second constant value, which can be set to 1.

[0115] Optionally, after comparing the second ground height with a preset height threshold, the method further includes:

[0116] When the second ground height is greater than or equal to the preset height threshold, a third constant value is obtained;

[0117] The empirical coefficient is calculated based on the third constant value, the number of left and right co-view feature point pairs, and the second ground height.

[0118] It should be noted that when the second ground height is greater than or equal to the preset height threshold, a pre-set third constant value can be obtained.

[0119] The third constant value can be set to 1000, so that the empirical coefficient can be calculated based on the third constant value, the number of left and right co-view feature point pairs, and the second ground height. The empirical coefficient alpha = 1000 / [k*sqrt(h2)], where k is the number of left and right co-view feature point pairs, h2 is the second ground height, and 1000 is the second constant value.

[0120] It should be noted that when the object is at a high altitude, the empirical coefficient is taken as the minimum value of the center half of the wide-angle region.

[0121] like Figure 11 As shown, Figure 11 This is a schematic diagram of the front-end processing flow of visual navigation. The input image pairs are used to detect and track feature points to obtain feature point data. The feature point data is then symmetrically processed to obtain processed feature point data. The processed feature point data is then dedistorted, including the parameters of the left and right sensors, to obtain target feature data. The target feature data is then packaged and output to perform visual navigation based on the target feature data.

[0122] In this embodiment, when the object is in a stationary state, a second ground height is obtained; the second ground height is compared with a preset height threshold; when the second ground height is less than the preset height threshold, an empirical coefficient is set as a second constant value. This allows for quick determination of the second constant value based on the object's motion state, thus enabling flexible setting of the empirical coefficient.

[0123] Reference Figure 12 , Figure 12 This is a structural block diagram of the first embodiment of the visual navigation device of the present invention.

[0124] like Figure 12 As shown, the visual navigation device proposed in this embodiment of the invention includes:

[0125] The acquisition module 10 is used to acquire the detected feature point data during visual navigation.

[0126] The processing module 20 is used to perform symmetry processing on the feature point data to obtain processed feature point data.

[0127] The processing module 20 is also used to perform distortion correction processing on the processed feature point data to obtain target feature data.

[0128] The calculation module 30 is used for visual navigation based on the target feature data.

[0129] This embodiment acquires detected feature point data during visual navigation; performs symmetry processing on the feature point data to obtain processed feature point data; performs distortion correction processing on the processed feature point data to obtain target feature data; and performs visual navigation based on the target feature data. By performing symmetry processing on the feature point data, multi-scene coverage of the visual odometry is achieved, solving the problem of visual odometry failure caused by the limited range of the visual sensor. This enables the visual odometry to output stably in different scenarios, improving its applicability.

[0130] In one embodiment, the processing module 20 is further configured to determine the object state based on the feature point data; set an empirical coefficient based on the object state; and perform symmetry processing on the feature point data using the empirical coefficient to obtain processed feature point data.

[0131] In one embodiment, the processing module 20 is further configured to obtain the number of left and right co-view feature point pairs based on the feature point data; when the object is in motion, obtain a first constant value and a first ground height; and calculate an empirical coefficient based on the first constant value, the number of left and right co-view feature point pairs, and the first ground height.

[0132] In one embodiment, the processing module 20 is further configured to: obtain a second ground height when the object is in a stationary state; compare the second ground height with a preset height threshold; and set an empirical coefficient to a second constant value when the second ground height is less than the preset height threshold.

[0133] In one embodiment, the processing module 20 is further configured to obtain a third constant value when the second ground height is greater than or equal to the preset height threshold; and calculate an empirical coefficient based on the third constant value, the number of left and right co-view feature point pairs, and the second ground height.

[0134] In one embodiment, the processing module 20 is further configured to acquire the size of the captured image; calculate the region of interest using the empirical coefficient and the image size; set an image mask based on the region of interest; and perform symmetry processing on the feature point data based on the image mask to obtain processed feature point data.

[0135] In one embodiment, the acquisition module 10 is further configured to acquire captured image pair data; and perform feature point detection on the image pair data to obtain feature point data.

[0136] Furthermore, embodiments of the present invention also propose a storage medium storing a visual navigation program, which, when executed by a processor, implements the steps of the visual navigation method described above.

[0137] Since this storage medium adopts all the technical solutions of all the above embodiments, it has at least all the beneficial effects brought about by the technical solutions of the above embodiments, which will not be repeated here.

[0138] It should be understood that the above are merely illustrative examples and do not constitute any limitation on the technical solutions of the present invention. In specific applications, those skilled in the art can make settings as needed, and the present invention does not impose any restrictions on this.

[0139] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of this invention. In practical applications, those skilled in the art can select some or all of the workflow to achieve the purpose of this embodiment according to actual needs, and no restrictions are imposed here.

[0140] In addition, for technical details not described in detail in this embodiment, please refer to the visual navigation method provided in any embodiment of the present invention, which will not be repeated here.

[0141] Furthermore, it should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0142] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0143] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as read-only memory (ROM) / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0144] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A visual navigation method, characterized in that, The visual navigation method includes: During visual navigation, acquire the detected feature point data; The object state is determined based on the feature point data; Set empirical coefficients based on the state of the object; Obtain the dimensions of the captured image; The region of interest is calculated using the empirical coefficients and the image size; Set an image mask according to the region of interest; The feature point data is symmetrically processed based on the image mask to obtain the processed feature point data. The processed feature point data is then subjected to distortion correction to obtain the target feature data. Visual navigation is performed based on the target feature data.

2. The visual navigation method as described in claim 1, characterized in that, The step of setting empirical coefficients based on the object's state includes: The number of left and right co-view feature point pairs is obtained based on the feature point data. When the object is in motion, a first constant value and a first ground height are obtained; An empirical coefficient is calculated based on the first constant value, the number of left and right co-view feature point pairs, and the first ground height.

3. The visual navigation method as described in claim 2, characterized in that, The step of setting empirical coefficients based on the object's state includes: When the object is in a stationary state, the second ground height is obtained; Compare the second ground height with a preset height threshold; When the second ground height is less than the preset height threshold, the empirical coefficient is set to a second constant value.

4. The visual navigation method as described in claim 3, characterized in that, After comparing the second ground height with the preset height threshold, the method further includes: When the second ground height is greater than or equal to the preset height threshold, a third constant value is obtained; The empirical coefficient is calculated based on the third constant value, the number of left and right co-view feature point pairs, and the second ground height.

5. The visual navigation method as described in any one of claims 1 to 4, characterized in that, The process of acquiring detected feature point data during visual navigation includes: During visual navigation, acquire image pairs data from captured images; Feature point detection is performed on the image data to obtain feature point data.

6. A visual navigation device, characterized in that, The visual navigation device includes: The acquisition module is used to acquire detected feature point data during visual navigation. The processing module is used to determine the object state based on the feature point data; set empirical coefficients based on the object state; acquire the captured image size; calculate the region of interest using the empirical coefficients and the image size; set an image mask based on the region of interest; and perform symmetry processing on the feature point data based on the image mask to obtain processed feature point data. The processing module is also used to perform distortion correction processing on the processed feature point data to obtain target feature data; The calculation module is used for visual navigation based on the target feature data.

7. A visual navigation device, characterized in that, The visual navigation device includes: a memory, a processor, and a visual navigation program stored in the memory and executable on the processor, the visual navigation program being configured to implement the visual navigation method as described in any one of claims 1 to 5.

8. A storage medium, characterized in that, The storage medium stores a visual navigation program, which, when executed by a processor, implements the visual navigation method as described in any one of claims 1 to 5.

Images