Map matching navigation method, device, equipment and storage medium based on decision analysis

Abstract

The present invention relates to the field of unmanned aerial vehicle navigation technology, and in particular to a map matching navigation method based on decision analysis. The method comprises the following steps: searching for corresponding environmental data in a preset map database according to the acquired positioning information; extracting an initial aerial image sequence from the acquired initial aerial video, performing lens distortion correction on the initial aerial image sequence, and obtaining a standard aerial image sequence; performing contour extraction on the standard aerial image sequence, and obtaining a standard environmental contour; performing overlap matching on the environmental data, and obtaining overlap environmental data; performing data correction on the overlap environmental data, and obtaining navigation environmental data; and performing position matching on the navigation environmental data using a preset displacement bag-of-words model. The present invention also proposes a map matching navigation device, an electronic device, and a storage medium based on decision analysis. The present invention can improve the accuracy of unmanned aerial vehicle map navigation.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a map matching navigation method, device, electronic device and computer-readable storage medium based on decision analysis. Background Art

[0002] With the popularization of drone technology, drones have been applied in more and more fields. However, in order to improve the operating experience of drones and reduce the manual operation of users, it is necessary to map the drones to facilitate user operation. Most of the existing drone navigation technologies are based on single features, such as inertial navigation that detects the motion trajectory of the drone and navigation through satellite positioning. In actual applications, the flight environment of drones is relatively complex. Navigation technology based on a single feature has great limitations in dealing with changing environments and has low robustness, which may lead to low accuracy in map navigation of drones. Summary of the invention

[0003] The present invention provides a map matching navigation method, device and computer-readable storage medium based on decision analysis, the main purpose of which is to solve the problem of low accuracy when performing map navigation on a drone.

[0004] To achieve the above object, the present invention provides a map matching navigation method based on decision analysis, comprising:

[0005] Obtaining positioning information uploaded by the positioning component of the drone, and searching for corresponding environmental data in a preset map database according to the positioning information;

[0006] Acquire an initial aerial video taken by a camera assembly of the drone, extract an initial aerial image sequence from the initial aerial video, and perform lens distortion correction on the initial aerial image sequence using a trained mean square correction model to obtain a standard aerial image sequence;

[0007] Using a preset contour extraction model to extract contours from the standard aerial image sequence to obtain a standard environment contour;

[0008] Using the standard environment contour to match the environment data for coincidence, to obtain coincident environment data, and using the standard environment contour to perform data correction on the coincident environment data, to obtain navigation environment data;

[0009] The motion data of the UAV is obtained, the navigation environment data is position-matched using a preset displacement bag-of-words model to obtain node positions, and the motion trajectory of the UAV is generated according to the motion data and the node positions using a preset fuzzy decision algorithm to complete map navigation.

[0010] Optionally, searching for corresponding environmental data in a preset map database according to the positioning information includes:

[0011] Extracting longitude and latitude information from the positioning information, and using the longitude and latitude information to retrieve corresponding terrain data in the map database;

[0012] Altitude information is extracted from the positioning information, and the terrain data is subjected to height mapping conversion using the altitude information to obtain environmental data.

[0013] Optionally, extracting an initial aerial image sequence from the initial aerial video includes:

[0014] Using a preset video conversion tool to convert the initial aerial video into a video frame sequence;

[0015] Marking video nodes of the video frame sequence at preset time intervals to obtain a video node set;

[0016] A video node is selected from the video node set one by one in time sequence as a target node, a video frame image corresponding to the target node is used as a target frame image, and the target frame images are arranged into the video frame sequence in time sequence.

[0017] Optionally, before the method uses the trained mean square correction model to perform lens distortion correction on the initial aerial image sequence to obtain a standard aerial image sequence, the method further includes:

[0018] Obtaining a preset normal sample atlas, and positioning and annotating the normal sample atlas to obtain a normal positioning atlas;

[0019] Using the camera assembly to shoot the normal positioning atlas to obtain a distorted positioning atlas;

[0020] Generate a normal coordinate set according to the positioning annotation points in the normal positioning atlas, and generate a distorted coordinate set according to the positioning annotation points in the distorted positioning atlas;

[0021] The distortion coordinate set is rectified using the following mean square correction algorithm to obtain a predicted coordinate set:

[0022]

[0023]

[0024] in, is the abscissa of the ith distorted coordinate in the distorted coordinate set, refers to the ordinate of the i-th distorted coordinate in the distorted coordinate set, X iis the horizontal coordinate of the i-th predicted coordinate in the predicted coordinate set, y i is the ordinate of the i-th predicted coordinate in the predicted coordinate set, i refers to the l-th in the predicted coordinate set and the distorted coordinate set, J1 is the first term of the Taylor series expansion, J2 is the second term of the Taylor series expansion, d refers to the internal parameter of the camera assembly, k1 is the first tangential parameter of the mean square correction algorithm, and K2 refers to the second tangential parameter of the mean square correction algorithm:

[0025] The distance difference between the normal coordinate set and the predicted coordinate set is calculated, and the parameters of the mean square correction algorithm are updated using the distance difference until the distance difference is less than a preset distance threshold, and the mean square correction algorithm after parameter update is used as the mean square correction model.

[0026] Optionally, the using the trained mean square correction model to perform lens distortion correction on the initial aerial image sequence to obtain a standard aerial image sequence includes:

[0027] Selecting one initial aerial picture in the initial aerial picture sequence as a target initial picture one by one;

[0028] Generate a corresponding unit matrix according to the image size of the target initial image, and calculate a mapping vector of the unit matrix using the mean square correction model;

[0029] Converting the target initial image into a target image matrix, multiplying the target image matrix by the mapping vector to obtain a target correction matrix, and converting the target correction matrix into a target correction image;

[0030] All the target corrected images are sorted according to the order of the initial aerial image sequence to obtain the standard aerial image sequence.

[0031] Optionally, the step of performing contour extraction on the standard aerial image sequence using a preset contour extraction model to obtain a standard environment contour includes:

[0032] Selecting one standard aerial picture in the standard aerial picture sequence as a target aerial picture one by one;

[0033] Performing grayscale processing on the target aerial image to obtain a target grayscale image;

[0034] Binarizing the target grayscale image to obtain a target binary image;

[0035] Performing image erosion on the target binary image using the contour extraction model to obtain a target eroded image;

[0036] The target erosion images are subjected to image denoising and texture enhancement respectively to obtain target environment contours, and all target environment contours are aggregated into a standard environment contour.

[0037] Optionally, performing overlap matching on the environmental data using the standard environmental profile to obtain overlap environmental data includes:

[0038] Determine the movement direction of the UAV according to the positioning information, and perform projection mapping on the environmental data in sequence according to the movement direction to obtain environmental projection data;

[0039] Performing contour extraction on the environment projection data to obtain an environment projection contour;

[0040] Extracting features from the standard environment contour to obtain a standard feature set, and performing normalization encoding on the standard feature set using a preset normalization function to obtain a standard feature encoding set;

[0041] Extracting features from the environmental projection contour to obtain an environmental feature set, and performing normalization encoding on the environmental feature set using the normalization function to obtain an environmental feature encoding set;

[0042] Selecting a first standard feature code from the standard feature code set as a target standard code, and calculating a Hamming distance between the target standard code and an environmental feature code in the environmental feature code set;

[0043] The environmental feature code whose Hamming distance is less than a preset Hamming threshold is selected as the target environmental code, and the environmental projection data corresponding to the target environmental code is used as the coincidence environmental data.

[0044] In order to solve the above problems, the present invention also provides a map matching navigation device based on decision analysis, the device comprising:

[0045] A positioning information module is used to obtain the positioning information uploaded by the positioning component of the drone, and search for corresponding environmental data in a preset map database according to the positioning information;

[0046] A video correction module is used to obtain an initial aerial video taken by the camera assembly of the drone, extract an initial aerial image sequence from the initial aerial video, and perform lens distortion correction on the initial aerial image sequence using a trained mean square correction model to obtain a standard aerial image sequence;

[0047] A contour extraction module, used to extract contours from the standard aerial image sequence using a preset contour extraction model to obtain a standard environment contour;

[0048] A data correction module, used to perform coincidence matching on the environmental data using the standard environmental profile to obtain coincident environmental data, and to perform data correction on the coincident environmental data using the standard environmental profile to obtain navigation environmental data;

[0049] The intelligent navigation module is used to obtain the motion data of the UAV, use the preset displacement bag-of-words model to match the position of the navigation environment data to obtain the node position, and use the preset fuzzy decision algorithm to generate the motion trajectory of the UAV according to the motion data and the node position to complete the map navigation, wherein the use of the preset displacement bag-of-words model to match the position of the navigation environment data to obtain the node position includes: using the displacement bag-of-words model to extract features of the navigation environment data to obtain the navigation bag features of the navigation environment data; using the following bag-of-words similarity formula to calculate the similarity between the navigation bag features and each bag-of-words feature in the preset bag-of-words feature library:

[0050]

[0051] Wherein, S refers to the similarity, is the preset adversarial coefficient, arccos is the inverse cosine function, refers to the navigation word bag feature, It refers to the bag-of-words features in the preset bag-of-words feature library, and T is the transposition symbol; the bag-of-words features whose similarity is less than a preset similarity threshold are selected in the preset bag-of-words feature library as target bag-of-words features, and the visual word set corresponding to the target bag-of-words features is used as the target word set; a bag-of-words tree model is constructed according to the target word set, and a loop detection is performed on the navigation environment data according to the bag-of-words tree model to obtain the node position.

[0052] In order to solve the above problem, the present invention further provides an electronic device, the electronic device comprising:

[0053] at least one processor; and,

[0054] a memory communicatively connected to the at least one processor; wherein,

[0055] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the above-mentioned map matching navigation method based on decision analysis.

[0056] In order to solve the above problems, the present invention also provides a computer-readable storage medium, in which at least one computer program is stored. The at least one computer program is executed by a processor in an electronic device to implement the above-mentioned map matching navigation method based on decision analysis.

[0057] The embodiment of the present invention can obtain the geographical environment information of the location of the UAV by searching for corresponding environmental data in a preset map database according to the positioning information, thereby performing positioning navigation according to the geographical environment information, and by using the trained mean square correction model to perform lens distortion correction on the initial aerial image sequence to obtain a standard aerial image sequence, it is possible to correct the distortion of the fisheye lens of the camera module, ensure the accuracy of the subsequent standard environment contour, and provide more accurate navigation for the UAV, and by using a preset contour extraction model to perform contour extraction on the standard aerial image sequence to obtain a standard environment contour, it is possible to effectively filter out redundant information of the image in the standard aerial image sequence, and perform image enhancement, and it is possible to effectively extract the image features of the scene described by the current UAV, so as to facilitate the subsequent coincidence matching of the environmental data. By using the standard environment contour to match the environment data for overlap, the overlap environment data is obtained, and the standard environment contour is used to correct the overlap environment data to obtain navigation environment data, so that the current environment model of the UAV can be determined, and the environment model after the flight route is simulated in combination with the environment data, thereby ensuring the accuracy of the navigation environment data, and by using the preset displacement word bag model to match the navigation environment data for position, the node position can be obtained, and the word bag tree model closest to the current environment data can be generated, and the path can be generated in the form of loop detection, thereby improving the efficiency of UAV navigation, and by using the positioning information, aerial video and motion data of the UAV, the positioning navigation and visual navigation can be integrated to provide more accurate navigation data for the UAV. Therefore, the map matching navigation method, device, electronic device and computer-readable storage medium based on decision analysis proposed in the present invention can solve the problem of low accuracy when performing map navigation on UAVs. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 A flowchart of a map matching navigation method based on decision analysis provided by an embodiment of the present invention;

[0059] Figure 2 A schematic diagram of a process for extracting an initial aerial image sequence provided by an embodiment of the present invention;

[0060] Figure 3 A schematic diagram of a process for calculating node positions provided by an embodiment of the present invention;

[0061] Figure 4A functional module diagram of a map matching navigation device based on decision analysis provided by an embodiment of the present invention;

[0062] Figure 5 A schematic diagram of the structure of an electronic device for implementing the map matching navigation method based on decision analysis provided by an embodiment of the present invention.

[0063] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0064] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0065] The embodiment of the present application provides a map matching navigation method based on decision analysis. The execution subject of the map matching navigation method based on decision analysis includes but is not limited to at least one of the electronic devices such as a server, a terminal, etc. that can be configured to execute the method provided by the embodiment of the present application. In other words, the map matching navigation method based on decision analysis can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc. The server can be an independent server, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (Content Delivery Network, CDN), and big data and artificial intelligence platforms.

[0066] Reference Figure 1 FIG. 1 is a flow chart of a map matching navigation method based on decision analysis provided by an embodiment of the present invention. In this embodiment, the map matching navigation method based on decision analysis includes:

[0067] S1. Obtaining positioning information uploaded by the positioning component of the drone, and searching for corresponding environmental data in a preset map database according to the positioning information;

[0068] In an embodiment of the present invention, the positioning component may be a GPS positioning component, a radio altimeter, an atmospheric data calculation component, or a Beidou navigation positioning component of the UAV.

[0069] In detail, the positioning information includes the real-time latitude and longitude information and real-time altitude information of the drone.

[0070] In detail, the map database may be a map database including terrain information, such as a Google map database or an Amap database.

[0071] In the embodiment of the present invention, searching for corresponding environmental data in a preset map database according to the positioning information includes:

[0072] Extracting longitude and latitude information from the positioning information, and using the longitude and latitude information to retrieve corresponding terrain data in the map database;

[0073] Altitude information is extracted from the positioning information, and the terrain data is subjected to height mapping conversion using the altitude information to obtain environmental data.

[0074] In the embodiment of the present invention, by searching for corresponding environmental data in a preset map database according to the positioning information, the geographical environment information of the location of the drone can be obtained, so that positioning navigation can be performed according to the geographical environment information.

[0075] S2, obtaining an initial aerial video taken by the camera assembly of the drone, extracting an initial aerial image sequence from the initial aerial video, and performing lens distortion correction on the initial aerial image sequence using a trained mean square correction model to obtain a standard aerial image sequence;

[0076] In the embodiment of the present invention, the camera component refers to a component in a drone used for shooting images, such as YT-A6500, GoPro, Insta360 and other components.

[0077] In the embodiment of the present invention, referring to Figure 2 As shown, the extracting of the initial aerial image sequence from the initial aerial video includes:

[0078] S21, using a preset video conversion tool to convert the initial aerial video into a video frame sequence;

[0079] S22, marking video nodes of the video frame sequence at preset time intervals to obtain a video node set;

[0080] S23. Select a video node from the video node set as a target node one by one according to the time sequence, use the video frame image corresponding to the target node as the target frame image, and arrange the target frame images into the video frame sequence according to the time sequence.

[0081] Specifically, the video conversion tool may be non-linear special effects production software (AE, Adobe AfterEffects).

[0082] In detail, the time interval may be 0.5s or 1s.

[0083] Specifically, before the lens distortion correction of the initial aerial image sequence is performed using the trained mean square correction model to obtain a standard aerial image sequence, the method further includes:

[0084] Obtaining a preset normal sample atlas, and positioning and annotating the normal sample atlas to obtain a normal positioning atlas;

[0085] Using the camera assembly to shoot the normal positioning atlas to obtain a distorted positioning atlas;

[0086] Generate a normal coordinate set according to the positioning annotation points in the normal positioning atlas, and generate a distorted coordinate set according to the positioning annotation points in the distorted positioning atlas;

[0087] The distortion coordinate set is rectified using the following mean square correction algorithm to obtain a predicted coordinate set:

[0088]

[0089]

[0090] in, is the abscissa of the ith distorted coordinate in the distorted coordinate set, refers to the ordinate of the i-th distorted coordinate in the distorted coordinate set, X i is the horizontal coordinate of the i-th predicted coordinate in the predicted coordinate set, y i is the ordinate of the i-th predicted coordinate in the predicted coordinate set, i refers to the i-th in the predicted coordinate set and the distorted coordinate set, j1 is the first term of the Taylor series expansion, j2 is the second term of the Taylor series expansion, d refers to the internal parameter of the camera assembly, k1 is the first tangential parameter of the mean square correction algorithm, and k2 refers to the second tangential parameter of the mean square correction algorithm:

[0091] The distance difference between the normal coordinate set and the predicted coordinate set is calculated, and the parameters of the mean square correction algorithm are updated using the distance difference until the distance difference is less than a preset distance threshold, and the mean square correction algorithm after parameter update is used as the mean square correction model.

[0092] In detail, the normal sample atlas may be an atlas consisting of pure color pictures or an atlas consisting of pictures containing regular patterns.

[0093] Specifically, the positioning annotation refers to marking a positioning annotation point on each normal sample picture in the normal sample atlas.

[0094] In detail, the distance difference between the normal coordinate set and the predicted coordinate set may be calculated using a two-point distance formula.

[0095] Specifically, a gradient descent algorithm may be used to update parameters of the mean square correction algorithm according to the distance difference.

[0096] In detail, the distance threshold may be 20 pixels or 25 pixels.

[0097] In an embodiment of the present invention, the distortion coordinate set is corrected by the mean square correction algorithm to obtain a predicted coordinate set, which can accurately simulate the lens distortion of the camera component and correct it, so that the drone can obtain more accurate current environment data.

[0098] In detail, the method of using the trained mean square correction model to perform lens distortion correction on the initial aerial image sequence to obtain a standard aerial image sequence includes:

[0099] Selecting one initial aerial picture in the initial aerial picture sequence as a target initial picture one by one;

[0100] Generate a corresponding unit matrix according to the image size of the target initial image, and calculate a mapping vector of the unit matrix using the mean square correction model;

[0101] Converting the target initial image into a target image matrix, multiplying the target image matrix by the mapping vector to obtain a target correction matrix, and converting the target correction matrix into a target correction image;

[0102] All the target corrected images are sorted according to the order of the initial aerial image sequence to obtain the standard aerial image sequence.

[0103] In detail, the target initial image can be converted into a target image matrix, and the target correction matrix can be converted into a target correction image using the PIL toolkit of Python.

[0104] In an embodiment of the present invention, by using the trained mean square correction model to perform lens distortion correction on the initial aerial image sequence, a standard aerial image sequence is obtained, which can correct the distortion of the fisheye lens of the camera module, ensure the accuracy of the subsequent standard environment contour, and provide more accurate navigation for the drone.

[0105] S3, using a preset contour extraction model to perform contour extraction on the standard aerial image sequence to obtain a standard environment contour;

[0106] In the embodiment of the present invention, the method of extracting contours from the standard aerial image sequence using a preset contour extraction model to obtain a standard environment contour includes:

[0107] Selecting one standard aerial picture in the standard aerial picture sequence as a target aerial picture one by one;

[0108] Performing grayscale processing on the target aerial image to obtain a target grayscale image;

[0109] Binarizing the target grayscale image to obtain a target binary image;

[0110] Performing image erosion on the target binary image using the contour extraction model to obtain a target eroded image;

[0111] The target erosion images are subjected to image denoising and texture enhancement respectively to obtain target environment contours, and all target environment contours are aggregated into a standard environment contour.

[0112] In detail, the target aerial image may be grayscaled using the cv2.cvtcolor function of OpenCV to obtain a target grayscale image.

[0113] Specifically, the target grayscale image may be binarized using the cv2.threshold function of OpenCV to obtain a target binary image.

[0114] Specifically, the contour extraction model may be a Canny edge detection algorithm or an ant colony algorithm.

[0115] In detail, the target erosion image may be denoised using a Primal-dual algorithm or a non-local means denoising algorithm.

[0116] Specifically, the drawContours function may be used to perform texture enhancement on the target erosion image after image denoising to obtain the target environment contour.

[0117] In an embodiment of the present invention, by utilizing a preset contour extraction model to perform contour extraction on the standard aerial image sequence to obtain a standard environmental contour, redundant information in the standard aerial image sequence can be effectively filtered out, and image enhancement can be performed. This can effectively extract image features of the scene described by the current drone, facilitating subsequent overlap matching of the environmental data.

[0118] S4, using the standard environment profile to perform overlap matching on the environment data to obtain overlap environment data, and using the standard environment profile to perform data correction on the overlap environment data to obtain navigation environment data;

[0119] In the embodiment of the present invention, the step of performing overlap matching on the environmental data using the standard environmental profile to obtain overlap environmental data includes:

[0120] Determine the movement direction of the UAV according to the positioning information, and perform projection mapping on the environmental data in sequence according to the movement direction to obtain environmental projection data;

[0121] Performing contour extraction on the environment projection data to obtain an environment projection contour;

[0122] Extracting features from the standard environment contour to obtain a standard feature set, and performing normalization encoding on the standard feature set using a preset normalization function to obtain a standard feature encoding set;

[0123] Extracting features from the environmental projection contour to obtain an environmental feature set, and performing normalization encoding on the environmental feature set using the normalization function to obtain an environmental feature encoding set;

[0124] Selecting a first standard feature code from the standard feature code set as a target standard code, and calculating a Hamming distance between the target standard code and an environmental feature code in the environmental feature code set;

[0125] The environmental feature code whose Hamming distance is less than a preset Hamming threshold is selected as the target environmental code, and the environmental projection data corresponding to the target environmental code is used as the coincidence environmental data.

[0126] The step of extracting the contour of the environment projection data to obtain the environment projection contour is consistent with the step of extracting the contour of the standard aerial image sequence using a preset contour extraction model to obtain the standard environment contour in the above step S3, and will not be repeated here.

[0127] Specifically, the VGGNet model may be used to extract features from the standard environment contour and the environment projection contour.

[0128] In detail, the normalization function may be a softmax function.

[0129] In detail, the use of the standard environment contour to correct the overlapping environment data to obtain navigation environment data refers to constructing a standard environment model based on the standard environment contour, and using the standard environment model to replace the 3D environment model in the overlapping environment data.

[0130] In an embodiment of the present invention, by using the standard environment contour to match the environmental data for overlap, obtaining overlap environment data, and using the standard environment contour to perform data correction on the overlap environment data to obtain navigation environment data, the current environment model of the UAV can be determined, and the environment model after the flight route is simulated in combination with the environment data, thereby ensuring the accuracy of the navigation environment data.

[0131] S5. Obtain the motion data of the UAV, use the preset displacement bag-of-words model to perform position matching on the navigation environment data to obtain the node position, and use the preset fuzzy decision algorithm to generate the motion trajectory of the UAV according to the motion data and the node position to complete the map navigation.

[0132] In the embodiment of the present invention, the motion data refers to the motion acceleration, gyroscope and other data of the drone.

[0133] In the embodiment of the present invention, refer to Figure 3 As shown, the use of a preset displacement bag-of-words model to perform position matching on the navigation environment data to obtain the node position includes:

[0134] S51, extracting features from the navigation environment data using the displacement bag-of-words model to obtain navigation bag-of-words features of the navigation environment data;

[0135] S52, using the following bag-of-words similarity formula to calculate the similarity between the navigation bag-of-words feature and each bag-of-words feature in the preset bag-of-words feature library:

[0136]

[0137] Wherein, S refers to the similarity, is the preset adversarial coefficient, arccos is the inverse cosine function, refers to the navigation word bag feature, refers to the bag-of-words feature in the preset bag-of-words feature library, and T is the transposition symbol;

[0138] S53, selecting the word bag feature whose similarity is less than a preset similarity threshold from the preset word bag feature library as the target word bag feature, and taking the visual word set corresponding to the target word bag feature as the target word set;

[0139] S54, constructing a bag-of-words tree model according to the target word set, and performing loop detection on the navigation environment data according to the bag-of-words tree model to obtain node positions.

[0140] Specifically, by using the bag-of-words similarity formula to calculate the similarity between the navigation bag-of-words feature and each bag-of-words feature in a preset bag-of-words feature library, the multidimensional feature vector can be compared in terms of vector direction and length, thereby ensuring the accuracy of the similarity calculation.

[0141] In detail, the shifted bag-of-words model may be a brief bag-of-words model.

[0142] Specifically, the preset bag-of-words feature library is obtained by extracting and annotating bag-of-words features from a large amount of environmental data using the shifted bag-of-words model.

[0143] In detail, the SLAM detection model can be used to perform loop detection on the navigation environment data according to the bag-of-words tree model to obtain the node position.

[0144] In detail, the generating the motion trajectory of the UAV according to the motion data and the node position by using a preset fuzzy decision algorithm includes:

[0145] Generate a path network diagram of the drone according to the node positions;

[0146] Extracting the steering delay of the UAV and the movement speed of the UAV from the movement data, and generating the path delay weight of the path network graph using the steering delay and the movement speed;

[0147] Extracting a navigation path set of the UAV from the path network graph, selecting one navigation path from the navigation path set as a target path one by one, and calculating a path delay corresponding to the target path according to the path delay weight to obtain a path delay set;

[0148] A path delay with the smallest value is selected from the path delay set as a target path delay, and a navigation path corresponding to the target path delay is used as a motion trajectory of the UAV.

[0149] Specifically, the path network graph is a directed graph starting from the current position of the drone and passing through the node positions.

[0150] Specifically, the steering delay refers to the delay time when the drone is turning, and the delay time is linearly related to the steering angle.

[0151] In detail, a topological sorting algorithm may be used to extract a navigation path set of the drone from the path network graph.

[0152] In an embodiment of the present invention, by using a preset displacement bag-of-words model to perform position matching on the navigation environment data to obtain node positions, a bag-of-words tree model that is closest to the current environment data can be generated, and a path can be generated in a loop detection manner, thereby improving the efficiency of drone navigation.

[0153] The embodiment of the present invention can obtain the geographical environment information of the location of the UAV by searching for corresponding environmental data in a preset map database according to the positioning information, thereby performing positioning navigation according to the geographical environment information, and by using the trained mean square correction model to perform lens distortion correction on the initial aerial image sequence to obtain a standard aerial image sequence, it is possible to correct the distortion of the fisheye lens of the camera module, ensure the accuracy of the subsequent standard environment contour, and provide more accurate navigation for the UAV, and by using a preset contour extraction model to perform contour extraction on the standard aerial image sequence to obtain a standard environment contour, it is possible to effectively filter out redundant information of the image in the standard aerial image sequence, and perform image enhancement, and it is possible to effectively extract the image features of the scene described by the current UAV, so as to facilitate the subsequent coincidence matching of the environmental data. By using the standard environment contour to match the environment data for overlap, the overlap environment data is obtained, and the standard environment contour is used to correct the overlap environment data to obtain navigation environment data, the current environment model of the UAV can be determined, and the environment model after the flight route is simulated in combination with the environment data, so as to ensure the accuracy of the navigation environment data, and by using the preset displacement word bag model to match the navigation environment data for position, the node position can be obtained, and the word bag tree model closest to the current environment data can be generated, and the path can be generated in the form of loop detection, thereby improving the efficiency of UAV navigation, and by using the positioning information, aerial video and motion data of the UAV, the positioning navigation and visual navigation can be integrated to provide more accurate navigation data for the UAV. Therefore, the map matching navigation method based on decision analysis proposed in the present invention can solve the problem of low accuracy when performing map navigation on UAVs.

[0154] like Figure 4 , which is a functional module diagram of a map matching navigation device based on decision analysis provided by an embodiment of the present invention.

[0155] The map matching navigation device 100 based on decision analysis of the present invention can be installed in an electronic device. According to the functions to be implemented, the map matching navigation device 100 based on decision analysis can include a positioning information module 101, a video correction module 102, a contour extraction module 103, a data correction module 104 and an intelligent navigation module 105. The module of the present invention can also be called a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0156] In this embodiment, the functions of each module / unit are as follows:

[0157] The positioning information module 101 is used to obtain the positioning information uploaded by the positioning component of the drone, and search for corresponding environmental data in a preset map database according to the positioning information;

[0158] The video correction module 102 is used to obtain the initial aerial video taken by the camera component of the drone, extract the initial aerial image sequence from the initial aerial video, and use the trained mean square correction model to perform lens distortion correction on the initial aerial image sequence to obtain a standard aerial image sequence;

[0159] The contour extraction module 103 is used to extract contours from the standard aerial image sequence using a preset contour extraction model to obtain a standard environment contour;

[0160] The data correction module 104 is used to perform coincidence matching on the environment data using the standard environment profile to obtain coincident environment data, and perform data correction on the coincident environment data using the standard environment profile to obtain navigation environment data;

[0161] The intelligent navigation module 105 is used to obtain the motion data of the drone, use a preset displacement bag-of-words model to position match the navigation environment data to obtain a node position, and use a preset fuzzy decision algorithm to generate a motion trajectory of the drone according to the motion data and the node position to complete map navigation, wherein the use of the preset displacement bag-of-words model to position match the navigation environment data to obtain a node position includes: using the displacement bag-of-words model to perform feature extraction on the navigation environment data to obtain a navigation bag-of-words feature of the navigation environment data; and using the following bag-of-words similarity formula to calculate the similarity between the navigation bag-of-words feature and each bag-of-words feature in a preset bag-of-words feature library:

[0162]

[0163] Wherein, S refers to the similarity, is the preset adversarial coefficient, arccos is the inverse cosine function, refers to the navigation word bag feature, It refers to the bag-of-words features in the preset bag-of-words feature library, and T is the transposition symbol; the bag-of-words features whose similarity is less than a preset similarity threshold are selected in the preset bag-of-words feature library as target bag-of-words features, and the visual word set corresponding to the target bag-of-words features is used as the target word set; a bag-of-words tree model is constructed according to the target word set, and a loop detection is performed on the navigation environment data according to the bag-of-words tree model to obtain the node position.

[0164] In detail, each module in the map matching navigation device 100 based on decision analysis in the embodiment of the present invention is used in the same manner as described above. Figures 1 to 3 The map matching navigation method based on decision analysis described in the present invention has the same technical means and can produce the same technical effects, so it will not be repeated here.

[0165] like Figure 5 , is a schematic diagram of the structure of an electronic device for implementing a map matching navigation method based on decision analysis provided by an embodiment of the present invention.

[0166] The electronic device 1 may include a processor 10, a memory 11, a communication bus 12, and a communication interface 13, and may also include a computer program stored in the memory 11 and executable on the processor 10, such as a map navigation program based on decision analysis.

[0167] In some embodiments, the processor 10 may be composed of an integrated circuit, for example, a single packaged integrated circuit, or a plurality of packaged integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and combinations of various control chips. The processor 10 is the control core (ControlUnit) of the electronic device, and uses various interfaces and lines to connect various components of the entire electronic device, and executes or executes programs or modules stored in the memory 11 (for example, executing a map navigation program based on decision analysis, etc.), and calls data stored in the memory 11 to execute various functions of the electronic device and process data.

[0168] The memory 11 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, mobile hard disk, multimedia card, card-type memory (for example: SD or DX memory, etc.), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 11 can be an internal storage unit of an electronic device, such as a mobile hard disk of the electronic device. In other embodiments, the memory 11 can also be an external storage device of an electronic device, such as a plug-in mobile hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device. Further, the memory 11 can also include both an internal storage unit of the electronic device and an external storage device. The memory 11 can not only be used to store application software and various types of data installed in the electronic device, such as the code of a map navigation program based on decision analysis, but also can be used to temporarily store data that has been output or is to be output.

[0169] The communication bus 12 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The bus is configured to realize connection and communication between the memory 11 and at least one processor 10, etc.

[0170] The communication interface 13 is used for communication between the above-mentioned electronic device and other devices, including a network interface and a user interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is generally used to establish a communication connection between the electronic device and other electronic devices. The user interface may be a display (Display), an input unit (such as a keyboard (Keyboard)), and optionally, the user interface may also be a standard wired interface, a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, and an OLED (Organic Light-Emitting Diode, organic light-emitting diode) touch device, etc. Among them, the display may also be appropriately referred to as a display screen or a display unit, which is used to display information processed in the electronic device and to display a visual user interface.

[0171] The figure only shows an electronic device with components. Those skilled in the art will understand that the structure shown in the figure does not constitute a limitation on the electronic device, and may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.

[0172] For example, although not shown, the electronic device may also include a power source (such as a battery) for supplying power to each component. Preferably, the power source may be logically connected to the at least one processor 10 through a power management device, so that the power management device can realize functions such as charging management, discharging management, and power consumption management. The power source may also include one or more DC or AC power sources, recharging devices, power failure detection circuits, power converters or inverters, power status indicators, and other arbitrary components. The electronic device may also include a variety of sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be repeated here.

[0173] It should be understood that the embodiment is for illustration only and the scope of the patent application is not limited to this structure.

[0174] The map navigation program based on decision analysis stored in the memory 11 of the electronic device 1 is a combination of multiple instructions. When running in the processor 10, it can achieve:

[0175] Obtaining positioning information uploaded by the positioning component of the drone, and searching for corresponding environmental data in a preset map database according to the positioning information;

[0176] Acquire an initial aerial video taken by a camera assembly of the drone, extract an initial aerial image sequence from the initial aerial video, and perform lens distortion correction on the initial aerial image sequence using a trained mean square correction model to obtain a standard aerial image sequence;

[0177] Using a preset contour extraction model to extract contours from the standard aerial image sequence to obtain a standard environment contour;

[0178] Using the standard environment contour to match the environment data for coincidence, to obtain coincident environment data, and using the standard environment contour to perform data correction on the coincident environment data, to obtain navigation environment data;

[0179] The motion data of the UAV is obtained, the navigation environment data is position-matched using a preset displacement bag-of-words model to obtain node positions, and the motion trajectory of the UAV is generated according to the motion data and the node positions using a preset fuzzy decision algorithm to complete map navigation.

[0180] Specifically, the specific implementation method of the processor 10 for the above instructions can refer to the description of the relevant steps in the corresponding embodiment of the accompanying drawings, which will not be repeated here.

[0181] Furthermore, if the module / unit integrated in the electronic device 1 is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a mobile hard disk, a magnetic disk, an optical disk, a computer memory, and a read-only memory (ROM).

[0182] The present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor of an electronic device, the computer program can implement:

[0183] Obtaining positioning information uploaded by the positioning component of the drone, and searching for corresponding environmental data in a preset map database according to the positioning information;

[0184] Acquire an initial aerial video taken by a camera assembly of the drone, extract an initial aerial image sequence from the initial aerial video, and perform lens distortion correction on the initial aerial image sequence using a trained mean square correction model to obtain a standard aerial image sequence;

[0185] Using a preset contour extraction model to extract contours from the standard aerial image sequence to obtain a standard environment contour;

[0186] Using the standard environment contour to match the environment data for coincidence, to obtain coincident environment data, and using the standard environment contour to perform data correction on the coincident environment data, to obtain navigation environment data;

[0187] The motion data of the UAV is obtained, the navigation environment data is position-matched using a preset displacement bag-of-words model to obtain node positions, and the motion trajectory of the UAV is generated according to the motion data and the node positions using a preset fuzzy decision algorithm to complete map navigation.

[0188] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

[0189] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0190] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0191] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0192] Therefore, no matter from which point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the appended claims rather than the above description, so it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any attached figure mark in the claims should not be regarded as limiting the claims involved.

[0193] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0194] In addition, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices stated in the system claim can also be implemented by one unit or device through software or hardware. The words first, second, etc. are used to indicate names, and do not indicate any particular order.

[0195] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A map matching navigation method based on decision analysis, characterized in that: The method comprises: S1: Obtaining positioning information uploaded by the positioning component of the drone, and searching for corresponding environmental data in a preset map database according to the positioning information; S2: obtaining an initial aerial video shot by the camera assembly of the drone, extracting an initial aerial image sequence from the initial aerial video, and performing lens distortion correction on the initial aerial image sequence using a trained mean square correction model to obtain a standard aerial image sequence; S3: using a preset contour extraction model to perform contour extraction on the standard aerial image sequence to obtain a standard environment contour; S4: using the standard environment profile to match the environment data with respect to the overlap, to obtain overlap environment data, and using the standard environment profile to perform data correction on the overlap environment data, to obtain navigation environment data; S5: acquiring the motion data of the drone, performing position matching on the navigation environment data using a preset displacement bag model to obtain a node position, and generating a motion trajectory of the drone according to the motion data and the node position using a preset fuzzy decision algorithm to complete map navigation, wherein performing position matching on the navigation environment data using a preset displacement bag model to obtain a node position includes: S51: extracting features from the navigation environment data using the displacement bag-of-words model to obtain navigation bag-of-words features of the navigation environment data; S52: Calculate the similarity between the navigation word bag feature and each word bag feature in the preset word bag feature library using the following word bag similarity formula: Wherein, S refers to the similarity, is the preset adversarial coefficient, arccos is the inverse cosine function, refers to the navigation word bag feature, refers to the bag-of-words feature in the preset bag-of-words feature library, and T is the transposition symbol; S53: selecting the bag-of-words features whose similarity is less than a preset similarity threshold from the preset bag-of-words feature library as target bag-of-words features, and taking the visual word set corresponding to the target bag-of-words features as the target word set; S54: constructing a bag-of-words tree model according to the target word set, and performing loop detection on the navigation environment data according to the bag-of-words tree model to obtain node positions.

2. The map matching navigation method based on decision analysis according to claim 1, characterized in that: The searching corresponding environmental data in a preset map database according to the positioning information includes: Extracting longitude and latitude information from the positioning information, and using the longitude and latitude information to retrieve corresponding terrain data in the map database; Altitude information is extracted from the positioning information, and the terrain data is subjected to height mapping conversion using the altitude information to obtain environmental data.

3. The map matching navigation method based on decision analysis according to claim 1, characterized in that: The step of extracting an initial aerial image sequence from the initial aerial video comprises: Using a preset video conversion tool to convert the initial aerial video into a video frame sequence; Marking video nodes of the video frame sequence at preset time intervals to obtain a video node set; A video node is selected from the video node set one by one in time sequence as a target node, a video frame image corresponding to the target node is used as a target frame image, and the target frame images are arranged into the video frame sequence in time sequence.

4. The map matching navigation method based on decision analysis according to claim 1, characterized in that: Before the trained mean square correction model is used to correct the lens distortion of the initial aerial image sequence to obtain a standard aerial image sequence, the method further includes: Obtaining a preset normal sample atlas, and positioning and annotating the normal sample atlas to obtain a normal positioning atlas; Using the camera assembly to shoot the normal positioning atlas to obtain a distorted positioning atlas; Generate a normal coordinate set according to the positioning annotation points in the normal positioning atlas, and generate a distorted coordinate set according to the positioning annotation points in the distorted positioning atlas; The distortion coordinate set is corrected using the following mean square correction algorithm to obtain a predicted coordinate set: in, is the abscissa of the ith distorted coordinate in the distorted coordinate set, refers to the ordinate of the i-th distorted coordinate in the distorted coordinate set, X i is the horizontal coordinate of the i-th predicted coordinate in the predicted coordinate set, y i is the ordinate of the i-th predicted coordinate in the predicted coordinate set, i refers to the i-th in the predicted coordinate set and the distorted coordinate set, j1 is the first term of the Taylor series expansion, j2 is the second term of the Taylor series expansion, d refers to the internal parameter of the camera assembly, k1 is the first tangential parameter of the mean square correction algorithm, and k2 refers to the second tangential parameter of the mean square correction algorithm: The distance difference between the normal coordinate set and the predicted coordinate set is calculated, and the parameters of the mean square correction algorithm are updated using the distance difference until the distance difference is less than a preset distance threshold, and the mean square correction algorithm after parameter update is used as the mean square correction model.

5. The map matching navigation method based on decision analysis according to claim 1, characterized in that: The method of using the trained mean square correction model to perform lens distortion correction on the initial aerial image sequence to obtain a standard aerial image sequence includes: Selecting one initial aerial picture in the initial aerial picture sequence as a target initial picture one by one; Generate a corresponding unit matrix according to the image size of the target initial image, and calculate a mapping vector of the unit matrix using the mean square correction model; Converting the target initial image into a target image matrix, multiplying the target image matrix by the mapping vector to obtain a target correction matrix, and converting the target correction matrix into a target correction image; All the target corrected images are sorted according to the order of the initial aerial image sequence to obtain the standard aerial image sequence.

6. The map matching navigation method based on decision analysis as claimed in claim 1, characterized in that: The method of extracting contours of the standard aerial image sequence using a preset contour extraction model to obtain a standard environment contour includes: Selecting one standard aerial picture in the standard aerial picture sequence as a target aerial picture one by one; Performing grayscale processing on the target aerial image to obtain a target grayscale image; Binarizing the target grayscale image to obtain a target binary image; Performing image erosion on the target binary image using the contour extraction model to obtain a target eroded image; The target erosion images are subjected to image denoising and texture enhancement respectively to obtain target environment contours, and all target environment contours are aggregated into a standard environment contour.

7. The map matching navigation method based on decision analysis according to claim 1, characterized in that: The using the standard environment profile to perform overlap matching on the environment data to obtain overlap environment data includes: Determine the movement direction of the UAV according to the positioning information, and perform projection mapping on the environmental data in sequence according to the movement direction to obtain environmental projection data; Performing contour extraction on the environment projection data to obtain an environment projection contour; Extracting features from the standard environment contour to obtain a standard feature set, and performing normalization encoding on the standard feature set using a preset normalization function to obtain a standard feature encoding set; Extracting features from the environmental projection contour to obtain an environmental feature set, and performing normalization encoding on the environmental feature set using the normalization function to obtain an environmental feature encoding set; Selecting a first standard feature code from the standard feature code set as a target standard code, and calculating a Hamming distance between the target standard code and an environmental feature code in the environmental feature code set; The environmental feature code whose Hamming distance is less than a preset Hamming threshold is selected as the target environmental code, and the environmental projection data corresponding to the target environmental code is used as the coincidence environmental data.

8. A map matching navigation device based on decision analysis, characterized in that: The device comprises: A positioning information module is used to obtain the positioning information uploaded by the positioning component of the drone, and search for corresponding environmental data in a preset map database according to the positioning information; A video correction module is used to obtain an initial aerial video taken by the camera assembly of the drone, extract an initial aerial image sequence from the initial aerial video, and perform lens distortion correction on the initial aerial image sequence using a trained mean square correction model to obtain a standard aerial image sequence; A contour extraction module, used to extract contours from the standard aerial image sequence using a preset contour extraction model to obtain a standard environment contour; A data correction module, used to perform coincidence matching on the environmental data using the standard environmental profile to obtain coincident environmental data, and to perform data correction on the coincident environmental data using the standard environmental profile to obtain navigation environmental data; The intelligent navigation module is used to obtain the motion data of the UAV, use the preset displacement bag-of-words model to match the position of the navigation environment data to obtain the node position, and use the preset fuzzy decision algorithm to generate the motion trajectory of the UAV according to the motion data and the node position to complete the map navigation, wherein the use of the preset displacement bag-of-words model to match the position of the navigation environment data to obtain the node position includes: using the displacement bag-of-words model to extract features of the navigation environment data to obtain the navigation bag features of the navigation environment data; using the following bag-of-words similarity formula to calculate the similarity between the navigation bag features and each bag-of-words feature in the preset bag-of-words feature library: Wherein, S refers to the similarity, is the preset adversarial coefficient, arccos is the inverse cosine function, refers to the navigation word bag feature, It refers to the bag-of-words features in the preset bag-of-words feature library, and T is the transposition symbol; the bag-of-words features whose similarity is less than a preset similarity threshold are selected in the preset bag-of-words feature library as target bag-of-words features, and the visual word set corresponding to the target bag-of-words features is used as the target word set; a bag-of-words tree model is constructed according to the target word set, and a loop detection is performed on the navigation environment data according to the bag-of-words tree model to obtain the node position.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the map matching navigation method based on decision analysis as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the map matching navigation method based on decision analysis as described in any one of claims 1 to 7 is implemented.

Images