Navigation map-based road closure event identification method, apparatus and device, and medium

Abstract

The embodiment of the invention provides a road closure event identification method and device based on a navigation map, equipment and a medium. The method comprises the following steps: acquiring an image frame of a navigation map, and identifying a road closure icon and a road closure section from the image frame; determining a target matching relationship according to the position information of each road closure icon and the position information of each road closure section in the image frame; wherein the target matching relation represents the matching relation between the road closure icon and the road closure section; determining a target road section from the closure road sections according to the target matching relationship; wherein the target road section represents a closure road section actually existing in the road corresponding to the image frame; according to the position information of the target road section, road closure event information is generated and stored; wherein the road closure event information represents the road closure condition of the road corresponding to the image frame. The road closure icons and the road closure sections are automatically matched and screened, so that the recognition efficiency and precision of the road closure events are improved.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a method, apparatus, device and medium for identifying road closure events based on navigation maps. Background Technology

[0002] In map navigation services, if users find incorrect road closure information on the map, they can upload a screenshot of the map containing the incorrect information through the client to prompt the backend staff to correct the incorrect information on the map.

[0003] Staff need to examine each screenshot, identify the location of the road closure in the image, and finally enter the processing results into the map database. This process suffers from low processing efficiency, high labor costs, and severe response delays, making it difficult to meet the real-time requirements of large-scale user reporting scenarios. Summary of the Invention

[0004] This application provides a method, apparatus, device, and medium for identifying road closure events based on navigation maps, in order to improve the accuracy and efficiency of road closure event identification.

[0005] In a first aspect, embodiments of this application provide a method for identifying road closure events based on navigation maps, including:

[0006] Acquire image frames from the navigation map and identify road closure icons and road closure sections from the image frames;

[0007] Based on the location information of each road closure icon and the location information of each road closure segment in the image frame, a target matching relationship is determined; wherein, the target matching relationship represents the matching relationship between the road closure icon and the road closure segment;

[0008] Based on the target matching relationship, a target road segment is determined from each of the road closure segments; wherein, the target road segment represents the actual road closure segment existing in the road corresponding to the image frame;

[0009] Based on the location information of the target road segment, road closure event information is generated and stored; wherein, the road closure event information represents the road closure status of the road corresponding to the image frame.

[0010] Secondly, embodiments of this application provide a road closure event identification device based on a navigation map, comprising:

[0011] The image acquisition unit is used to acquire image frames of the navigation map and identify road closure icons and road closure sections from the image frames;

[0012] The relationship determination unit is used to determine the target matching relationship based on the position information of each road closure icon and the position information of each road closure segment in the image frame; wherein, the target matching relationship represents the matching relationship between the road closure icon and the road closure segment;

[0013] A road segment determination unit is used to determine a target road segment from each of the road closure segments according to the target matching relationship; wherein, the target road segment represents the actual road closure segment existing in the road corresponding to the image frame;

[0014] An event generation unit is used to generate road closure event information based on the location information of the target road segment and store it; wherein the road closure event information represents the road closure status of the road corresponding to the image frame.

[0015] Thirdly, embodiments of this application provide an electronic device, including: a memory and a processor;

[0016] The memory stores computer-executed instructions;

[0017] The processor executes computer execution instructions stored in the memory, causing the processor to perform the implementation method described in the first aspect above.

[0018] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the embodiments described in the first aspect above.

[0019] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the implementation methods described in the first aspect above.

[0020] This application provides a method, apparatus, device, and medium for identifying road closure events based on navigation maps. It acquires image frames from navigation maps and identifies road closure icons and road sections from these frames, achieving automated capture of key visual elements in user-reported screenshots. Based on the location information of each road closure icon and road section in the image frame, corresponding road closure icons and road sections are determined, establishing a target matching relationship between them. According to the target matching relationship, target road sections are determined from each road section, effectively filtering out actually existing road closures and avoiding interference from falsely detected road sections. Finally, road closure event information is automatically generated based on the location information of the target road section. Automatic association and filtering through location information avoids the tedious process of manually analyzing screenshots, significantly shortening the information processing cycle. The establishment of target matching relationships ensures spatial consistency between road closure icons and road sections, improving the accuracy and efficiency of road closure event identification. Attached Figure Description

[0021] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0022] Figure 1 A flowchart illustrating a method for identifying road closure events based on a navigation map, provided in an embodiment of this application;

[0023] Figure 2 A schematic diagram of a road closure icon provided in an embodiment of this application;

[0024] Figure 3 A schematic diagram of an image frame provided in an embodiment of this application;

[0025] Figure 4 A flowchart illustrating a method for identifying road closure events based on a navigation map, provided in an embodiment of this application;

[0026] Figure 5 A flowchart illustrating a method for identifying road closure events based on a navigation map, provided in an embodiment of this application;

[0027] Figure 6 A schematic diagram of a road closure event identification device based on a navigation map, provided in an embodiment of this application;

[0028] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0029] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0030] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0031] First, let me explain the terms used in this application:

[0032] Road closure icon: Used in navigation maps to represent road closure signs. Road closure signs are traffic signs erected beside roads to inform drivers that the road ahead is closed to traffic.

[0033] Road closure section: refers to a specific section of road where vehicles and pedestrians are prohibited from passing for some reason. In navigation maps, the road closure icon is usually located at the starting point of the road closure section.

[0034] ROI: Region of Interest.

[0035] In map navigation services, if users find incorrect road closure information on the map, such as construction or temporary traffic control, they can upload a map screenshot containing the road closure information through the client. These screenshots typically include key information such as road boundaries, scale indicators (e.g., "50m", "100m"), road closure icons (e.g., red circle with a horizontal bar), and the closed road segment (e.g., red dashed line). The backend needs to process these user-reported images to quickly extract the road closure location, the shape of the closed road segment, and their relationships, in order to dynamically update the map data and ensure the accuracy of navigation route planning.

[0036] The current mainstream processing method relies on manual review. Staff need to examine each screenshot, identify the location of the road closure icon, analyze the route shape, and retrieve the geographic coordinates through the POI (Point of Interest) information in the screenshot before finally entering the processing results into the map database. Traditional manual processing of a single screenshot takes about 10 minutes, which has problems such as low processing efficiency, high labor costs, and serious response delays, making it difficult to meet the real-time requirements of large-scale user reporting scenarios.

[0037] This application provides a method, apparatus, device, and medium for identifying road closure events based on navigation maps, aiming to solve the above-mentioned technical problems in the prior art.

[0038] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0039] Figure 1 This is a flowchart illustrating a method for identifying road closure events based on a navigation map, provided in an embodiment of this application. This method can be executed by a device for identifying road closure events based on a navigation map. Figure 1 As shown, the method includes:

[0040] S101. Obtain image frames of the navigation map and identify road closure icons and road closure sections from the image frames.

[0041] Exemplarily, an image frame of a navigation map refers to a static screen image captured by a user terminal in a navigation or map roaming state. The content in the image frame includes a road network, POIs, scale indicators, interface controls, road closure visualization elements, etc. The resolution range of the image frame can be from 720×1280 to 1440×3200 pixels, supporting the screenshot formats of mainstream systems. The user can generate an image frame of the navigation map by taking a screenshot and upload the image frame. For example, when the user terminal displays the navigation map and notices that the road closure icons and road closure sections in the navigation map do not match the actual road closure situation, the user takes a screenshot by clicking a button on the user terminal and uploads the obtained image frame to a preset cloud platform.

[0042] After obtaining the image frame uploaded by the user, road closure icons can be recognized from the image frame. There are graphical symbols with road closure semantics displayed in the image frame, that is, road closure icons. The typical forms of road closure icons include but are not limited to a red circle with a white horizontal bar inside, a red triangle with the black character "closed" inside, a yellow diamond with the black words "construction", etc. Figure 2 is a schematic diagram of a road closure icon. There is also a road closure section displayed in the image frame. The road closure section represents a section that cannot be passed. For example, the road closure section can be represented by a red dotted line surrounding a gray line. Figure 3 is a schematic diagram of the image frame. Figure 3 In, the logo of a red circle with a white horizontal bar inside is a road closure icon, and the line segment with a gray area inside the red dotted line is a road closure section.

[0043] When performing the recognition process of road closure icons on the image frame, a multi-scale sliding window + lightweight convolutional neural network model or an OpenCV (Open Source Computer Vision Library) template matching algorithm can be used to output the bounding box coordinates and class confidence of each icon in the image frame. The confidence threshold can be set to 0.85 - 0.92, and the result higher than this threshold is regarded as a road closure icon; when recognizing the road closure section, continuous, connected red or orange dotted / solid line segments with road direction characteristics can be extracted from the image frame. For example, the image frame can be subjected to channel separation, focusing on the red channel, and after Otsu adaptive threshold binarization, morphological closing operations are used to fill the gaps in the line segments, and then line segments with a length ≥ 30 pixels, a width ≤ 8 pixels, and a direction angle within ±30° of the main road direction are jointly extracted through Hough line transformation and connected component analysis as the road closure section. The recognition of the two can be executed synchronously or serially in stages. In this embodiment, the recognition methods of road closure icons and road closure sections are not specifically limited.

[0044] S102. Determine the target matching relationship based on the location information of each road closure icon and each road closure segment in the image frame; wherein, the target matching relationship represents the matching relationship between the road closure icon and the road closure segment.

[0045] For example, after identifying each road closure icon and each road closure segment in the image frame, the location information of each road closure icon and each road closure segment is determined. The location information can be image coordinates within the image frame or an actual geographical location. For instance, the image coordinates can be determined first, and then the actual geographical location can be calculated based on the image coordinates and the scale of the image frame. The location information of a road closure icon can be a point, and the location information of a road closure segment can be a region, such as the start and end points of the road closure segment.

[0046] In this embodiment, for an icon, location information can refer to its center point coordinates and bounding box size; for a road segment, it can refer to its geometric center point coordinates, main direction vector (calculated from the starting point to the ending point), and minimum bounding rectangle, etc.

[0047] Target matching relationships refer to the matching relationships between road closure icons and road closure segments. For example, one road closure icon corresponds to one road closure segment. For instance, the Hungarian algorithm can be used to obtain the maximum weighted matching result, outputting a unique set of mapping relationships. That is, each road closure icon is associated with at most one road closure segment, and each road closure segment is associated with at most one road closure icon; this set is the target matching relationship. Alternatively, graph neural network modeling can be used to encode the identified road closure icons and road closure segments into node feature vectors. Edge features are constructed using spatial relative position, visual semantic similarity, and contextual road connectivity. After aggregation through a 2-layer GraphSAGE (graph sampling and aggregation network), a matching probability matrix is ​​output, improving the ambiguity resolution capability in complex intersection scenarios with multiple icons and road segments.

[0048] In this embodiment, the spatial relationship between a road closure icon and a road closure segment can be determined based on their location information, thereby establishing a target matching relationship. The spatial relationship characterizes the spatial distance and directional consistency between the road closure icon and the road closure segment. If the spatial relationship indicates that the road closure icon is located at the starting point of the road closure segment, then the road closure icon is considered to match the road closure segment. If there are redundant road closure icons or road closure segments, these are filtered out.

[0049] S103. Based on the target matching relationship, determine the target road segment from each road closure segment; wherein, the target road segment represents the actual road closure segment in the road corresponding to the image frame.

[0050] For example, the road closure icons and road closure segments that have been mapped in the target matching relationship are subsequently verified. Unmatched road closure icons and road closure segments are not included in the target road segment determination at this time. The target road segment can be determined by performing a double reasonableness check process.

[0051] The first layer of verification can be geometric rationality, which involves calculating the length, connected region area, and rate of curvature change of the road closure segment, and determining whether these parameters meet preset geometric conditions. For example, is the length within a preset range? If the preset geometric conditions are met, a second layer of verification is performed; otherwise, the road closure segment is eliminated, meaning it is not the target road segment. The second layer can be spatial semantic verification, which determines whether the road closure segment is located within the road area corresponding to the road closure icon, and whether the road closure icon is located at the starting point of the road closure segment. If so, the verification passes; otherwise, it fails. If both verifications pass, the road closure segment can be marked as the target road segment.

[0052] Furthermore, a cross-frame consistency enhancement mechanism can be introduced. If the current image frame belongs to a multi-frame sequence continuously reported by the user, i.e., the user reports multiple consecutive screenshots, then adjacent screenshots are aligned and stitched together to obtain a complete road closure segment, generating a longer and smoother road closure segment. Based on the position information of each road closure icon and the position information of each stitched road closure segment, the target matching relationship is determined, and then the target road segment is determined from the stitched road closure segments based on the target matching relationship.

[0053] S104. Based on the location information of the target road segment, generate road closure event information and store it; wherein, the road closure event information represents the road closure status of the road corresponding to the image frame.

[0054] For example, the location information of the target road segment can be determined, such as by determining the image coordinates of the target road segment in the image frame, or by determining the actual latitude and longitude of the target road segment. That is, a coordinate dimensionality upgrade transformation can be performed on the target road segment. For example, based on the identified scale marker and the latitude and longitude of the map center point in the image frame, an affine transformation model from the pixel coordinate system to the spherical Mercator projection coordinate system is established. The pixel endpoint coordinates of the target road segment are substituted into this model to obtain the corresponding Mercator coordinates. Finally, the inverse Mercator transformation is performed to convert it into a latitude and longitude coordinate string.

[0055] The format or template for road closure event information can be pre-defined, and road closure event information can be automatically generated based on the location information of the target road segment. For example, the road closure event information can be a structured JSON object. The road closure event information can include the location information of the target road segment, the shape information of the target road segment, and the encoding of the corresponding road closure icon. The road closure event information can be serialized and written to a distributed key-value store or time-series database for real-time retrieval by the map dynamic update engine. In this embodiment, the road closure event information can also include semantic tag fields, impact level fields, etc. The semantic tag field can characterize the reason for the road closure determined by recognizing the road closure icon, such as construction, traffic control, accident, etc., and the impact level field can characterize the impact of the road closure on road traffic.

[0056] In this embodiment, multi-source feature collaborative recognition simultaneously captures the semantic identifier of the road closure icon and the spatial morphology of the road closure segment, thereby constructing a target matching relationship. This breaks the limitation of isolated icon and road segment recognition and establishes a visual semantic relationship between the two. Based on geometric constraints and spatial semantic verification mechanisms, the target road segment that truly corresponds to the icon and physically exists is accurately selected from all identified road closure segments. Finally, end-to-end conversion and structured encapsulation of pixel coordinates to geographic coordinates are completed, generating road closure event information that can directly drive incremental updates of the map database. This significantly improves the accuracy and automation of road closure event recognition and reduces the workload of manual review.

[0057] This application provides a method for identifying road closure events based on navigation maps. It acquires image frames from navigation maps and identifies road closure icons and road segments from these frames, automating the capture of key visual elements in user-reported screenshots. Based on the location information of each road closure icon and road segment in the image frame, corresponding road closure icons and segments are determined, establishing a target matching relationship between them. According to this target matching relationship, target road segments are identified from each road closure segment, effectively filtering out actually existing road closures and avoiding interference from falsely detected segments. Finally, road closure event information is automatically generated based on the location information of the target road segments. Automatic association and filtering using location information avoids the tedious process of manually analyzing screenshots, significantly shortening the information processing cycle. The establishment of target matching relationships ensures spatial consistency between road closure icons and road segments, improving the accuracy and efficiency of road closure event identification.

[0058] Figure 4 A flowchart illustrating a method for identifying road closure events based on navigation maps, as provided in this application embodiment, is shown below. Figure 4 As shown, this embodiment, based on the above embodiments, provides a detailed description of a method for identifying road closure events based on navigation maps. The method includes:

[0059] S401. Obtain image frames of the navigation map, identify candidate icons from the image frames according to preset first feature information, and identify candidate road segments from the image frames according to preset second feature information; wherein, the preset first feature information represents the display characteristics of the road closure icon, the preset second feature information represents the display characteristics of the road closure segment, the candidate icon is the icon initially selected, and the candidate road segment is the road segment initially selected.

[0060] For example, a first feature information is pre-set. The pre-set first feature information refers to the visual discrimination criteria used to quickly locate road closure icons in navigation map image frames, including display features of road closure icons such as color features, shape features, and spatial layout features. For example, the color feature is limited to a red main color, corresponding to the standardized warning color scheme of road closure icons such as "red circle + white horizontal bar" or "red triangle + black cross"; the shape feature is limited to a circle; and the spatial layout feature is limited to being located near the center line of the road.

[0061] A second set of pre-defined features is used as the visual criteria for locating road closures in an image frame. These features include color channel response features, topological connectivity features, and geometric scale features, which are the display characteristics of road closures. For example, the color channel response feature focuses on the enhanced response of the red channel, extracting highly saturated red dashed lines; the topological connectivity feature requires that candidate road segments appear as 8-neighborhood connected regions in the binarized image, with the number of pixels within the connected region being ≥200; the geometric scale feature limits the candidate road segment's main direction length to be ≥5 times its width, and its centerline curvature radius to be >50 pixels, in order to exclude non-closed road segments such as winding paths and arrow symbols.

[0062] All icons are identified from the image frames, and their display characteristics are determined, such as color and shape. Based on preset first feature information, candidate icons are identified from all icons. That is, candidate icons are a set of suspected road closure icons obtained through coarse screening using the first feature information, and each candidate icon can be represented as a minimum bounding rectangle. All road segments are identified from the image frames, and their display characteristics are determined, such as connectivity and length. Based on preset second feature information, candidate road segments are identified from all road segments. That is, candidate road segments are a set of suspected road closure road segments obtained through coarse screening using the second feature information, and each candidate road segment is represented as a sequence of polygonal vertices or a simplified polyline segment. Candidate icons and candidate road segments are intermediate calculation results and do not directly participate in the generation of road closure events; they can serve as input sources for subsequent fine-tuning stages.

[0063] In this embodiment, identifying candidate icons from an image frame based on preset first feature information and identifying candidate road segments from an image frame based on preset second feature information includes: performing binarization processing on the image frame to obtain a binarized image, and extracting a region of interest from the binarized image; wherein the region of interest represents the area where the road is located in the image frame; identifying candidate icons from the region of interest based on preset first feature information, and identifying candidate road segments from the region of interest based on preset second feature information.

[0064] Specifically, the image frame is a static screen image captured by the navigation map application when a user reports a road closure event. It contains multiple layers of visual elements, including the map base map, road network, POI labels, interface controls, advertising banners, and road closure icons and road closure segments. To improve the recognition accuracy of road closure icons and road closure segments, the original image frame can be binarized, thereby identifying candidate icons and candidate road segments in the binarized image.

[0065] Alternatively, the original image frames can be preprocessed before binarization. For example, preprocessing could involve using OpenCV for grayscale conversion, Gaussian blur denoising, and removing irrelevant elements such as pop-up ads and interface buttons from the screenshot. Furthermore, the resolution and size of each image frame can be standardized.

[0066] The image obtained after binarization is called a binarized image. Regions of interest (ROIs) are extracted from the binarized image. ROIs represent the approximate area where roads are located in the image frame, thus focusing on road and surrounding annotation information. ROIs can be created by connecting broken road connected components using morphological closing operations, combined with contour detection to extract the largest area closed contour. This contour, after convex hull filling, covers the vector directions of all main roads, secondary roads, and auxiliary roads in the map, excluding non-road areas such as the top status bar, bottom navigation bar, and side menu bar. In this embodiment, the binarization process and ROI extraction process are not specifically limited.

[0067] The preset first feature information characterizes the display features of the road closure icon, including geometric constraints, color features, and texture features. For example, the geometric constraints are set as follows: the aspect ratio of the bounding rectangle is 1:1±0.15, and the area ranges from 32×32 to 128×128 pixels; the color features are set as follows: R value ≥ 200 and G / B ≤ 80 in the RGB three channels, or H∈[0,15]∪[345,360], S≥ 120, V≥ 150 in the HSV space; the texture mode is set as follows: there is a directional horizontal bar or diagonal bar in the central area, and the edge intensity ratio in the horizontal / 45° direction is ≥ 60% detected by the Sobel gradient operator. All icons are identified from the region of interest, and it is determined whether the display features of the identified icons conform to the first feature information. If yes, they are candidate icons; otherwise, they are not candidate icons. For example, a scanning window can be slid across the binarized image behind the ROI mask. The sliding step size can be 4 pixels and the window size can be 64×64. First feature information matching is performed on each pixel within the window, and all rectangular coordinate boxes and confidence scores that meet the conditions are output.

[0068] The preset second feature information characterizes the display features of the closed road segment, including connectivity constraints, color consistency, and dashed line periodicity. Connectivity constraints can be set as follows: the number of pixels in the 8-neighbor connected region ≥ 200 and the aspect ratio ≥ 5:1; color consistency can be set as: more than 90% of pixels on a continuous line segment satisfy R ≥ 180, G ≤ 60, and B ≤ 60; dashed line periodicity can be set as: the center-to-center distance between adjacent red line segments is 12-28 pixels, and the gap width is 8-20 pixels. All road segments are identified from the region of interest (ROI), and it is determined whether the display features of the identified road segments conform to the second feature information. If yes, they are candidate road segments; otherwise, they are not. For example, within the ROI, connected component analysis is performed on the binarized image. After removing noise spots with an area < 200 pixels, the second feature information index is calculated for the remaining connected components. Connected components with a comprehensive index score ≥ 0.7 are retained as candidate road segments.

[0069] The advantage of this setup is that there are a large number of non-road interference elements in the original navigation map image frame, such as the top status bar, bottom operation buttons, advertising pop-ups, and text labels. By binarizing, the key color channels are highlighted and the processing space is limited to the road area, which improves the recognition accuracy of candidate icons and candidate road segments, thereby improving the accuracy of road closure events.

[0070] S402. Based on the first preset condition, determine the road closure icon from each candidate icon, and based on the second preset condition, determine the road closure section from each candidate road section.

[0071] For example, the first preset condition is a criterion for identifying genuine road closure icons from candidate icons. This can be determined by assessing the confidence level of a candidate icon as a road closure icon. For instance, the cropped area of ​​a candidate icon can be normalized to 64×64 pixels, and after grayscale conversion and LBP (Local Binary Pattern) texture encoding, a 128-dimensional feature vector can be generated. Then, the cosine similarity, or confidence level, is calculated with each pre-stored standard road closure icon template. If the similarity with any template is ≥0.9, the first preset condition is considered met. In this embodiment, a lightweight CNN (Convolutional Neural Network) classifier can also be introduced to perform binary classification on the candidate icons, classifying them as either road closure icons or non-road closure icons. An output probability ≥0.95 is considered to meet the first preset condition. Alternatively, OCR recognition results can be further integrated. When there is Chinese text such as "road closure," "construction," or "control" around the candidate icon and the confidence level is ≥0.8, the candidate icon is considered a road closure icon.

[0072] The second precondition is a criterion used to confirm the actual road closure from candidate road segments. It can be determined by assessing the geometric rationality and contextual consistency of the candidate road segments. For example, a simplified DP (Douglas-Peucker) algorithm is applied to the candidate road segments. After retaining key inflection points, the ratio of the total length L of the candidate road segment to the Euclidean distance D between its first and last endpoints is calculated. If L / D > 1.2 and L ≥ 150 pixels, then the geometric rationality check is passed. Next, the pixel distance d from the centroid of the candidate road segment to the centroid of the nearest confirmed road closure icon is calculated. If d ≤ 80 pixels, then the contextual consistency check is passed. If both conditions are met, the second precondition is satisfied.

[0073] The preset first feature information and preset second feature information constitute a parallel coarse screening path, which is applied to two types of heterogeneous targets, icons and road segments, respectively, reducing the computational complexity of subsequent steps. Candidate icons and candidate road segments serve as intermediate products, providing structured input for subsequent fine-tuning. The first preset condition and the second preset condition constitute the fine-tuning mechanism. The former focuses on the semantic authenticity of icons, while the latter takes into account the geometric rationality and spatial context of road segments. Together, they ensure the accuracy and robustness of the recognition results, enabling the system to stably output high-quality sets even in typical noise scenarios such as advertising obstruction, interface button interference, local color shift, and low contrast.

[0074] This embodiment enables hierarchical and differentiated recognition of road closure elements in navigation map image frames, allowing for rapid filtering of icons and road segments. It significantly compresses the search space, reduces processing time, and improves recognition accuracy, providing highly reliable basic data support for subsequent target matching and event generation.

[0075] In this embodiment, determining a road closure icon from candidate icons according to a first preset condition includes: extracting feature vectors of candidate icons from the region of interest; wherein the feature vector is a matrix representing the candidate icon; if the feature vector of a candidate icon satisfies the first preset condition, then the candidate icon is determined as a road closure icon.

[0076] Specifically, feature extraction is performed on the Region of Interest (ROI) to identify areas suspected of being road closure icons, thus determining candidate icons. For example, candidate icons typically possess features such as a red circle with a horizontal bar. The feature matrix of the candidate icons is then determined, meaning the feature vectors of the candidate icons are extracted from the ROI. For instance, the SIFT (Scale-Invariant Feature Transform) tool from the OpenCV library can be used to extract keypoints and descriptors, and the descriptor sequence can be normalized and organized into a fixed-dimensional numerical matrix, i.e., the feature vector.

[0077] A first preset condition is set, which is the condition for determining whether a candidate icon is a road closure icon. If the feature vector of a candidate icon meets the first preset condition, the candidate icon is determined to be a road closure icon; otherwise, it is not a road closure icon. For example, the first preset condition specifies the format of the feature vector of the road closure icon. The system determines whether the feature vector of the candidate icon meets the format specified in the first preset condition. If it does, the first preset condition is met; otherwise, it is not met.

[0078] The advantage of this setup is that it determines the feature matrix of candidate icons and filters them based on the feature matrix, achieving highly robust discrimination of road closure icons in user-reported navigation screenshots, and improving the accuracy and scene adaptability of road closure event recognition.

[0079] In this embodiment, a vector library is pre-set, which stores multiple candidate vectors. The candidate vectors are matrices representing standard road closure icons. If the feature vector of a candidate icon satisfies a first preset condition, the candidate icon is determined as a road closure icon. This includes: for each candidate vector in the vector library, determining the first similarity between the feature vector of the candidate icon and the candidate vector; if the first similarity is greater than a preset first threshold, the feature vector of the candidate icon is determined to satisfy the first preset condition, and the candidate icon is determined as a road closure icon.

[0080] Specifically, the extracted feature matrix can be matched with pre-stored road closure icon templates to calculate the matching confidence of candidate icons with the templates of each road closure icon, thereby filtering out road closure icons.

[0081] The vector library is a structured storage unit used to persistently store high-dimensional feature vectors corresponding to standard road closure icon samples that have been manually labeled and verified by algorithms. The vector library stores multiple candidate vectors, each representing a pre-stored template for a road closure icon. Different candidate vectors represent different templates. The candidate vectors are also in matrix form; that is, the candidate vectors represent a matrix of standard road closure icons.

[0082] The first similarity is a quantitative metric that measures the degree of matching between the feature vector of a candidate icon and any candidate vector in the vector library; it is essentially a matching confidence level. For example, cosine similarity can be used to calculate the first similarity. For each feature vector of a candidate icon, the first similarity between that feature vector and each candidate vector can be calculated. A first threshold is preset, for example, 0.9. For each candidate icon corresponding to multiple first similarities, if any one of the first similarities is greater than the preset first threshold, the feature vector of the candidate icon is considered to meet the first preset condition, and the candidate icon can be identified as a road closure icon. If none of the first similarities of a candidate icon are greater than the first threshold, then the candidate icon is determined not to meet the first preset condition, and the candidate icon is not a road closure icon.

[0083] The advantage of this setting is that, provided that the first similarity between the feature vector of the candidate icon and any candidate vector in the vector library exceeds a preset first threshold, the candidate icon is determined to be a real road closure icon, thereby improving the accuracy of road closure icon recognition and thus improving the accuracy of road closure event determination.

[0084] In this embodiment, the road closure section is determined from each candidate road segment according to the second preset condition, including: extracting the geometric information of the candidate road segment from the region of interest; wherein, the geometric information represents the connectivity and length of the road segment; if the geometric information of the candidate road segment satisfies the second preset condition, the candidate road segment is determined as the road closure section.

[0085] Specifically, candidate icons and candidate road segments can be identified simultaneously, followed by the simultaneous identification of road closure icons and road closure segments. Alternatively, candidate icons and road closure icons can be identified first, and after determining the road closure icon, its coverage area can be subtracted from the image frame to avoid interfering with subsequent extraction of candidate line segments and road closure segments. Red dashed line features can be extracted from the binarized image, and small red spots and other noise can be removed to obtain candidate road segments.

[0086] Geometric information of candidate road segments is extracted from the region of interest (ROI) of the binarized image. This geometric information specifically refers to two core parameters calculated for the connected component of the candidate road segment: connectivity and length. Connectivity can be defined as the topological continuity of the longest skeleton path within the connected component. For example, an 8-neighborhood connectivity traversal can be used to count the proportion of continuous paths without branching; if the proportion is less than 70%, it is considered a break, spur, or isolated spot. Length can be defined as the sum of the Euclidean lengths of the skeleton paths within the connected component, after weighted average filtering to remove endpoint jitter noise. In other words, the continuity of the candidate road segment and its length are determined. In this embodiment, the method for determining the geometric information is not specifically limited.

[0087] A second preset condition is pre-set, which may include conditions for determining connectivity and conditions for determining length. For example, the second preset condition may include connectivity ≥ 0.75, meaning that the proportion of continuous segments without branching in the skeleton path is not less than 75% of the total skeleton length; it may also include length ≥ L_min, where L_min is a preset length threshold. For each candidate road segment, it is determined whether the geometric information of the candidate road segment meets the second preset condition. If yes, the candidate road segment is determined to be a road closure segment; if not, the candidate road segment is determined not to be a road closure segment. For example, if the connectivity ratio of a candidate road segment is greater than a preset ratio threshold and the length is greater than a preset length threshold, then the second preset condition is met.

[0088] The advantage of this setup is that, based on the initial extraction of candidate road segments, road closure segments are filtered by connectivity and length, which improves the authenticity and usability of road closure segment identification, thereby increasing the accuracy of road closure event determination.

[0089] S403. Determine the target matching relationship based on the location information of each road closure icon and each road closure segment in the image frame; wherein, the target matching relationship represents the matching relationship between the road closure icon and the road closure segment.

[0090] For example, this step can refer to step S102 above, and will not be repeated here.

[0091] S404. Based on the target matching relationship, determine the target road segment from each road closure segment; wherein, the target road segment represents the actual road closure segment in the road corresponding to the image frame.

[0092] For example, this step can refer to step S103 above, and will not be repeated here.

[0093] S405. Based on the location information of the target road segment, generate road closure event information and store it; wherein, the road closure event information represents the road closure status of the road corresponding to the image frame.

[0094] For example, this step can refer to step S104 above, and will not be repeated here.

[0095] This application provides a method for identifying road closure events based on navigation maps. It acquires image frames from navigation maps and identifies road closure icons and road segments from these frames, automating the capture of key visual elements in user-reported screenshots. Based on the location information of each road closure icon and road segment in the image frame, corresponding road closure icons and segments are determined, establishing a target matching relationship between them. According to this target matching relationship, target road segments are identified from each road closure segment, effectively filtering out actually existing road closures and avoiding interference from falsely detected segments. Finally, road closure event information is automatically generated based on the location information of the target road segments. Automatic association and filtering using location information avoids the tedious process of manually analyzing screenshots, significantly shortening the information processing cycle. The establishment of target matching relationships ensures spatial consistency between road closure icons and road segments, improving the accuracy and efficiency of road closure event identification.

[0096] Figure 5 A flowchart illustrating a method for identifying road closure events based on navigation maps, as provided in this application embodiment, is shown below. Figure 5 As shown, this embodiment, based on the above embodiments, provides a detailed description of a method for identifying road closure events based on navigation maps. The method includes:

[0097] S501. Obtain image frames of the navigation map and identify road closure icons and road closure sections from the image frames.

[0098] For example, this step can refer to step S101 above, and will not be repeated here.

[0099] S502. Determine the target matching relationship based on the location information of each road closure icon and each road closure segment in the image frame; wherein, the target matching relationship represents the matching relationship between the road closure icon and the road closure segment.

[0100] For example, this step can refer to step S102 above, and will not be repeated here.

[0101] S503. For each road closure icon, determine the road segment corresponding to the road closure icon based on the target matching relationship.

[0102] For example, in the target matching relationship, one road closure icon corresponds to one road closure segment. A road closure icon refers to a graphic element extracted from a navigation map image frame that possesses typical visual features (such as a red circle with a white horizontal bar, a yellow triangle with a black "No Entry" symbol, etc.); a road closure segment refers to a road section in the same image frame, determined after candidate segment screening and geometric verification, and presented as a continuous red dashed or solid line. For each road closure icon in the target matching relationship, the road closure segment matching that icon is determined.

[0103] S504. Determine the geometric information of the road segment corresponding to the road closure icon from the image frame; wherein, the geometric information represents the connectivity and length of the road segment.

[0104] For example, for a road closure segment that matches a road closure icon, the geometric information of that road closure segment is extracted. The geometric information can characterize the connectivity and length of the road closure segment, that is, determine whether the road closure segment is continuous and determine the length of the road closure segment.

[0105] S505. If the geometric information of the road segment corresponding to the road closure icon meets the third preset condition, then the road segment corresponding to the road closure icon is determined as the target road segment.

[0106] For example, a third preset condition is pre-set to determine the validity of a road closure segment. If the geometric information of the road closure segment corresponding to the road closure icon meets the third preset condition, then the road closure segment is determined to be a valid road closure segment, that is, the road closure segment is determined to be the target road segment. For example, the third preset condition may include a length constraint and a connectivity constraint. The length constraint requires that the length of the road closure segment falls within a preset length range; the connectivity constraint requires that the number of connected components of the road closure segment is 1. If the geometric information of the road closure segment corresponding to the road closure icon does not meet the third preset condition, then the road closure segment is determined not to be the target road segment.

[0107] In this embodiment, road closures in the image frame are effectively filtered to improve the accuracy of road closure identification, avoid false reporting of road closure events, and improve the business availability and geographical accuracy of road closure event identification results.

[0108] In this embodiment, if the geometric information of the road segment corresponding to the road closure icon meets a third preset condition, then the road segment corresponding to the road closure icon is determined as the target road segment. This includes: obtaining the actual information corresponding to the road closure icon from a preset database; wherein, the actual information represents the geometric information of the road segment actually corresponding to the road closure icon; determining a second similarity between the geometric information of the road segment corresponding to the road closure icon and the actual information corresponding to the road closure icon; if the second similarity is greater than a preset second threshold, then determining that the geometric information of the road segment corresponding to the road closure icon meets the third preset condition, and determining the road segment corresponding to the road closure icon as the target road segment.

[0109] Specifically, the pre-defined database is a structured geographic semantic knowledge base that stores the geometric information of the actual road closures that different road closure icons should correspond to. This actual geometric information is referred to as the actual information. This pre-defined database is constructed through a triple mechanism of historical manual annotation, high-precision map backtesting, and crowdsourced verification. Each record contains a unique icon identifier to represent the type of road closure icon. For example, in an urban expressway scenario, the standard length range of a certain type of red circle with a horizontal bar icon in the corresponding actual information is [180m, 350m], and the average curvature is 0.02 rad / m.

[0110] For each road closure icon in the target matching relationship, determine its icon identifier, that is, determine the type of road closure icon. For example, the type of road closure icon can be determined based on information such as its color and shape. Based on the icon identifier, retrieve the actual information corresponding to the road closure icon from a preset database, that is, determine the geometric information that the road segment corresponding to the road closure icon should possess.

[0111] The geometric information of the road closure segment corresponding to the road closure icon in the image frame is determined. The similarity between the geometric information of the road closure segment corresponding to the road closure icon and the actual information corresponding to the road closure icon is calculated as a second similarity. For example, the second similarity can be a weighted cosine similarity. This embodiment does not specifically limit the calculation method of the second similarity.

[0112] A second threshold is preset, which is a configurable confidence criterion and can be set to 0.8. The second similarity is compared with the second threshold. If the second similarity is greater than the preset second threshold, it is determined that the geometric information of the road segment corresponding to the road closure icon meets the third preset condition, that is, the pre-stored geometric information is consistent with the extracted geometric information, and the road segment corresponding to the road closure icon can be identified as the target road segment; if the second similarity is less than or equal to the preset second threshold, it is determined that the road segment corresponding to the road closure icon is not the target road segment.

[0113] The beneficial effect of this setting is that it obtains the standard geometric prior corresponding to the road closure icon in the real geographical scene from the preset database. When the geometric information of the identified road closure segment deviates significantly from the prior knowledge, it can actively identify the anomaly and refuse to output it, thereby avoiding writing the erroneous segment as the target road segment into the road closure event information, and significantly improving the accuracy of the final generated road closure event information.

[0114] S506. Based on the location information of the target road segment, generate road closure event information and store it; wherein, the road closure event information represents the road closure status of the road corresponding to the image frame.

[0115] For example, this step can refer to step S104 above, and will not be repeated here.

[0116] This application provides a method for identifying road closure events based on navigation maps. It acquires image frames from navigation maps and identifies road closure icons and road segments from these frames, automating the capture of key visual elements in user-reported screenshots. Based on the location information of each road closure icon and road segment in the image frame, corresponding road closure icons and segments are determined, establishing a target matching relationship between them. According to this target matching relationship, target road segments are identified from each road closure segment, effectively filtering out actually existing road closures and avoiding interference from falsely detected segments. Finally, road closure event information is automatically generated based on the location information of the target road segments. Automatic association and filtering using location information avoids the tedious process of manually analyzing screenshots, significantly shortening the information processing cycle. The establishment of target matching relationships ensures spatial consistency between road closure icons and road segments, improving the accuracy and efficiency of road closure event identification.

[0117] Figure 6 A schematic diagram of a road closure event identification device based on a navigation map, provided in an embodiment of this application, is shown below. Figure 6 As shown, the road closure event identification device 60 based on a navigation map provided in this embodiment includes:

[0118] The image acquisition unit 601 is used to acquire image frames of the navigation map and identify road closure icons and road closure sections from the image frames;

[0119] The relationship determination unit 602 is used to determine the target matching relationship based on the position information of each road closure icon and the position information of each road closure segment in the image frame; wherein, the target matching relationship represents the matching relationship between the road closure icon and the road closure segment;

[0120] The road segment determination unit 603 is used to determine the target road segment from each road closure segment according to the target matching relationship; wherein, the target road segment represents the actual road closure segment in the road corresponding to the image frame;

[0121] The event generation unit 604 is used to generate road closure event information based on the location information of the target road segment and store it; wherein, the road closure event information represents the road closure status of the road corresponding to the image frame.

[0122] In one possible implementation, the image acquisition unit 601 includes:

[0123] The candidate recognition module is used to identify candidate icons from an image frame based on preset first feature information, and to identify candidate road segments from an image frame based on preset second feature information; wherein, the preset first feature information represents the display characteristics of the road closure icon, the preset second feature information represents the display characteristics of the road closure segment, the candidate icon is a preliminarily selected icon, and the candidate road segment is a preliminarily selected road segment;

[0124] The road closure determination module is used to determine the road closure icon from each candidate icon according to a first preset condition, and to determine the road closure section from each candidate road section according to a second preset condition.

[0125] In one possible implementation, the candidate identification module is specifically used for:

[0126] The image frame is binarized to obtain a binarized image, and the region of interest is extracted from the binarized image; where the region of interest represents the area where the road is located in the image frame;

[0127] Based on the preset first feature information, candidate icons are identified from the region of interest, and based on the preset second feature information, candidate road segments are identified from the region of interest.

[0128] In one possible implementation, the road closure determination module is specifically used for:

[0129] Extract feature vectors of candidate icons from image frames; where the feature vector is a matrix representing the candidate icon;

[0130] If the feature vector of a candidate icon satisfies the first preset condition, then the candidate icon is determined as a road closure icon.

[0131] In one possible implementation, a vector library is pre-set, storing multiple candidate vectors, which are matrices representing standard road closure icons; the road closure determination module is specifically used for:

[0132] For each candidate vector in the vector library, determine the first similarity between the feature vector of the candidate icon and the candidate vector.

[0133] If there exists a first similarity greater than a preset first threshold, then the feature vector of the candidate icon is determined to satisfy the first preset condition, and the candidate icon is determined to be a road closure icon.

[0134] In one possible implementation, the road closure determination module is specifically used for:

[0135] Extract geometric information of candidate road segments from image frames; whereby geometric information represents the connectivity and length of road segments;

[0136] If the geometric information of a candidate road segment meets the second preset condition, the candidate road segment will be designated as a road closure segment.

[0137] In one possible implementation, the road segment determination unit 603 includes:

[0138] The first determination module is used to determine the road segment corresponding to each road closure icon based on the target matching relationship.

[0139] The second determining module is used to determine the geometric information of the road segment corresponding to the road closure icon from the image frame; wherein, the geometric information represents the connectivity and length of the road segment;

[0140] The third determination module is used to determine the road closure segment corresponding to the road closure icon as the target road segment if the geometric information of the road closure segment corresponding to the road closure icon meets the third preset condition.

[0141] In one possible implementation, the third determining module is specifically used for:

[0142] Retrieve the actual information corresponding to the road closure icon from the preset database; whereby the actual information represents the geometric information of the road segment actually closed by the road closure icon.

[0143] Determine the second similarity between the geometric information of the road segment corresponding to the road closure icon and the actual information corresponding to the road closure icon;

[0144] If the second similarity is greater than the preset second threshold, then the geometric information of the road closure segment corresponding to the road closure icon is determined to meet the third preset condition, and the road closure segment corresponding to the road closure icon is determined as the target road segment.

[0145] This embodiment provides a road closure event identification device based on a navigation map, which can execute the method provided in the above method embodiment. Its implementation principle and technical effect are similar, and will not be described in detail here.

[0146] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 7 As shown, the electronic device 70 provided in this embodiment includes at least one processor 701 and a memory 702. Optionally, the device 70 further includes a communication component 703. The processor 701, memory 702, and communication component 703 are connected via a bus 704.

[0147] In a specific implementation, at least one processor 701 executes computer execution instructions stored in memory 702, causing at least one processor 701 to perform the above-described method.

[0148] The specific implementation process of processor 701 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0149] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0150] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0151] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0152] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0153] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method.

[0154] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0155] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0156] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0157] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0158] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0159] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0160] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0161] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A method for identifying road closure events based on navigation maps, characterized in that, include: Acquire image frames from the navigation map and identify road closure icons and road closure sections from the image frames; Based on the location information of each road closure icon and the location information of each road closure segment in the image frame, a target matching relationship is determined; wherein, the target matching relationship represents the matching relationship between the road closure icon and the road closure segment; Based on the target matching relationship, a target road segment is determined from each of the road closure segments; wherein, the target road segment represents the actual road closure segment existing in the road corresponding to the image frame; Based on the location information of the target road segment, road closure event information is generated and stored; wherein, the road closure event information represents the road closure status of the road corresponding to the image frame.

2. The method according to claim 1, characterized in that, Identifying road closure icons and road closure sections from the image frames includes: Candidate icons are identified from the image frame based on preset first feature information, and candidate road segments are identified from the image frame based on preset second feature information; wherein, the preset first feature information characterizes the display characteristics of the road closure icon, the preset second feature information characterizes the display characteristics of the road closure segment, the candidate icon is a preliminarily selected icon, and the candidate road segment is a preliminarily selected road segment; Based on a first preset condition, the road closure icon is determined from each of the candidate icons, and based on a second preset condition, the road closure section is determined from each of the candidate road sections.

3. The method according to claim 2, characterized in that, Identifying candidate icons from the image frame based on preset first feature information, and identifying candidate road segments from the image frame based on preset second feature information, including: The image frame is binarized to obtain a binarized image, and a region of interest is extracted from the binarized image; wherein, the region of interest represents the area where the road is located in the image frame; Candidate icons are identified from the region of interest based on a preset first feature information, and candidate road segments are identified from the region of interest based on a preset second feature information.

4. The method according to claim 2, characterized in that, Based on a first preset condition, the road closure icon is determined from the candidate icons, including: Extract the feature vector of the candidate icon from the image frame; wherein the feature vector is a matrix representing the candidate icon; If the feature vector of the candidate icon satisfies the first preset condition, then the candidate icon is determined as the road closure icon.

5. The method according to claim 4, characterized in that, A vector library is pre-set, which stores multiple candidate vectors, and the candidate vectors are matrices representing standard road closure icons; If the feature vector of the candidate icon satisfies the first preset condition, then the candidate icon is determined as the road closure icon, including: For each candidate vector in the vector library, determine the first similarity between the feature vector of the candidate icon and the candidate vector; If the first similarity is greater than a preset first threshold, then the feature vector of the candidate icon is determined to satisfy the first preset condition, and the candidate icon is determined to be the road closure icon.

6. The method according to claim 2, characterized in that, Based on the second preset condition, the road closure sections are determined from the candidate road sections, including: Geometric information of the candidate road segment is extracted from the image frame; wherein the geometric information represents the connectivity and length of the road segment; If the geometric information of the candidate road segment is determined to meet the second preset condition, then the candidate road segment is determined as the road closure segment.

7. The method according to any one of claims 1-6, characterized in that, Based on the target matching relationship, the target road segment is determined from each of the road closure segments, including: For each road closure icon, the road segment corresponding to the road closure icon is determined according to the target matching relationship; The geometric information of the road segment corresponding to the road closure icon is determined from the image frame; wherein the geometric information represents the connectivity and length of the road segment; If the geometric information of the road segment corresponding to the road closure icon meets the third preset condition, then the road segment corresponding to the road closure icon is determined as the target road segment.

8. The method according to claim 7, characterized in that, If the geometric information of the road segment corresponding to the road closure icon meets the third preset condition, then the road segment corresponding to the road closure icon is determined as the target road segment, including: The actual information corresponding to the road closure icon is obtained from a preset database; wherein, the actual information represents the geometric information of the road segment actually corresponding to the road closure icon; Determine the second similarity between the geometric information of the road segment corresponding to the road closure icon and the actual information corresponding to the road closure icon; If the second similarity is greater than the preset second threshold, then the geometric information of the road closure segment corresponding to the road closure icon is determined to meet the third preset condition, and the road closure segment corresponding to the road closure icon is determined as the target road segment.

9. A road closure event identification device based on a navigation map, characterized in that, include: The image acquisition unit is used to acquire image frames of the navigation map and identify road closure icons and road closure sections from the image frames; The relationship determination unit is used to determine the target matching relationship based on the position information of each road closure icon and the position information of each road closure segment in the image frame; wherein, the target matching relationship represents the matching relationship between the road closure icon and the road closure segment; A road segment determination unit is used to determine a target road segment from each of the road closure segments according to the target matching relationship; wherein, the target road segment represents the actual road closure segment existing in the road corresponding to the image frame; An event generation unit is used to generate road closure event information based on the location information of the target road segment and store it; wherein the road closure event information represents the road closure status of the road corresponding to the image frame.

10. An electronic device / computer-readable storage medium / computer program product, characterized in that, The electronic device includes: a memory, a processor; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory, causing the processor to perform the method as described in any one of claims 1-8; and / or, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1-8; and / or, The computer program product includes a computer program that, when executed by a processor, implements the method described in any one of claims 1-8.

Images