Signboard discoloration detection method and system based on dynamic patrol

Abstract

This invention discloses a method and system for detecting road sign fading based on dynamic inspection. The method includes: acquiring video of road signs on both sides of a road and recording time and location information; preprocessing the road sign video to obtain a preprocessing result; identifying and tracking each road sign in the preprocessing result, constructing a time series of the road sign's continuous appearance in different frames to obtain a tracking sequence; based on the tracking sequence, dividing each road sign into multiple functional areas, calculating the color features of each functional area, comparing them with standard samples, and marking suspected fading areas to obtain suspected fading markers; summarizing the suspected fading markers in all frames to determine whether the road sign is continuously fading; when the road sign is continuously fading, verifying the reflectivity, readability, and texture sharpness of the suspected fading areas, and determining the fading level based on the verification results. By implementing the method of this invention, the degree of fading of road signs can be accurately and stably identified and evaluated.

Description

Technical Field

[0001] This invention relates to computer vision, and more specifically to a method and system for detecting fading of road signs based on dynamic inspection. Background Technology

[0002] As an indispensable information transmission tool in the road traffic system, road signs rely heavily on their color and reflective properties for drivers to correctly and quickly read traffic information. However, due to long-term exposure to the outdoor environment, road signs are inevitably affected by various factors such as ultraviolet radiation, acid rain erosion, and vehicle exhaust pollution, leading to aging of the reflective film, fading of the pigment layer, and surface corrosion, resulting in color deviation and reduced reflectivity. Especially at night or in environments with complex lighting conditions, these problems can increase the difficulty of reading road signs, thus creating safety hazards.

[0003] Currently, the inspection and maintenance of road signs mainly rely on manual patrols. Inspectors directly observe whether road signs have faded or have poor reflectivity. However, this method suffers from low inspection frequency, high labor intensity, and strong subjectivity, making it difficult to meet the needs of dynamic monitoring of urban road facilities. With the advancement of computer vision technology and intelligent transportation systems, some research has begun to explore the use of image recognition algorithms to achieve automatic detection and visibility assessment of traffic signs. For example, deep learning models are used to identify, locate, and analyze the visibility of road signs, as well as automatic identification methods for road marking wear and damage. Although some progress has been made, most of these studies focus on sign detection and classification or visibility assessment, and the automated detection of the specific quality level of "color degradation and reflectivity deterioration" remains insufficient. Especially when dealing with the "fading" feature, existing methods are usually based on single-frame images, which are easily affected by factors such as ambient lighting, shooting angle, and equipment differences, lacking stability and objectivity.

[0004] Therefore, it is necessary to design a new method to accurately and stably identify and assess the degree of fading of road signs, and to solve the problems of low efficiency, strong subjectivity, insufficient stability and lack of comprehensive evaluation mechanism in the current technology. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method and system for detecting road sign fading based on dynamic inspection.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a road sign fading detection method based on dynamic inspection, comprising:

[0007] Acquire videos of road signs on both sides of the road and record the time and location information;

[0008] The road sign video is preprocessed to obtain the preprocessing result;

[0009] Identify and track each road sign in the preprocessing results, and construct a time series of the road signs' continuous appearance in different frames to obtain a tracking sequence;

[0010] Based on the tracking sequence, each road sign is divided into multiple functional areas, the color features of each functional area are calculated, and compared with standard samples to mark suspected fading areas to obtain suspected fading marks;

[0011] Summarize all suspected faded markers in all frames to determine if the road sign is fading continuously;

[0012] When the road sign is continuously fading, the reflectivity, readability, and texture sharpness of the suspected fading area are verified, and the fading level is determined based on the verification results.

[0013] The further technical solution is as follows: The step of summarizing suspected fading markers in all frames to determine whether the road sign is continuously fading includes:

[0014] Collect color difference values ​​and quality scores for each functional area in each frame, and record timestamps and suspected fading markers;

[0015] Calculate the proportion of suspected faded frames in each functional area, adjust the weights according to the quality score, and calculate the weighted proportion to obtain the proportion of suspected faded frames.

[0016] Based on the proportion of suspected faded frames, a threshold is set to determine whether the corresponding functional area is continuously fading;

[0017] When the corresponding functional area shows continuous fading, the functional area corresponding to the road sign is marked as continuously fading.

[0018] The further technical solution is as follows: The calculation of the proportion of suspected faded frames in each functional area, and the adjustment of weights based on the quality score to calculate the weighted proportion, to obtain the proportion of suspected faded frames, includes:

[0019] Calculate the sum of suspected faded markers for each functional region in the tracking sequence, and divide the sum by the total number of frames for each functional region in the tracking sequence to obtain the proportion of suspected faded frames for each functional region;

[0020] The weights of the corresponding frames are adjusted based on the quality score.

[0021] The weights are weighted and accumulated based on the proportion of suspected faded frames in each functional area to obtain the proportion of suspected faded frames.

[0022] The further technical solution is as follows: the step of determining whether there is continuous fading in the corresponding functional area based on the percentage of suspected faded frames by setting a threshold includes:

[0023] If the proportion of suspected faded frames exceeds a set threshold, and the fluctuation of the suspected faded marker in the tracking sequence meets the requirements, then it is determined that the corresponding functional area has continuous fading.

[0024] The further technical solution is as follows: Based on the tracking sequence, each road sign is divided into multiple functional areas, the color features of each functional area are calculated, and compared with standard samples to mark suspected fading areas, thereby obtaining suspected fading marks, including:

[0025] Based on the design specifications for traffic signs, the road sign images in the tracking sequence are segmented into background areas, text areas, and symbol areas to obtain multiple functional areas;

[0026] Calculate the average color value, standard deviation, and dominant color distribution for each functional area to obtain color statistical characteristics;

[0027] Based on the corresponding standard samples in the database, the color difference value is calculated for the color statistical features;

[0028] Based on the color difference value and the preset color difference value threshold, suspected fading areas with significant color deviation are identified and marked to obtain suspected fading marks.

[0029] Its further technical solution is as follows: Based on the design specifications for traffic signs, the road sign images in the tracking sequence are segmented into background areas, text areas, and symbol areas to obtain multiple functional areas, including:

[0030] Based on traffic sign design specifications, template matching or deep learning methods are used to segment the road sign images in the tracking sequence into background areas, text areas, and symbol areas to obtain multiple functional areas.

[0031] The further technical solution is as follows: when the road sign is continuously fading, the reflectivity, readability, and texture sharpness of the suspected fading area are verified, and the fading level is determined based on the verification results, including:

[0032] The intensity of reflected light in the suspected faded area is detected at night or in a strong light environment, the reflectivity recovery ratio is calculated, and the fading is classified according to the calculation results to obtain the reflectivity performance verification results.

[0033] The text content in the suspected faded area is identified, the recognition rate is calculated, and the readability of the text content in the suspected faded area is classified based on the recognition rate to obtain the text readability verification result.

[0034] The aging status of the surface coating in the suspected fading area is analyzed, the texture sharpness is evaluated, and the suspected fading area is classified according to the texture clarity to obtain the texture status verification result.

[0035] Based on the results of the reflectivity verification, text readability verification, and texture status verification, each suspected fading area is comprehensively evaluated to determine the fading level.

[0036] The further technical solution is as follows: the preprocessing of the road sign video to obtain the preprocessing result includes:

[0037] The image sequence is extracted frame by frame from the road sign video and numbered and synchronized according to time.

[0038] Automatically detect and remove blurry, jittery, or poorly exposed frames to obtain the removal results;

[0039] The removal result is adjusted by distortion correction and geometric transformation to obtain a preprocessed result, wherein the preprocessed result includes the processed image frame and the corresponding time and position information.

[0040] The further technical solution is as follows: Identifying and tracking each road sign in the preprocessing result, and constructing a time series of the road signs' continuous appearance in different frames to obtain a tracking sequence, includes:

[0041] The road sign features in each frame of the preprocessed image are analyzed and located using a deep learning object detection algorithm to obtain the road sign;

[0042] Using a target tracking mechanism, a tracking ID is assigned to each road sign by comparing the position, size, direction of movement, and color similarity between adjacent frames;

[0043] For each road sign being tracked, an observation sequence is created containing the cropped image, timestamp, and location information to obtain the tracking sequence.

[0044] This invention also provides a road sign fading detection system based on dynamic inspection, comprising:

[0045] The acquisition unit is used to acquire video of road signs on both sides of the road and record the time and location information;

[0046] A preprocessing unit is used to preprocess the road sign video to obtain a preprocessing result;

[0047] The tracking sequence construction unit is used to identify and track each road sign in the preprocessing result, and construct a time series of the road signs continuously appearing in different frames to obtain the tracking sequence;

[0048] A marking unit is used to divide each road sign into multiple functional areas based on the tracking sequence, calculate the color features of each functional area, compare them with standard samples, and mark suspected fading areas to obtain suspected fading marks.

[0049] The persistence determination unit is used to summarize the suspected fading marks in all frames to determine whether the road sign is fading continuously;

[0050] The fine verification unit is used to verify the reflectivity, readability, and texture sharpness of the suspected faded area when the road sign is continuously fading, and to determine the fading level based on the verification results.

[0051] The advantages of this invention compared to existing technologies are as follows: This invention acquires and preprocesses video data of road signs on both sides of the road, combines time and location information, and uses advanced image processing technology to identify and track each road sign, constructing its time-series performance in different frames. Subsequently, each road sign is divided into multiple functional areas, color features are calculated and compared with standard samples to mark suspected fading areas; the marks in all frames are summarized to determine whether the road sign is continuously fading, and the suspected areas are comprehensively verified for reflectivity, readability, and texture sharpness to finally determine the fading level. This method achieves accurate and stable identification and evaluation of the degree of road sign fading, effectively solving the problems of low efficiency, strong subjectivity, insufficient stability, and lack of multi-dimensional comprehensive evaluation mechanisms in existing technologies, providing a scientific basis and technical support for the maintenance of road traffic facilities.

[0052] The present invention will be further described below with reference to the accompanying drawings and specific embodiments. Attached Figure Description

[0053] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0054] Figure 1 A flowchart illustrating the road sign fading detection method based on dynamic inspection provided in an embodiment of the present invention;

[0055] Figure 2 A schematic diagram of the tracking sequence provided in the embodiments of the present invention. Figure 1 ;

[0056] Figure 3 A schematic diagram of the tracking sequence provided in the embodiments of the present invention. Figure 2 ;

[0057] Figure 4 This is a schematic diagram of regional color contrast provided in an embodiment of the present invention;

[0058] Figure 5 This is a partial schematic diagram of regional color contrast provided in an embodiment of the present invention;

[0059] Figure 6 This is a detailed schematic diagram of a road sign provided in an embodiment of the present invention;

[0060] Figure 7 A schematic block diagram of a road sign fading detection system based on dynamic patrol provided in an embodiment of the present invention;

[0061] Figure 8 A schematic block diagram of a computer device provided for an embodiment of the present invention. Detailed Implementation

[0062] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0063] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0064] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0065] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0066] Please see Figure 1 , Figure 1This is a flowchart illustrating a road sign fading detection method based on dynamic inspection, provided in an embodiment of the present invention. This method is applied to a server. By acquiring and preprocessing videos of road signs on both sides of the road, a deep learning object detection algorithm is used to identify and track the road signs in each frame. A time series is constructed to track the changes of each road sign. Based on traffic sign design specifications, the road signs are divided into multiple functional areas, and color features are calculated and compared with standard samples to mark suspected fading areas. Subsequently, the suspected fading marks in all frames are summarized, and a set threshold is used to determine whether persistent fading exists. For areas confirmed as persistently fading, reflectivity, readability, and texture sharpness are verified to determine the fading level. This method solves the problems of low efficiency, strong subjectivity, insufficient stability, and lack of a comprehensive evaluation mechanism in existing technologies, achieving accurate and stable identification and evaluation of the degree of road sign fading, and improving the scientificity and effectiveness of road sign maintenance and management.

[0067] Figure 1 This is a flowchart illustrating the road sign fading detection method based on dynamic inspection provided in an embodiment of the present invention. Figure 1 As shown, the method includes the following steps S110 to S160.

[0068] S110: Acquire video of road signs on both sides of the road and record the time and location information.

[0069] In this embodiment, road sign video refers to video data of road signs on both sides of the road continuously collected by a high-definition camera installed on the vehicle during normal vehicle operation. This process not only enables dynamic acquisition of road sign status but also effectively covers a large-scale road network, avoiding the inefficiency and limitations of traditional manual point-by-point inspections.

[0070] First, install high-resolution (e.g., 1080p or higher) in-vehicle cameras. These cameras can maintain stable images under different lighting conditions, ensuring that the captured video is clear enough for subsequent analysis.

[0071] As the vehicle travels along the predetermined route, the onboard cameras begin to operate, continuously capturing images of road signs on both sides of the road. Simultaneously, the system records the timestamp and approximate geographical location information for each frame of video in real time. This information is crucial for subsequent temporal-spatial correlation of images and scenes, helping to accurately determine the location of each road sign and its changes.

[0072] The collected video data is immediately tagged with timestamps and location information, forming a complete and traceable video sequence. This step ensures that even in subsequent processing, the geographical location and shooting time of each video segment can be accurately traced, greatly improving the efficiency and accuracy of data analysis.

[0073] To ensure the efficiency of subsequent processing and the reliability of results, the system performs basic preprocessing operations at the edge. For example, it automatically detects and removes overexposed, jittery, or blurry frames, or performs lightweight compression for easier uploading and storage. The aim is to ensure that the sample frames used for analysis are clear, stable, and complete, providing a high-quality data foundation for subsequent image preprocessing and road sign recognition.

[0074] In summary, step S110 is the foundation of the entire vehicle-mounted video-based road sign fading detection mechanism. Through efficient automated data acquisition, it provides essential raw data support for subsequent color stability analysis, regional color comparison, persistent fading determination, and fine-grained verification. This method significantly improves detection efficiency, reduces labor costs, and enhances the objectivity and accuracy of the detection results.

[0075] S120. Preprocess the road sign video to obtain the preprocessing result.

[0076] In this embodiment, the preprocessing result refers to a high-quality, standardized set of image frames obtained after a series of image processing steps. These image frames are not only clear and stable, but also contain corresponding timestamps and location information, providing reliable input data for subsequent road sign detection, tracking, and fading analysis.

[0077] In one embodiment, step S120 described above may include steps S121 to S123.

[0078] S121. Extract the image sequence frame by frame from the road sign video, and number and synchronize them according to time.

[0079] In this embodiment, this step involves breaking down the raw video stream into individual image frames and numbering and synchronizing each frame according to its capture time. This process ensures that all image frames are accurately arranged in chronological order, facilitating precise association with specific time points and geographical locations during subsequent processing.

[0080] The system first analyzes the video stream captured by the vehicle-mounted camera frame by frame, extracting a series of image frames.

[0081] Each frame of the image is assigned a precise timestamp, which records the exact moment the image was captured.

[0082] To facilitate management and processing, each image frame is assigned a unique number for easy retrieval and referencing in subsequent operations.

[0083] S122. Automatically detect and remove blurry, jittery, or poorly exposed frames to obtain the removal result.

[0084] In this embodiment, the removal result refers to the set of image frames that have had most of the quality problems removed.

[0085] In this step, the system automatically identifies and removes image frames that are of low quality due to poor shooting conditions (such as vibrations during vehicle movement, changes in lighting, etc.). This significantly improves the accuracy of subsequent analysis and reduces errors caused by low-quality data.

[0086] By calculating inter-frame differences and edge sharpness indicators, the system automatically evaluates the quality of each frame of the image, paying particular attention to whether there are problems such as blurring, jitter, or overexposure.

[0087] Based on the above evaluation results, the system will automatically filter out frames that do not meet the quality requirements and retain only clear and stable images for subsequent processing.

[0088] After this step, what you get is a set of image frames with most quality issues removed, which is the "removal result".

[0089] S123. The removal result is adjusted by distortion correction and geometric transformation to obtain a preprocessing result, wherein the preprocessing result includes the processed image frame and the corresponding time and position information.

[0090] The final step is to further optimize the selected image frames, mainly including distortion correction and geometric transformation, to eliminate the effects of lens distortion and correct image deformation caused by changes in viewing angle.

[0091] To address potential lens distortions in vehicle-mounted cameras (such as barrel or pincushion distortion), the system performs corresponding distortion correction processing on the image, making the shape of the road sign closer to reality.

[0092] Taking into account the changes in the vehicle's perspective during driving, the system also needs to perform slight geometric transformations or perspective adjustments on the image to maintain the consistency of the road sign's shape across different frames.

[0093] After completing all the above steps, the final result is a collection of high-quality, standardized image frames, each with its own timestamp and location information. This is the so-called "preprocessing result".

[0094] In summary, before entering the analysis stage, the video sequences acquired by the vehicle undergo image preprocessing to ensure the accuracy and stability of subsequent recognition and analysis. The system parses the original video frames one by one, extracts the image frame sequence, and numbers and synchronizes them according to timestamps. Subsequently, a series of standardization and quality optimization operations are performed on each frame, including but not limited to:

[0095] Sharpness and stability screening: Frames that are blurry, jittery, or severely overexposed are automatically removed by calculating the inter-frame difference and edge sharpness index, ensuring that the sample frames used for analysis are clear, stable, and complete.

[0096] Distortion and perspective correction: The lens distortion of the vehicle camera is corrected, and the image is slightly geometrically transformed or perspective adjusted according to the vehicle's driving direction and perspective changes to maintain the basic consistency of the road sign shape.

[0097] After this processing stage, the video data is transformed into a standardized set of high-quality image frames, retaining their corresponding time and location metadata. This provides a reliable visual input foundation for the next stage of road sign detection and localization. In other words, meticulous processing and optimization of the raw video data lays a solid foundation for subsequent road sign recognition, tracking, and fading analysis. This not only improves detection efficiency but also enhances the accuracy and reliability of the detection results.

[0098] S130. Identify and track each road sign in the preprocessing results, and construct a time series of the road signs continuously appearing in different frames to obtain a tracking sequence.

[0099] In this embodiment, the tracking sequence refers to a series of continuous records containing cropped images, timestamps, and location information created for each identified road sign in the video through deep learning object detection and tracking mechanisms, in order to achieve continuous tracking and state analysis of the road signs in different frames.

[0100] like Figure 2 as well as Figure 3 As shown, through deep learning object detection algorithms and object tracking mechanisms, accurate identification, localization, and cross-frame association of each road sign appearing in the video are achieved, forming a tracking sequence containing multi-dimensional information (such as image, timestamp, and location). This process first applies a deep learning object detection algorithm to analyze each frame, accurately locating the road sign based on its shape, edges, color, and text features, and outputting bounding box coordinates. Next, through the object tracking mechanism, a unique tracking ID is assigned to each physical road sign based on comparisons of position, size, motion direction, and color similarity between adjacent frames, ensuring cross-frame identity consistency and the formation of the observation sequence. Each tracking sequence contains a cropped image, timestamp, and location information, recording the road sign's performance under different conditions, providing data support for subsequent analysis. Furthermore, the system evaluates the quality of each trajectory, marking sequences with low quality or severe frame drops as low-confidence sequences, and appropriately downweighting or removing them in further analysis to ensure the reliability and stability of the final evaluation results. This series of steps collectively achieves efficient detection and accurate tracking of road signs in the video.

[0101] In one embodiment, step S130 described above may include steps S131 to S133.

[0102] S131. Use a deep learning object detection algorithm to analyze and locate the road sign features in each frame of the preprocessed image to obtain the road sign.

[0103] In this embodiment, this step involves using advanced computer vision technology to automatically identify and locate road signs in each frame of an image.

[0104] The system employs pre-trained deep learning models that can effectively identify various objects in images, with specific optimizations for road signs. The model analyzes each frame of the image to find regions that might belong to a road sign.

[0105] Once a potential road sign area is detected, the system further analyzes its shape, edges, color, and text features to ensure accurate positioning. The final output is the exact bounding box coordinates of each road sign, providing basic data support for subsequent steps.

[0106] S132. Using a target tracking mechanism, a tracking ID is assigned to each road sign by comparing the position, size, direction of movement, and color similarity between adjacent frames.

[0107] In this embodiment, the system applies a target tracking algorithm in this step, aiming to maintain consistent tracking of a specific road sign across frames, and to maintain stability even when the viewing angle changes or the lighting conditions change.

[0108] For each pair of adjacent frames, the system compares the attributes of the road signs contained in them, such as position, size, direction of movement, and color similarity, to determine whether they represent the same physical road sign.

[0109] Based on the comparison results above, the system assigns a unique tracking ID to each identified road sign, ensuring that the road signs maintain consistent identity throughout the tracking process, even within the video stream. This method effectively addresses the perspective change problem caused by vehicle movement, ensuring correct matching of road signs in consecutive frames.

[0110] S133. For each tracked road sign, create an observation sequence containing the cropped image, timestamp, and location information to obtain the tracking sequence.

[0111] In this embodiment, the final step is to integrate all relevant information to form a structured tracking sequence, which facilitates subsequent detailed analysis.

[0112] The area corresponding to the road sign is cropped from the original image, unnecessary background interference is removed, and the focus is on the road sign itself.

[0113] In addition to the cropped image, it includes the timestamp of the time of capture (accurate to the second or even finer granularity) and geolocation information (if available). This additional information helps establish the temporal and spatial distribution of road signs.

[0114] All the above information is combined into an observation sequence, with each sequence corresponding to an independent road sign object, containing all observation records for that road sign throughout the entire video stream. Such sequences not only improve the efficiency of data analysis but also provide a solid data foundation for subsequent color stability analysis and fading assessment.

[0115] In summary, the process of "S130, identifying and tracking each road sign in the preprocessed result, and constructing a time series of the road sign's continuous appearance in different frames to obtain a tracking sequence" successfully extracts key information about road signs from vehicle-mounted videos through deep learning object detection and tracking mechanisms, and organizes it into a form that is easy to manage and analyze, greatly improving the automation level and accuracy of road sign status monitoring.

[0116] S140. Based on the tracking sequence, each road sign is divided into multiple functional areas, the color features of each functional area are calculated, and compared with standard samples to mark suspected fading areas to obtain suspected fading marks.

[0117] In this embodiment, the suspected fading mark refers to identifying and marking road sign areas that may have significant color deviations by comparing the calculated color difference value of the functional area with a preset threshold.

[0118] In one embodiment, step S140 described above may include steps S141 to S144.

[0119] S141. Based on the design specifications for traffic signs, the road sign images in the tracking sequence are divided into background area, text area, and symbol area to obtain multiple functional areas.

[0120] In this embodiment, the functional area refers to dividing the road sign image into different parts according to its function, such as background, text, and symbols, based on traffic sign design specifications. This step can be achieved through template matching (i.e., geometrically registering the detected road sign with a standard template and then partitioning it) or deep learning methods (such as using a semantic segmentation network). The aim is to ensure that the analysis focuses on the color representation of different functional areas, without being affected by font or deformation.

[0121] Specifically, based on the design specifications of traffic signs, template matching or deep learning methods are used to segment the road sign images in the tracking sequence into background areas, text areas, and symbol areas to obtain multiple functional areas.

[0122] Please see Figure 4 and Figure 5 After establishing the tracking sequence, the system performs regionalized color feature extraction and standard comparison on the keyframe images of each road sign to evaluate its color retention and potential fading. This step first requires dividing the road sign image into multiple functional regions, such as background, text, and symbol areas, according to traffic sign design specifications and visual structure. The division can be achieved using template matching, which compares the detected road sign with a standard template through geometric registration; or by using a segmentation network in deep learning to automatically identify and segment semantic regions. This approach ensures that the analysis process focuses on the color representation of different functional regions, avoiding interference from font styles or image distortions.

[0123] S142. Calculate the average color value, standard deviation, and dominant color distribution of each functional area to obtain color statistical characteristics.

[0124] In this embodiment, color statistical features refer to the calculation of the average color value (reflecting the basic hue of the color), standard deviation (showing the consistency or range of color variation) and dominant color distribution (determining the main colors present in the area and their proportions) for each functional area. These features are converted to the Lab color space to enhance the perceived consistency of color difference analysis.

[0125] For each defined functional region, the system further extracts its color statistical features, including but not limited to calculating the average color value, standard deviation of color distribution, and dominant hue distribution within that region. These features are obtained from the RGB color space and further converted to the Lab color space to improve the perceptual consistency of color difference analysis. The Lab color space is closer to the perceptual characteristics of the human visual system, making the quantification of color difference more accurate and reliable, providing a solid foundation for subsequent assessment of fading levels.

[0126] S143. Calculate the color difference value based on the corresponding standard samples in the database for the color statistical features.

[0127] In this embodiment, the color difference value refers to the ΔE value between functional areas calculated by comparing it with standard samples stored in the database. The formula is as follows: Where L, a, and b represent the brightness, red-green difference, and yellow-blue difference components, respectively. This step aims to quantify the color deviation between the actual road sign and the ideal state.

[0128] The system retrieves color samples of standard road signs of the corresponding category from the database as a reference benchmark and compares them with the color features of each functional area of ​​the current road sign. This method not only assesses the overall color retention of the road sign but also pinpoints which areas show significant color deviations, thus determining whether the road sign is in a suspected fading state. This method, based on regional color feature extraction and standard comparison, greatly improves the efficiency and accuracy of road sign maintenance, helping to promptly identify and repair information recognition difficulties that may result from fading.

[0129] S144. Based on the color difference value and the preset color difference value threshold, identify and mark the suspected fading areas with significant color deviation to obtain suspected fading marks.

[0130] When the color difference value of a certain functional area exceeds a preset threshold (e.g., ΔE>5 is considered a significant color shift), that area is marked as potentially fading. The system records these marks to provide a basis for subsequent judgment on whether the fading is persistent. In addition, to improve accuracy, the system also considers the consistency across multiple frames to ensure the stability and reliability of the markings.

[0131] This process enables precise assessment of the color retention of various functional areas of the road sign, improving not only the accuracy and objectivity of the inspection but also providing solid data support for further detailed verification. In this way, the degree of fading of the road sign can be effectively identified, allowing for appropriate maintenance measures to be taken.

[0132] S150. Summarize the suspected faded markers in all frames to determine whether the road sign is fading continuously.

[0133] In one embodiment, step S150 described above may include steps S151 to S154.

[0134] S151. Collect the color difference value and quality score of each functional area in each frame, and record the timestamp and suspected fading markers.

[0135] In this embodiment, at this stage, the system records in detail the color difference value (ΔE) of each functional area (such as the background area, text area, and symbol area) in each frame of video, as well as the quality score of that frame. This information includes, but is not limited to: the color difference value (ΔE) of each functional area; the quality score of the frame (e.g., sharpness, exposure uniformity, etc.); timestamp; and suspected fading markers (Boolean values ​​or rating values).

[0136] This process ensures that subsequent analyses are based on a comprehensive dataset, providing a foundation for accurate evaluation.

[0137] S152. Calculate the proportion of suspected faded frames in each functional area, adjust the weights according to the quality score, and calculate the weighted proportion to obtain the proportion of suspected faded frames.

[0138] In this embodiment, the percentage of suspected faded frames refers to the ratio of the number of frames marked as suspected faded in each functional area to the total number of frames in that area, after being adjusted by a quality score weighting.

[0139] In one embodiment, step S152 described above may include steps S1521 to S1523.

[0140] S1521. Calculate the sum of suspected faded markers for each functional region in the tracking sequence, and divide the sum by the total number of frames for each functional region in the tracking sequence to obtain the proportion of suspected faded frames for each functional region.

[0141] By counting the number of times each functional region was marked as suspected of fading throughout the entire tracking sequence and comparing it with the total number of times that functional region appeared in the entire sequence, a preliminary ratio indicator was obtained.

[0142] S1522. Adjust the weight of the corresponding frame according to the quality score.

[0143] By combining the weights calculated in the previous step, a weighted average is applied to the proportion of suspected faded frames, thereby more accurately reflecting the true fading situation of each functional area.

[0144] S1523. The weighted summation is performed based on the weight and the proportion of suspected faded frames in each functional area to obtain the proportion of suspected faded frames.

[0145] By combining the weights calculated in the previous step, a weighted average is applied to the proportion of suspected faded frames, thereby more accurately reflecting the true fading situation of each functional area.

[0146] S153. Based on the proportion of suspected faded frames, a threshold is set to determine whether the corresponding functional area is continuously fading.

[0147] In this embodiment, when the proportion of suspected faded frames exceeds a set threshold, and the fluctuation of the suspected faded marker in the tracking sequence meets the requirements, it is determined that the corresponding functional area has continuous fading.

[0148] In this crucial step, the system uses a pre-set threshold to determine whether a functional area has experienced persistent fading. If the proportion of suspected faded frames in a functional area exceeds this threshold, and the distribution of these suspected faded markers in the tracking sequence is relatively stable (i.e., without obvious fluctuations), then it can be determined that there is persistent fading in that area.

[0149] S154. When the corresponding functional area is continuously fading, mark the functional area corresponding to the road sign as continuously fading.

[0150] Finally, once a persistent fading is confirmed in a functional area, the system will mark that area as "persistently faded." This marking not only helps with subsequent detailed verification work but also provides the maintenance department with a clear target, facilitating them to take timely and effective remedial measures.

[0151] In this embodiment, for all tracking frames of the same road sign, the system summarizes the region information (i.e., ΔE values) of all functional areas (e.g., background area, text area, symbol area, etc.) in each frame to generate a marker sequence for each functional area. Each marker contains the following information:

[0152] Timestamp and frame number: Used to identify the position of the frame in the tracking sequence.

[0153] Suspected faded area marker: Indicates whether the area is marked as suspected faded in the form of a Boolean value or a rating value.

[0154] Frame quality score: a comprehensive evaluation including factors such as sharpness and exposure, used for subsequent weighted calculations.

[0155] Based on the markers in each frame, the system determines the persistence of each area of ​​the road sign throughout the entire tracking sequence. The key is to distinguish between environmental interference (such as changes in lighting, changes in viewing angle, etc.) and actual fading.

[0156] Cumulative suspected faded frames: The number of frames in the entire tracking sequence that each region is marked as "suspected faded".

[0157] Calculate the proportion: Assuming that the total number of frames in a certain region in the tracking sequence is N, and the number of frames marked as suspected fading is M, then the proportion of suspected fading in that region is P = M / N.

[0158] If the proportion of suspected faded frames in a certain area exceeds a set threshold (e.g., P ≥ 50% or 60%), and the fluctuations between these marked frames are small, then the area is considered to have persistent fading, meeting the requirements for further fine-grained verification.

[0159] If the proportion of suspected faded frames in the region is lower than the set threshold (e.g., P<20%), or if suspected fading only appears in a very small number of frames, the system marks the region as "intermittent deviation" or "environmental interference" and does not proceed to fine verification.

[0160] To reduce interference from low-quality frames, the system performs quality weighting on each frame: evaluating the sharpness, exposure, and occlusion of each frame and assigning a weight to each frame (e.g., a high-sharpness frame has a weight of 1, and a low-sharpness frame has a weight of 0.5). When calculating the proportion of suspected faded frames, the weight of each frame is considered to calculate the weighted proportion, thus more accurately reflecting the true fading situation in each area.

[0161] If the proportion of suspected faded frames in a certain area meets the condition of continuous fading and there are no obvious abnormalities (such as extremely low quality frames or discontinuous suspected markers), the system determines that the area is "continuously faded" and submits the result to the fine verification module (such as reflectivity, readability and other verifications).

[0162] If the proportion of suspected faded frames in a certain area is below a threshold, or if there are obvious interferences such as lighting fluctuations or occlusions, the system determines that the area is "non-persistent fading" or "environmental interference" and records it as an item to be reviewed or observed. This ensures that only areas that truly require attention enter the more in-depth analysis stage, improving overall efficiency and accuracy.

[0163] This demonstrates that the system can effectively and automatically identify the persistent fading problem of road signs from a large amount of vehicle video data, greatly improving detection efficiency and accuracy and reducing the need for manual intervention.

[0164] S160. When the road sign is continuously fading, the reflectivity, readability and texture sharpness of the suspected fading area are verified, and the fading level is determined based on the verification results.

[0165] In this embodiment, the fading level refers to a comprehensive assessment based on reflectivity, text readability, and texture condition, classifying the degree of fading of road signs into three levels: slight, moderate, and severe.

[0166] like Figure 6 As shown, to ensure that the detected suspected fading areas do indeed exist and to assess their impact, the system further performs verifications of reflectivity, readability, and texture sharpness. These tests aim to provide a comprehensive evaluation framework to accurately classify and quantify the fading condition of each road sign.

[0167] In one embodiment, step S160 described above may include steps S161 to S164.

[0168] S161. Detect the intensity of reflected light on the suspected fading area at night or in a strong light environment, calculate the reflection recovery ratio, classify the fading according to the calculation results, and obtain the reflection performance verification results.

[0169] In this embodiment, the reflectivity verification result refers to measuring the intensity of reflected light in the suspected faded area at night or in a strong light environment, and calculating its reflectivity recovery ratio (IRR) compared with the standard sample, thereby determining the retention status of reflectivity.

[0170] This verification is only conducted at night or in bright light conditions, because only under these conditions can the reflectivity of road signs be effectively evaluated.

[0171] The intensity of reflected light in suspected fading areas at night or under strong light was measured using specific equipment.

[0172] Calculate the Infrared Retrieval Ratio (IRR), which is the ratio of the actual reflected light intensity to the reflected light intensity of a standard sample. Based on the IRR value, reflectivity is classified into three levels:

[0173] Grade I: Normal reflectivity, with no significant degradation.

[0174] Grade II: Slight degradation in reflectivity, suitable for daytime use.

[0175] Grade III: Reflectivity is severely degraded, resulting in poor visibility at night or in low light conditions, requiring urgent repair.

[0176] S162. Identify the text content in the suspected faded area, calculate the recognition rate, and classify the readability of the text content in the suspected faded area based on the recognition rate to obtain the text readability verification result.

[0177] In this embodiment, the text readability verification result refers to identifying the text content in the suspected faded area and evaluating the clarity of the text and the effectiveness of information transmission based on the recognition rate.

[0178] The text content in suspected faded areas is automatically identified using optical character recognition technology.

[0179] The text recognition rate is calculated based on the OCR results. A higher recognition rate indicates that the text is clearer and more readable.

[0180] Classify the readability of text content based on recognition rate:

[0181] Level I: The text is clear and has a high recognition rate.

[0182] Level II: Text is slightly blurry, recognition rate is moderate.

[0183] Level III: The text is severely blurry, with low recognition rate, affecting the transmission of traffic information.

[0184] S163. Analyze the aging status of the surface coating in the suspected fading area, evaluate the texture sharpness, and classify the suspected fading area according to the texture clarity to obtain the texture status verification result.

[0185] In this embodiment, the texture state verification result refers to the assessment of the degree of degradation of the road sign surface coating by analyzing the aging condition and texture sharpness of the surface coating.

[0186] The aging status of the surface coating in suspected fading areas was analyzed using visual inspection and image processing techniques.

[0187] Assess the clarity of the texture to determine if there is significant degradation. Based on texture clarity, fading can be categorized as follows:

[0188] Grade I: Clear texture with no obvious degradation.

[0189] Level II: The texture is slightly blurry and there is slight degradation.

[0190] Level III: Texture is severely blurred or lost, and the surface coating has obviously degraded.

[0191] S164. Combining the reflectivity verification results, text readability verification results, and texture status verification results, a comprehensive evaluation is conducted on each suspected fading area to determine the fading level.

[0192] Based on the above verification results of reflectivity, text readability, and texture status, a comprehensive evaluation was conducted on each suspected fading area.

[0193] The final determination of the fading level of each road sign is usually divided into three categories:

[0194] Grade I (Slight Fading): Slight color deviation, normal reflectivity and readability, and clear texture.

[0195] Grade II (Moderate Fading): Color, reflectivity, and readability are all reduced, and surface texture is slightly degraded.

[0196] Level III (Severe Fading): Color, reflectivity, and readability are significantly reduced, and surface texture is severely degraded.

[0197] Through such a multi-level and multi-dimensional verification and evaluation mechanism, the actual condition of road signs can be reflected more accurately, providing a scientific basis for subsequent maintenance and updating work.

[0198] This embodiment utilizes a vehicle-mounted camera to capture road sign images in real time during driving, replacing the traditional manual inspection mode with an automated approach, significantly improving data collection efficiency. Advanced image processing algorithms are employed to reduce the impact of lighting changes and shadows on color assessment through fusion analysis of multiple consecutive frames, ensuring the objectivity and reliability of color degradation detection results. Traffic signs are divided into different functional areas, and their color characteristics are compared with a pre-set standard sample library to achieve precise quantitative assessment of the degree of fading. Combining multi-dimensional indicators such as color change, reflectivity, and text readability, fading assessment reports of slight, moderate, and severe levels are automatically generated, providing a scientific basis for maintenance decisions. This approach improves detection efficiency and coverage, meeting the needs of large-scale road facility monitoring; solves the problem of inconsistent manual inspection results through objective quantitative methods; introduces advanced algorithms to improve the robustness and accuracy of the detection system, especially under complex environmental conditions; and constructs a comprehensive evaluation model covering reflectivity decay, color difference, and recognition rate, enhancing the comprehensiveness of road sign quality detection.

[0199] The method in this embodiment promotes a more scientific and efficient approach to road traffic facility maintenance through the application of intelligent technologies. Specifically, automation significantly reduces labor costs while ensuring the efficiency and accuracy of inspection work. Using a standardized evaluation system and multi-dimensional indicator fusion technology effectively improves the consistency and repeatability of inspection results; it not only focuses on color changes but also considers factors such as reflectivity and text clarity, comprehensively evaluating the actual condition of signs from multiple perspectives and enhancing the practical value of the inspection plan.

[0200] The aforementioned dynamic patrol-based road sign fading detection method acquires and preprocesses video data of road signs on both sides of the road, combines time and location information, and utilizes advanced image processing technology to identify and track each road sign, constructing its time-series performance across different frames. Subsequently, each road sign is divided into multiple functional areas, color features are calculated and compared with standard samples to mark suspected fading areas. The marks from all frames are then aggregated to determine whether the road sign is continuously fading, and the suspected areas are comprehensively verified for reflectivity, readability, and texture sharpness to ultimately determine the fading level. This method achieves accurate and stable identification and evaluation of the degree of road sign fading, effectively solving the problems of low efficiency, strong subjectivity, insufficient stability, and lack of multi-dimensional comprehensive evaluation mechanisms in existing technologies, providing a scientific basis and technical support for the maintenance of road traffic facilities.

[0201] Figure 8 This is a schematic block diagram of a road sign fading detection system 300 based on dynamic patrol provided in an embodiment of the present invention. Figure 8As shown, corresponding to the above-described road sign fading detection method based on dynamic patrol, the present invention also provides a road sign fading detection system 300 based on dynamic patrol. This road sign fading detection system 300 based on dynamic patrol includes a unit for executing the above-described road sign fading detection method based on dynamic patrol, and the system can be configured in a server. Specifically, please refer to... Figure 8 The road sign fading detection system 300 based on dynamic inspection includes an acquisition unit 301, a preprocessing unit 302, a tracking sequence construction unit 303, a marking unit 304, a persistence judgment unit 305, and a fine verification unit 306.

[0202] The system comprises: an acquisition unit 301 for acquiring videos of road signs on both sides of the road and recording time and location information; a preprocessing unit 302 for preprocessing the road sign videos to obtain preprocessing results; a tracking sequence construction unit 303 for identifying and tracking each road sign in the preprocessing results, constructing a time series of the road sign's continuous appearance in different frames to obtain a tracking sequence; a marking unit 304 for dividing each road sign into multiple functional areas based on the tracking sequence, calculating the color features of each functional area, comparing them with standard samples, and marking suspected fading areas to obtain suspected fading marks; a persistence judgment unit 305 for summarizing suspected fading marks in all frames to determine whether the road sign is persistently fading; and a fine verification unit 306 for verifying the reflectivity, readability, and texture sharpness of the suspected fading areas when the road sign is persistently fading, and determining the fading level based on the verification results.

[0203] In one embodiment, the persistence determination unit 305 includes:

[0204] The collection subunit is used to collect the color difference value and quality score of each functional area in each frame, and record the timestamp and suspected fading mark; the proportion calculation subunit is used to calculate the proportion of suspected fading frames in each functional area, and adjust the weight according to the quality score to calculate the weighted proportion to obtain the proportion of suspected fading frames; the judgment subunit is used to apply a set threshold based on the proportion of suspected fading frames to determine whether the corresponding functional area has continuous fading; the marking subunit is used to mark the functional area corresponding to the road sign as continuously fading when the corresponding functional area has continuous fading.

[0205] In one embodiment, the percentage calculation subunit includes:

[0206] The proportion calculation module is used to calculate the sum of suspected faded markers for each functional region in the tracking sequence, and divide the sum by the total number of frames for each functional region in the tracking sequence to obtain the proportion of suspected faded frames for each functional region; the adjustment module is used to adjust the weight of the corresponding frame according to the quality score; the accumulation module is used to perform weighted accumulation based on the weight and the proportion of suspected faded frames for each functional region to obtain the percentage of suspected faded frames.

[0207] In one embodiment, the judgment subunit is used to determine that the corresponding functional area has continuous fading when the proportion of suspected faded frames exceeds a set threshold and the fluctuation of the suspected faded marker in the tracking sequence meets the requirements.

[0208] In one embodiment, the marking unit 304 includes:

[0209] The segmentation subunit is used to segment the road sign image in the tracking sequence into a background area, a text area, and a symbol area based on the design specifications of traffic signs, so as to obtain multiple functional areas; the feature determination subunit is used to calculate the average color value, standard deviation, and dominant color distribution of each functional area to obtain color statistical features; the color difference value calculation subunit is used to calculate the color difference value based on the corresponding standard samples in the database; the identification subunit is used to identify and mark suspected fading areas with significant color deviation according to the color difference value and the preset color difference value threshold, so as to obtain suspected fading marks.

[0210] In one embodiment, the segmentation subunit is used to segment the road sign image in the tracking sequence into a background area, a text area, and a symbol area based on the design specifications of traffic signs, using template matching or deep learning methods, to obtain multiple functional areas.

[0211] In one embodiment, the fine verification unit 306 includes:

[0212] The reflectivity verification subunit is used to detect the intensity of reflected light in the suspected fading area at night or in a strong light environment, calculate the reflectivity recovery ratio, and classify the fading based on the calculation results to obtain the reflectivity performance verification result. The text verification subunit is used to identify the text content in the suspected fading area, calculate the recognition rate, and classify the readability of the text content in the suspected fading area based on the recognition rate to obtain the text readability verification result. The texture verification subunit is used to analyze the aging status of the surface coating in the suspected fading area, evaluate the texture sharpness, and classify the suspected fading area based on the texture clarity to obtain the texture status verification result. The comprehensive evaluation subunit is used to combine the reflectivity performance verification result, the text readability verification result, and the texture status verification result to comprehensively evaluate each suspected fading area and determine the fading level.

[0213] In one embodiment, the preprocessing unit 302 includes:

[0214] The extraction subunit is used to extract image sequences frame by frame from the road sign video and number and synchronize them according to time; the removal subunit is used to automatically detect and remove blurry, jittery or poorly exposed frames to obtain removal results; the adjustment subunit is used to adjust the removal results through distortion correction and geometric transformation to obtain preprocessed results, wherein the preprocessed results include processed image frames and corresponding time and position information.

[0215] In one embodiment, the tracking sequence construction unit 303 includes:

[0216] The localization subunit is used to analyze and locate the road sign features in each frame of the preprocessed image using a deep learning object detection algorithm to obtain the road sign; the comparison subunit is used to assign a tracking ID to each road sign by comparing the position, size, movement direction, and color similarity of road signs between adjacent frames using an object tracking mechanism; the creation subunit is used to create an observation sequence containing the cropped image, timestamp, and location information for each tracked road sign to obtain the tracking sequence.

[0217] It should be noted that those skilled in the art can clearly understand that the specific implementation process of the above-mentioned road sign fading detection system 300 based on dynamic inspection and each unit can be referred to the corresponding description in the foregoing method embodiments. For the sake of convenience and brevity, it will not be repeated here.

[0218] The aforementioned road sign fading detection system 300 based on dynamic inspection can be implemented as a computer program, which can, for example... Figure 8 It runs on the computer device shown.

[0219] Please see Figure 8 , Figure 8 This is a schematic block diagram of a computer device provided in an embodiment of this application. The computer device 500 can be a server, wherein the server can be a standalone server or a server cluster composed of multiple servers.

[0220] See Figure 8 The computer device 500 includes a processor 502, a memory, and a network interface 505 connected via a system bus 501. The memory may include a non-volatile storage medium 503 and internal memory 504.

[0221] The non-volatile storage medium 503 may store an operating system 5031 and a computer program 5032. The computer program 5032 includes program instructions that, when executed, cause the processor 502 to perform a road sign fading detection method based on dynamic inspection.

[0222] The processor 502 provides computing and control capabilities to support the operation of the entire computer device 500.

[0223] The internal memory 504 provides an environment for the operation of the computer program 5032 in the non-volatile storage medium 503. When the computer program 5032 is executed by the processor 502, the processor 502 can execute a road sign fading detection method based on dynamic inspection.

[0224] This network interface 505 is used for network communication with other devices. Those skilled in the art will understand that... Figure 8 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device 500 to which the present application is applied. The specific computer device 500 may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0225] The processor 502 is used to run the computer program 5032 stored in the memory to implement all the steps of the road sign fading detection method based on dynamic inspection.

[0226] It should be understood that in the embodiments of this application, the processor 502 may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0227] It will be understood by those skilled in the art that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program includes program instructions and can be stored in a storage medium, which is a computer-readable storage medium. The program instructions are executed by at least one processor in the computer system to implement the process steps of the embodiments of the above methods.

[0228] Therefore, the present invention also provides a storage medium. This storage medium can be a computer-readable storage medium. The storage medium stores a computer program, wherein when executed by a processor, the computer program causes the processor to perform all the steps of the dynamic patrol-based road sign fading detection method.

[0229] The storage medium can be any computer-readable storage medium capable of storing program code, such as a USB flash drive, portable hard drive, read-only memory (ROM), magnetic disk, or optical disk.

[0230] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0231] In the embodiments provided by this invention, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the division of each unit is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.

[0232] The steps in the method of this invention can be adjusted, merged, or reduced in order according to actual needs. The units in the system of this invention can be merged, divided, or reduced according to actual needs. Furthermore, the functional units in the various embodiments of this 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.

[0233] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or 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, a terminal, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention.

[0234] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for detecting road sign fading based on dynamic inspection, characterized in that, include: Acquire videos of road signs on both sides of the road and record the time and location information; The road sign video is preprocessed to obtain the preprocessing result; Identify and track each road sign in the preprocessing results, and construct a time series of the road signs' continuous appearance in different frames to obtain a tracking sequence; Based on the tracking sequence, each road sign is divided into multiple functional areas, the color features of each functional area are calculated, and compared with standard samples to mark suspected fading areas to obtain suspected fading marks; Summarize all suspected faded markers in all frames to determine if the road sign is fading continuously; When the road sign is continuously fading, the reflectivity, readability, and texture sharpness of the suspected fading area are verified, and the fading level is determined based on the verification results. The process of summarizing suspected faded markers from all frames to determine whether the road sign is continuously fading includes: Collect color difference values ​​and quality scores for each functional area in each frame, and record timestamps and suspected fading markers; Calculate the proportion of suspected faded frames in each functional area, adjust the weights according to the quality score, and calculate the weighted proportion to obtain the proportion of suspected faded frames. Based on the proportion of suspected faded frames, a threshold is set to determine whether the corresponding functional area is continuously fading; When the corresponding functional area shows continuous fading, the functional area corresponding to the road sign is marked as continuously fading.

2. The method for detecting road sign fading based on dynamic inspection according to claim 1, characterized in that, The calculation of the proportion of suspected faded frames in each functional area, and the adjustment of weights based on the quality score to calculate the weighted proportion to obtain the proportion of suspected faded frames, includes: Calculate the sum of suspected faded markers for each functional region in the tracking sequence, and divide the sum by the total number of frames for each functional region in the tracking sequence to obtain the proportion of suspected faded frames for each functional region; The weights of the corresponding frames are adjusted based on the quality score. The weights are weighted and accumulated based on the proportion of suspected faded frames in each functional area to obtain the proportion of suspected faded frames.

3. The method for detecting road sign fading based on dynamic inspection according to claim 2, characterized in that, The step of determining whether there is continuous fading in the corresponding functional area based on the percentage of suspected faded frames by setting a threshold includes: If the proportion of suspected faded frames exceeds a set threshold, and the fluctuation of the suspected faded marker in the tracking sequence meets the requirements, then it is determined that the corresponding functional area has continuous fading.

4. The method for detecting road sign fading based on dynamic inspection according to claim 1, characterized in that, Based on the tracking sequence, each road sign is divided into multiple functional areas, the color features of each functional area are calculated, and compared with standard samples to mark suspected fading areas, thereby obtaining suspected fading markers, including: Based on the design specifications for traffic signs, the road sign images in the tracking sequence are segmented into background areas, text areas, and symbol areas to obtain multiple functional areas; Calculate the average color value, standard deviation, and dominant color distribution for each functional area to obtain color statistical characteristics; Based on the corresponding standard samples in the database, the color difference value is calculated for the color statistical features; Based on the color difference value and the preset color difference value threshold, suspected fading areas with significant color deviation are identified and marked to obtain suspected fading marks.

5. The method for detecting road sign fading based on dynamic inspection according to claim 4, characterized in that, The traffic sign-based design specifications segment the road sign images in the tracking sequence into background, text, and symbol areas to obtain multiple functional areas, including: Based on traffic sign design specifications, template matching or deep learning methods are used to segment the road sign images in the tracking sequence into background areas, text areas, and symbol areas to obtain multiple functional areas.

6. The method for detecting road sign fading based on dynamic inspection according to claim 1, characterized in that, When the road sign is continuously fading, the reflectivity, readability, and texture sharpness of the suspected fading areas are verified, and the fading level is determined based on the verification results, including: The intensity of reflected light in the suspected faded area is detected at night or in a strong light environment, the reflectivity recovery ratio is calculated, and the fading is classified according to the calculation results to obtain the reflectivity performance verification results. The text content in the suspected faded area is identified, the recognition rate is calculated, and the readability of the text content in the suspected faded area is classified based on the recognition rate to obtain the text readability verification result. The aging status of the surface coating in the suspected fading area is analyzed, the texture sharpness is evaluated, and the suspected fading area is classified according to the texture clarity to obtain the texture status verification result. Based on the results of the reflectivity verification, text readability verification, and texture status verification, each suspected fading area is comprehensively evaluated to determine the fading level.

7. The method for detecting road sign fading based on dynamic inspection according to claim 1, characterized in that, The preprocessing of the road sign video to obtain the preprocessing result includes: The image sequence is extracted frame by frame from the road sign video and numbered and synchronized according to time. Automatically detect and remove blurry, jittery, or poorly exposed frames to obtain the removal results; The removal result is adjusted by distortion correction and geometric transformation to obtain a preprocessed result, wherein the preprocessed result includes the processed image frame and the corresponding time and position information.

8. The method for detecting road sign fading based on dynamic inspection according to claim 1, characterized in that, The process of identifying and tracking each road sign in the preprocessing results, and constructing a time series of the road signs' continuous appearance in different frames to obtain a tracking sequence, includes: The road sign features in each frame of the preprocessed image are analyzed and located using a deep learning object detection algorithm to obtain the road sign; Using a target tracking mechanism, a tracking ID is assigned to each road sign by comparing the position, size, direction of movement, and color similarity between adjacent frames; For each road sign being tracked, an observation sequence is created containing the cropped image, timestamp, and location information to obtain the tracking sequence.

9. A road sign fading detection system based on dynamic inspection, characterized in that, include: The acquisition unit is used to acquire video of road signs on both sides of the road and record the time and location information; A preprocessing unit is used to preprocess the road sign video to obtain a preprocessing result; The tracking sequence construction unit is used to identify and track each road sign in the preprocessing result, and construct a time series of the road signs continuously appearing in different frames to obtain the tracking sequence; A marking unit is used to divide each road sign into multiple functional areas based on the tracking sequence, calculate the color features of each functional area, compare them with standard samples, and mark suspected fading areas to obtain suspected fading marks. The persistence determination unit is used to summarize the suspected fading markers in all frames and determine whether the road sign is persistently fading. This includes: collecting the color difference value and quality score of each functional area in each frame, and recording the timestamp and suspected fading marker; calculating the proportion of suspected fading frames in each functional area, adjusting the weights according to the quality score, and calculating the weighted proportion to obtain the proportion of suspected fading frames; applying a set threshold based on the proportion of suspected fading frames to determine whether the corresponding functional area is persistently fading; and marking the functional area corresponding to the road sign as persistently fading when the corresponding functional area is persistently fading. The fine verification unit is used to verify the reflectivity, readability, and texture sharpness of the suspected faded area when the road sign is continuously fading, and to determine the fading level based on the verification results.

Images