Calibration-free monitoring distance estimation method using lane line scale and vehicle tracking

Abstract

This invention provides a calibration-free method for estimating monitoring distance using lane line markers and vehicle tracking, relating to the fields of computer vision and intelligent traffic monitoring technology. The method includes: based on monitoring video, using directional bounding boxes to detect lane lines of known true length and obtaining their pixel lengths; within the image area corresponding to the lane line, defining virtual detection lines, and using the YOLO algorithm and a tracker to detect and track moving vehicles; calculating the vehicle's true speed based on the pixel displacement and time difference of its base midpoint crossing the virtual line, combined with the proportional relationship of the lane line markers; finally, automatically estimating the effective coverage distance of the monitoring camera based on this speed and the total duration of the vehicle's presence in the image from appearance to disappearance, and providing a visual output. This invention completely eliminates the need for pre-calibrating camera parameters, achieving efficient, automated, and low-cost assessment of the effective coverage distance of large-scale deployed monitoring systems.

Description

Technical Field

[0001] This invention relates to the fields of computer vision and intelligent traffic monitoring technology, and in particular to a method for estimating uncalibrated monitoring distances using lane line rulers and vehicle tracking. Background Technology

[0002] In the field of traffic monitoring, accurately determining the effective coverage distance of surveillance cameras is crucial for evaluating system performance, optimizing camera placement, and ensuring comprehensive monitoring without blind spots. Especially in critical traffic scenarios such as highways, tunnels, and bridges, understanding the furthest range at which cameras can clearly capture and continuously track vehicles directly impacts the reliability of event detection, violation evidence collection, traffic flow analysis, and emergency command.

[0003] Currently, the mainstream method for obtaining the effective coverage distance of surveillance cameras heavily relies on the precise calibration of the camera's parameters. These parameters include the camera's internal parameters (such as focal length, sensor size, and distortion coefficient) and external parameters (such as installation height, pitch angle, and yaw angle). Theoretically, once these parameters are known, the real-world distance corresponding to any position in the field of view can be calculated using a geometric projection model. However, this method faces severe challenges in the actual deployment and operation of large-scale traffic monitoring networks. First, the precise acquisition of camera parameters is itself a professional and tedious process, requiring the use of specialized calibration equipment or specific calibration references for careful on-site measurement and calculation. For a city-level or road network-level monitoring system with hundreds or thousands of cameras, calibrating each camera individually would require enormous manpower, resources, and time, making it almost impossible. Second, many existing monitoring systems lack complete installation parameter records, especially for early-installed equipment, whose specific internal and external parameters are often unverifiable. Furthermore, camera parameters are not static and may undergo slight changes due to equipment aging, lens loosening, reinstallation after repair, or human touch. This can cause the calculation model based on the initial calibration values ​​to gradually deviate from the actual situation, rendering the calculation results unreliable.

[0004] To circumvent complex parameter calibration, some alternatives attempt to use static references in the scene for distance estimation, such as traffic signs or stationary objects of known size. However, these methods often require manual specification of references in the image, making automation difficult, and the actual size of the references may not be uniform or precisely known, limiting their large-scale application. Other deep learning-based methods attempt to regress distance information directly from images, but these models require a large amount of precisely labeled data for training, and their generalization ability is limited by the distribution of the training data. Stability and accuracy are difficult to guarantee when facing different camera models, installation locations, and scenes.

[0005] Therefore, in the daily management and effectiveness evaluation of traffic monitoring, management departments often face a dilemma: on the one hand, they urgently need to understand the actual monitoring capabilities of each camera; on the other hand, they are limited by the high cost, low efficiency, or unreliability of existing technologies, often relying solely on equipment design parameters or experience-based estimates, which frequently deviate significantly from reality. This makes it impossible to scientifically assess monitoring blind spots, optimize camera deployment, or accurately determine whether the system meets control requirements. Developing an effective coverage distance estimation method that is independent of camera calibration, highly automated, and applicable to large-scale existing monitoring systems has become a pressing practical need in this field. Summary of the Invention

[0006] To address the technical problems of existing technologies that rely on precise camera calibration, resulting in high implementation costs, difficulty in applying to large-scale deployed systems, inability to dynamically adapt to parameter changes, low automation of existing uncalibrated methods, reliance on specific reference objects or large amounts of labeled data, and insufficient generalization ability, thus failing to efficiently, accurately, and automatically obtain the true effective coverage distance of traffic monitoring cameras, this invention provides an uncalibrated monitoring distance estimation method using lane line rulers and vehicle tracking.

[0007] The technical solution provided by this invention is as follows: The present invention provides a calibration-free monitoring distance estimation method using lane line markers and vehicle tracking, comprising: S1. Calibration information acquisition steps: Based on the video image output by the monitoring camera, detect at least one lane line with a known real length and determine the pixel length of the lane line in the image; S2. Short-range vehicle speed calculation steps: Within the image area corresponding to the lane line of known true length, by detecting and tracking moving vehicles, and combining the true length of the lane line and the pixel length, calculate the true speed of the moving vehicle when it passes through the area. S3. Effective coverage distance estimation step: Obtain the total duration from the appearance to the disappearance of the moving vehicle in the monitoring video frame, and estimate the effective coverage distance of the monitoring camera based on the actual movement speed and the total duration.

[0008] Furthermore, in step S1, the lane line of known actual length is a longitudinal deceleration marking on the road, with a standard length of 6 meters.

[0009] Further, in step S1, the lane line is detected using a target detection method based on directional bounding boxes to obtain its precise pixel length and orientation angle in the image.

[0010] Furthermore, step S2 specifically includes the following sub-steps: S21. Draw the first virtual detection line and the second virtual detection line at the positions corresponding to the two ends of the lane line of known real length in the image; S22. Detect and track moving vehicles in the video sequence, determine the time points when the midpoint of the bottom edge passes through the first virtual detection line and the second virtual detection line in sequence, and calculate the pixel distance that the midpoint of the bottom edge moves between the two lines. S23. Based on the ratio between the actual length of the lane line and its pixel length, and the pixel distance of the movement of the midpoint of the bottom edge, calculate the actual distance the moving vehicle moves between the two virtual detection lines. S24. Calculate the actual speed of the moving vehicle based on the actual distance and the time difference between the two virtual detection lines.

[0011] Furthermore, in step S22, the detection and tracking of moving vehicles are achieved in collaboration with a YOLO series target detection algorithm and a tracker.

[0012] Further, in step S3, the total duration is obtained by recording the first global timestamp when the moving vehicle is first detected and the second global timestamp when it finally disappears from the screen, and calculating the difference between them.

[0013] Further, in step S3, the effective coverage distance is estimated using the following formula: in, For effective coverage distance, The actual velocity calculated in step S2, The total duration is given.

[0014] Furthermore, the method also includes step S4 and a result output step: The estimated effective coverage distance is associated with the corresponding video frame or camera identification information, and then visualized or stored.

[0015] Furthermore, the method is executed by a computer program deployed on an edge computing device or server for batch estimation of the effective coverage distance of one or more surveillance cameras in a traffic monitoring system that have not undergone internal parameter calibration.

[0016] Furthermore, the method is used to evaluate the actual monitoring range of surveillance cameras in tunnel, bridge, or highway scenarios, providing data support for traffic management departments to optimize surveillance deployment and evaluate effectiveness.

[0017] The beneficial effects of the technical solution provided by this invention include at least the following: (1) In this invention, by utilizing lane lines (such as longitudinal deceleration markings) that are common and standardized in traffic scenarios as a natural length benchmark, the dependence on internal and external parameters of the camera is completely eliminated. This method requires no on-site calibration operations, nor does it require querying or measuring difficult-to-obtain parameters such as the camera's focal length and installation height; it only requires analysis using the video stream output by the camera itself. This significantly reduces the implementation threshold and cost, making it possible to quickly assess the effective coverage distance of a large number of deployed surveillance cameras with unknown or potentially changing parameters, thus solving the core bottleneck of traditional methods that cannot be applied on a large scale due to strong parameter dependence.

[0018] (2) In this invention, by combining pixel measurement of standard-length lane lines with vehicle detection and tracking technology, dynamic conversion and distance estimation from image pixels to real-world distance are achieved. First, the pixel length of the lane lines is accurately obtained based on rotating target detection technology, establishing a stable proportional relationship under the current image viewpoint. Then, by tracking the movement of the vehicle within a known scale area, its true speed is directly calculated. Finally, using this speed and the vehicle's dwell time in the full field of view, the effective coverage distance of the camera is automatically estimated. The entire process is fully automated, requiring no manual intervention to specify reference objects, and by dynamically tracking the vehicle to calculate the speed, the estimation results can truly reflect the monitoring capabilities under the current scene, avoiding errors caused by using fixed models or empirical values.

[0019] (3) The proposed method in this invention has a clear logic, and each step can be efficiently executed by mature computer vision algorithms, making it easy to integrate into a software system. This system can be deployed on servers or edge computing devices to achieve parallel and batch processing of multiple surveillance video streams. This provides traffic management departments with an efficient tool that can periodically or irregularly scan and evaluate the effectiveness of the entire monitoring network and automatically generate coverage distance reports. Based on this accurate and objective data, managers can scientifically identify monitoring blind spots, optimize camera deployment strategies, and assess system upgrade needs, thereby greatly improving the management refinement and proactive maintenance capabilities of the traffic monitoring system, laying a reliable data foundation for the precise control of intelligent transportation. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1This is a flowchart illustrating the uncalibrated monitoring distance estimation method using lane line rulers and vehicle tracking, as provided in an embodiment of the present invention. Detailed Implementation

[0022] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0023] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0024] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0025] In embodiments of the present invention, sometimes the subscript is as follows: It may be mistakenly written as a non-subscript form such as W1. When the distinction is not emphasized, the meaning they express is the same.

[0026] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0027] Reference manual attached Figure 1 The diagram illustrates a flowchart of an uncalibrated monitoring distance estimation method using lane line rulers and vehicle tracking, provided by an embodiment of the present invention.

[0028] This invention provides a calibration-free monitoring distance estimation method using lane line markers and vehicle tracking. The processing flow may include the following steps: S1. Calibration information acquisition steps: Based on the video image output by the monitoring camera, detect at least one lane line with a known real length and determine the pixel length of the lane line in the image.

[0029] This step operates based on video image sequences directly output by the surveillance camera, without relying on any pre-measured internal or external camera parameters. Keyframes containing clear lane lines are selected from the video, and an object detection algorithm automatically identifies at least one lane line with a known physical length. The algorithm outputs the pixel range occupied by this lane line in the image coordinate system, and then calculates its complete pixel length. This pixel length, together with the known physical length, constitutes the calibration basis information required for subsequent calculations.

[0030] S2. Steps for calculating the speed of close-range vehicles: Within the image area corresponding to a lane line of known actual length, moving vehicles are detected and tracked. Combining the actual length of the lane line with the pixel length, the actual speed of the moving vehicle when passing through the area is calculated.

[0031] This step is performed within the image area defined by the lane lines of known actual lengths. First, the moving vehicle in the video is continuously detected and steadily tracked to obtain its trajectory. Next, using the ratio between actual length and pixel length established in step S1, the pixel displacement of the vehicle moving within this specific area is converted into real-world physical displacement. Simultaneously, the time taken for the vehicle to generate this physical displacement is calculated based on the video timestamp. Finally, by removing the physical bits and adding time, the actual speed of the moving vehicle in the vicinity of the current monitored scene is directly calculated.

[0032] S3. Effective coverage distance estimation steps: Obtain the total duration from when the moving vehicle appears to when it disappears in the monitoring video frame, and estimate the effective coverage distance of the monitoring camera based on the actual speed of movement and the total duration.

[0033] This step first records the complete time period from when the target vehicle first appears in the monitoring frame until it finally disappears from the frame through vehicle tracking. This time period is the total duration for which the vehicle is within the camera's effective field of view. Then, the vehicle's actual speed at close range, calculated in step S2, is multiplied by its total duration in the frame. The resulting product is estimated as the effective coverage distance of the current monitoring camera in that scene direction. This distance represents the farthest range at which the camera can clearly and continuously track moving targets.

[0034] In one possible implementation, in step S1, the lane line of known actual length is the longitudinal deceleration marking on the road, with a standard length of 6 meters.

[0035] In traffic monitoring scenarios, longitudinal deceleration markings painted on the road surface are selected as benchmarks with known actual lengths. These markings consist of a set of parallel rhomboid blocks, and their design specifications conform to national standards. The actual physical length of each segment of the marking is fixed and known. In a preferred embodiment of this invention, the marking used is a standardized longitudinal deceleration marking with a length of 6 meters. This fixed length serves as an absolute physical scale benchmark, used to establish the conversion relationship between the image pixel domain and the real physical domain.

[0036] In one possible implementation, in step S1, a target detection method based on directional bounding boxes is used to detect lane lines in order to obtain their precise pixel length and orientation angle in the image.

[0037] To accurately determine the pixel length of the sloping lane line in an image, a target detection method based on oriented bounding boxes is employed. This method abandons the traditional detection box with edges parallel to the image coordinate axes, instead using a freely rotatable rectangular box to closely fit the actual direction of the lane line. This oriented bounding box is characterized by five parameters: center point coordinates, width, height, and rotation angle. Through image inference, the algorithm outputs an oriented bounding box that encloses the lane line, where the length direction of the box aligns with the lane line direction. The pixel span along the length direction of this oriented bounding box is taken as the precise pixel length of the lane line. At the same time, the rotation angle of the frame The inclination of the lane lines was recorded, providing geometric information for subsequent calculations.

[0038] The oriented bounding box used is a flexible target bounding box representation in computer vision. It breaks the limitation of traditional axis-aligned bounding boxes where edges must be parallel to the image coordinate axes, allowing for free rotation based on the target's actual tilt angle. This bounding box achieves precise target wrapping through a combination of center position, width, height, and rotation angle, thus solving the problem of efficient target localization for tilted targets. In this method, OBB technology enables high-precision detection and pixel length extraction of tilted lane lines, providing an accurate foundation for subsequent geometric calculations.

[0039] In one possible implementation, step S2 specifically includes the following sub-steps: S21. Draw the first virtual detection line and the second virtual detection line at the positions corresponding to the two ends of the lane line of known real length in the image; S22. Detect and track moving vehicles in the video sequence, determine the time points when the midpoint of its bottom edge passes through the first virtual detection line and the second virtual detection line in sequence, and calculate the pixel distance that the midpoint of its bottom edge moves between the two lines. S23. Based on the ratio between the actual length of the lane line and its pixel length, and the pixel distance of the movement of the bottom edge midpoint, calculate the actual distance the moving vehicle moves between the two virtual detection lines. S24. Calculate the actual speed of the moving vehicle based on the actual distance and the time difference between the two virtual detection lines.

[0040] Speed ​​calculation is implemented through the following specific process: First, within the image plane, along positions corresponding to the start and end points of lane lines of known length, a first virtual detection line (Line A) and a second virtual detection line (Line B) are defined, respectively. Both lines are drawn perpendicular to the extension direction of the lane lines. Subsequently, the video sequence is processed frame by frame, and a target detection algorithm is used to locate the vehicle and track its trajectory. The midpoint of the bottom edge of the vehicle's bounding box is selected as the feature tracking point. When this point crosses Line A and Line B, a timestamp accurate to the frame is recorded. and Calculate the time difference Simultaneously, the pixel distance in the lane line direction along the trajectory traversed by the tracking point from Line A to Line B is measured on the image. The pixel distance here This is the pixel length corresponding to the projection of the vehicle's bottom edge midpoint's trajectory onto the lane line of known length in the image. This mapping process ensures that the measured pixel displacement and the lane line, which serves as a benchmark, are in the same image geometric reference frame, making subsequent scaling based on lane line pixel lengths geometrically consistent. Combined with the known actual lane line length of 6 meters and its corresponding total pixel length... Through proportional relationships Calculate the actual distance the vehicle travels on the actual road. Ultimately, the vehicle speed was... The calculation yielded the result.

[0041] In one possible implementation, in step S22, the detection and tracking of moving vehicles is achieved by using a YOLO-based target detection algorithm in conjunction with a tracker.

[0042] The detection of moving vehicles is accomplished by the YOLO object detection algorithm. This algorithm divides the input image into an S x S grid, with each grid cell responsible for predicting targets whose center falls within that region. For each frame, the YOLO model outputs the bounding boxes of all detected vehicles at once, including their position, size, and confidence score. To achieve vehicle association and continuous tracking across different frames, the detection results of each YOLO frame are input into a separate tracker. This tracker assigns a unique ID to each detected target based on motion prediction, appearance features, or data association algorithms, and generates continuous trajectories across frames. The position of the bottom edge midpoint is derived from the bottom center coordinates of the YOLO detection boxes, and the tracker performs inter-frame smoothing and association to ensure timestamp accuracy. , and pixel displacement The accuracy of the calculation.

[0043] The YOLO object detection algorithm employed is a landmark algorithm that refactors the detection task into a single regression problem. Its core idea is to divide the input image into an S×S grid, with each grid cell responsible for predicting targets whose center point falls within it, and outputting the bounding box coordinates, confidence scores, and class probabilities of all targets in one step. This "one-step" architecture gives the algorithm extremely fast processing speed, meeting the needs of real-time video analysis, while effectively reducing false detections because it can see the entire image context. In this project, YOLO serves as the visual perception engine, providing fast and accurate target location information for vehicle detection and tracking initialization.

[0044] In one possible implementation, in step S3, the total duration is obtained by recording the first global timestamp when the moving vehicle is first detected and the second global timestamp when it finally disappears from the screen, and calculating the difference between them.

[0045] The total duration a vehicle remains in the video feed is determined by the lifecycle of its tracking ID. The moment a tracker successfully associates and creates a new vehicle tracking ID for the first time is recorded as the first global timestamp. Subsequently, as long as the vehicle target continues to be detected and associated, the tracking ID remains active. When the vehicle leaves the monitored area or disappears, after several consecutive frames when the ID cannot be associated, the tracker determines that the target has disappeared and records the moment of the last successful association as the second global timestamp. Total duration That is and The difference. The process is fully automated and requires no human intervention.

[0046] In one possible implementation, in step S3, the effective coverage distance is estimated using the following formula: in, For effective coverage distance, The actual velocity calculated in step S2, This represents the total duration.

[0047] Obtain vehicle speed and total duration Subsequently, the effective coverage distance of the surveillance camera on the vehicle. Through formula It is calculated directly. This formula is based on the assumption of uniform motion, treating the vehicle's speed measured at the near end of the camera's field of view as its average speed over the entire tracked road segment. The calculated... The physical meaning is clear: the vehicle is moving at a speed of Under the condition of constant speed travel, at time The distance traveled within the vehicle is equal to the furthest distance the camera can effectively track.

[0048] In one possible implementation, the method further includes step S4 and a result output step: The estimated effective coverage distance is associated with the corresponding video frame or camera identification information, and then visualized or stored.

[0049] After estimating the effective coverage distance, the system executes the result output step. This step outputs the distance value. The data is bound to the source video stream information that generated it, the corresponding camera device identifier, and the timestamp used for calculation, and then stored in a structured manner in a database or file. Simultaneously, the system supports visualization, overlaying the estimated maximum coverage distance onto the video frame. A typical visualization method involves drawing a red line perpendicular to the road direction on the image plane at the location where the vehicle disappeared, and labeling the distance value next to the line, thus visually representing the physical range of the camera's effective coverage.

[0050] In one possible implementation, the method is executed by a computer program deployed on an edge computing device or server for batch estimation of the effective coverage distance of one or more surveillance cameras in a traffic monitoring system that have not undergone internal parameter calibration.

[0051] The method is implemented as a computer program that can be deployed on embedded computing devices at the network edge or on cloud servers. The program operates by receiving video streams from one or more cameras in a traffic monitoring system. Because the method does not rely on internal camera parameter calibration, it is particularly suitable for the rapid evaluation of a large number of historical cameras or newly added cameras whose parameters are unknown. The system supports parallel processing of multiple video streams, sequentially performing calibration object detection, vehicle tracking, velocity calculation, and distance estimation through an automated pipeline, ultimately outputting a batch evaluation report containing the effective coverage distances of all cameras in the system.

[0052] In one possible implementation, the method is used to evaluate the actual monitoring range of surveillance cameras in tunnel, bridge, or highway scenarios, providing data support for traffic management departments to optimize surveillance deployment and evaluate effectiveness.

[0053] This method is specifically applicable to structured traffic scenarios such as highways, tunnels, and bridges. In these scenarios, lane markings are clear and standardized, meeting the criteria for use as markers. By implementing this method, traffic management departments can obtain the real, dynamically updated effective monitoring radius of each surveillance camera within their jurisdiction, rather than relying on theoretical design values. This data is directly used to assess the blind spots and overlaps of existing monitoring points, providing quantitative data support for optimizing and adding monitoring points, and evaluating the overall effectiveness of the entire monitoring system, thereby improving the efficiency of traffic management and emergency response.

[0054] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: (1) In this invention, by utilizing lane lines (such as longitudinal deceleration markings) that are common and standardized in traffic scenarios as a natural length benchmark, the dependence on internal and external parameters of the camera is completely eliminated. This method requires no on-site calibration operations, nor does it require querying or measuring difficult-to-obtain parameters such as the camera's focal length and installation height; it only requires analysis using the video stream output by the camera itself. This significantly reduces the implementation threshold and cost, making it possible to quickly assess the effective coverage distance of a large number of deployed surveillance cameras with unknown or potentially changing parameters, thus solving the core bottleneck of traditional methods that cannot be applied on a large scale due to strong parameter dependence.

[0055] (2) In this invention, by combining pixel measurement of standard-length lane lines with vehicle detection and tracking technology, dynamic conversion and distance estimation from image pixels to real-world distance are achieved. First, the pixel length of the lane lines is accurately obtained based on rotating target detection technology, establishing a stable proportional relationship under the current image viewpoint. Then, by tracking the movement of the vehicle within a known scale area, its true speed is directly calculated. Finally, using this speed and the vehicle's dwell time in the full field of view, the effective coverage distance of the camera is automatically estimated. The entire process is fully automated, requiring no manual intervention to specify reference objects, and by dynamically tracking the vehicle to calculate the speed, the estimation results can truly reflect the monitoring capabilities under the current scene, avoiding errors caused by using fixed models or empirical values.

[0056] (3) The proposed method in this invention has a clear logic, and each step can be efficiently executed by mature computer vision algorithms, making it easy to integrate into a software system. This system can be deployed on servers or edge computing devices to achieve parallel and batch processing of multiple surveillance video streams. This provides traffic management departments with an efficient tool that can periodically or irregularly scan and evaluate the effectiveness of the entire monitoring network and automatically generate coverage distance reports. Based on this accurate and objective data, managers can scientifically identify monitoring blind spots, optimize camera deployment strategies, and assess system upgrade needs, thereby greatly improving the management refinement and proactive maintenance capabilities of the traffic monitoring system, laying a reliable data foundation for the precise control of intelligent transportation.

[0057] 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 variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included 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.

[0058] The following points need to be explained: (1) The accompanying drawings of the embodiments of the present invention only involve the structures involved in the embodiments of the present invention. Other structures can refer to the general design.

[0059] (2) For clarity, the thickness of layers or regions is enlarged or reduced in the drawings used to describe embodiments of the invention, i.e., these drawings are not drawn to scale. It is understood that when an element such as a layer, film, region or substrate is referred to as being “above” or “below” another element, the element may be “directly” located “above” or “below” the other element or there may be intermediate elements.

[0060] (3) Where there is no conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other to obtain new embodiments.

[0061] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. The scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for estimating uncalibrated monitoring distances using lane line markers and vehicle tracking, characterized in that: include: S1. Calibration information acquisition steps: Based on the video image output by the monitoring camera, detect at least one lane line with a known real length and determine the pixel length of the lane line in the image; S2. Short-range vehicle speed calculation steps: Within the image area corresponding to the lane line of known true length, by detecting and tracking moving vehicles, and combining the true length of the lane line and the pixel length, calculate the true speed of the moving vehicle when it passes through the area. S3. Effective coverage distance estimation step: Obtain the total duration from the appearance to the disappearance of the moving vehicle in the monitoring video frame, and estimate the effective coverage distance of the monitoring camera based on the actual movement speed and the total duration.

2. The method for estimating uncalibrated monitoring distance using lane line markers and vehicle tracking according to claim 1, characterized in that, include: In step S1, the lane line of known actual length is the longitudinal deceleration marking on the road, and its standard length is 6 meters.

3. The method for estimating uncalibrated monitoring distance using lane line markers and vehicle tracking according to claim 1, characterized in that, include: In step S1, the lane line is detected using a target detection method based on directional bounding boxes to obtain its precise pixel length and orientation angle in the image.

4. The method for estimating uncalibrated monitoring distance using lane line markers and vehicle tracking according to claim 1, characterized in that, Step S2 specifically includes the following sub-steps: S21. Draw the first virtual detection line and the second virtual detection line at the positions corresponding to the two ends of the lane line of known real length in the image; S22. Detect and track moving vehicles in the video sequence, determine the time points when the midpoint of the bottom edge passes through the first virtual detection line and the second virtual detection line in sequence, and calculate the pixel distance that the midpoint of the bottom edge moves between the two lines. S23. Based on the ratio between the actual length of the lane line and its pixel length, and the pixel distance of the movement of the midpoint of the bottom edge, calculate the actual distance the moving vehicle moves between the two virtual detection lines. S24. Calculate the actual speed of the moving vehicle based on the actual distance and the time difference between the two virtual detection lines.

5. The method for estimating uncalibrated monitoring distance using lane line markers and vehicle tracking according to claim 4, characterized in that, include: In step S22, the detection and tracking of moving vehicles are achieved by using a YOLO-based target detection algorithm in conjunction with a tracker.

6. The method for estimating uncalibrated monitoring distance using lane line markers and vehicle tracking according to claim 1, characterized in that, include: In step S3, the total duration is obtained by recording the first global timestamp when the moving vehicle is first detected and the second global timestamp when it finally disappears from the screen, and calculating the difference between them.

7. The method for estimating uncalibrated monitoring distance using lane line markers and vehicle tracking according to claim 1 or 6, characterized in that, include: In step S3, the effective coverage distance is estimated using the following formula: in, For effective coverage distance, The actual velocity calculated in step S2, The total duration is given.

8. The method for estimating uncalibrated monitoring distance using lane line markers and vehicle tracking according to claim 1, characterized in that, The method further includes step S4 and a result output step: The estimated effective coverage distance is associated with the corresponding video frame or camera identification information, and then visualized or stored.

9. The method for estimating uncalibrated monitoring distance using lane line markers and vehicle tracking according to claim 1, characterized in that, include: The method is executed by a computer program deployed on an edge computing device or server, and is used to perform batch estimation of the effective coverage distance of one or more surveillance cameras in a traffic monitoring system that have not undergone internal parameter calibration.

10. The method for estimating uncalibrated monitoring distance using lane line markers and vehicle tracking according to claim 1, characterized in that, include: The method described above is used to evaluate the actual monitoring range of surveillance cameras in tunnel, bridge, or highway scenarios, providing data support for traffic management departments to optimize surveillance deployment and evaluate effectiveness.

Images