Intelligent vehicle intervention system

Abstract

The application discloses an intelligent vehicle driving intervention system, which comprises a road blocking device arranged at a road side, the road blocking device comprising a camera, an image processing device, a flexible intercepting device and a control device; the flexible intercepting device comprises a mounting bracket, a flexible intercepting belt and a driving motor, two ends of the flexible intercepting belt are connected to the left bracket and the right bracket through two rotating shafts respectively, the flexible intercepting belt is parallel and abuts on a road surface in an initial state, and the flexible intercepting belt is perpendicular to the road surface in an intercepting state; the camera is used for collecting image data of a monitored road section, the image processing device processes the image data, identifies whether there is an accident image in the image data, and if yes, the control device is used for controlling the driving motor to work. The application can intercept a vehicle when an accident occurs, and improves the safety of vehicle driving.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic equipment, and in particular to an intelligent driving intervention system. Background Art

[0002] Currently, traditional road blocking methods rely on manual on-site installation of warning signs. However, manual operation is inefficient, and in poor weather (such as heavy rain, dense fog, and dust) and low-visibility conditions at night, the warning effect is significantly reduced, failing to convey danger information to drivers in a timely and accurate manner. Therefore, developing a device that can automatically detect accidents in real time and quickly block roads is of great practical significance. Summary of the Invention

[0003] An object of the present invention is to solve at least the above problems and / or disadvantages and to provide at least the advantages which will be described hereinafter.

[0004] One object of the present invention is to provide an intelligent driving intervention system, which can intercept vehicles in front of the accident section when an accident occurs on the road, thereby improving the safety of vehicle driving.

[0005] In order to achieve these objectives and other advantages according to the present invention, an intelligent driving intervention system is provided, comprising:

[0006] A road blocking device is provided on the side of a highway, and includes a camera, an image processing device, a flexible intercepting device, and a control device;

[0007] The flexible interception device includes a mounting bracket, a flexible interception belt and a drive motor, the mounting bracket includes a left bracket and a right bracket respectively arranged on both sides of the road, the two ends of the flexible interception belt are respectively connected to the left bracket and the right bracket through two rotating shafts, the drive motor is connected to one of the rotating shafts, the flexible interception belt is parallel to and abuts against the road surface in an initial state, the flexible interception belt is perpendicular to the road surface in an intercepting state, the angle of the flexible interception belt relative to the road surface in a pre-lifting state is 30°, and the drive motor drives one of the rotating shafts to rotate so that the flexible interception belt switches between the initial state, the pre-lifting state and the intercepting state;

[0008] The camera is used to collect image data of the monitored road section, and the image processing device processes the image data to identify whether the image data contains an accident image. If so, the control device is used to control the operation of the drive motor;

[0009] The image processing device integrates a vehicle trajectory prediction algorithm, and a Kalman filter model is used to recursively calculate vehicle motion parameters, including speed, acceleration and steering angle, in 5 consecutive images to predict the vehicle position and attitude in the next 3-5 seconds; when the prediction trajectory overlaps with the accident area with a probability of >90%, a pre-interception instruction is triggered;

[0010] The control device dynamically adjusts the triggering condition of the flexible interception belt according to the predicted vehicle arrival time:

[0011] If the predicted arrival time is >8 seconds, only an audible and light warning is triggered;

[0012] If the predicted arrival time is ≤8 seconds and >3 seconds, the control device is used to control the driving motor to work, so that the flexible interception belt is switched from the initial state to the pre-lifting state;

[0013] If the predicted arrival time is ≤3 seconds, full vertical interception is performed, and the control device is used to control the driving motor to work, so that the flexible interception belt is switched from the initial state to the interception state.

[0014] Preferably, in the intelligent vehicle driving intervention system, the road blocking device comprises a pressure sensor arranged on the road surface behind the flexible interception device, which generates a vehicle detection signal when a vehicle drives over the pressure sensor, and the control device is electrically connected to the pressure sensor; when the image processing device detects an accident image in the image data and the pressure sensor detects the vehicle detection signal, the control device is used to control the driving motor to work, so that the flexible interception belt is switched from the initial state to the interception state.

[0015] Preferably, in the intelligent vehicle driving intervention system, the two ends of the flexible interception belt are connected to the two rotating shafts through two buffer assemblies respectively, the buffer assembly comprises a plurality of springs and dampers, one end of the spring is connected to one end of the flexible interception belt, the other end of the spring is connected to one end of the rotating shaft, and the damper is arranged between one end of the flexible interception belt and one end of the rotating shaft.

[0016] Preferably, the intelligent vehicle driving intervention system further comprises a central control system, which is communicatively connected with a traffic monitoring system and the control device, sends a preliminary interception notification to the control device after obtaining an accident occurrence notification from the traffic monitoring system, and controls the driving motor to work to switch the flexible interception belt from the initial state to the interception state when the preliminary interception notification and the vehicle detection signal are received.

[0017] Preferably, the intelligent driving intervention system further includes a roadside sound and light alarm terminal, which is arranged on the side of the highway and in front of the road blocking device; the central control system is communicatively connected with the roadside sound and light alarm terminal and the traffic monitoring system, and after obtaining the accident notification from the traffic monitoring system, controls the roadside sound and light alarm terminal to emit an sound and light alarm, and sends a preliminary interception notification to the control device. When the control device receives the preliminary interception notification and the vehicle detection signal, it controls the drive motor to operate and switches the flexible interception belt from the initial state to the interception state.

[0018] Preferably, the intelligent driving intervention system further includes a roadside directional high-pitched horn, which is arranged on the side of the highway, and the roadside directional high-pitched horn is arranged between the road blocking equipment and the roadside sound and light alarm terminal; the central control system is communicatively connected with the roadside directional high-pitched horn, and after obtaining the accident notification from the traffic monitoring system, the central control system controls the roadside directional high-pitched horn to issue a voice alarm.

[0019] Preferably, the intelligent driving intervention system also includes a driving induction system, which is arranged near the roadside sound and light alarm terminal and is communicatively connected to the traffic monitoring system, for obtaining the road conditions, recommended driving speed and detour route of the accident section from the traffic monitoring system, and displaying the road conditions, recommended driving speed and detour route of the accident section through an electronic display screen.

[0020] Preferably, the intelligent driving intervention system further includes a lane indicating device, which is arranged near the roadside sound and light alarm terminal and is communicatively connected to the traffic monitoring system, and is used to obtain and display the open and closed status of the lanes on the highway from the traffic monitoring system.

[0021] Preferably, in the intelligent driving intervention system, the vehicle trajectory prediction algorithm of the image processing device includes the following steps:

[0022] The five consecutive frames of images captured by the camera are grayscaled in sequence, Gaussian filtering is used to remove image noise, and the vehicle outline is extracted using an edge detection algorithm;

[0023] A deep learning target detection algorithm is used to identify vehicles in each frame. Feature point matching technology is used to establish the correspondence between feature points of the same vehicle in five consecutive frames. The RANSAC algorithm is used to filter out mismatched points and obtain the vehicle's motion trajectory.

[0024] Define the vehicle's position, velocity, acceleration, and steering angle as four motion parameters as state variables, and construct a dynamic model including a state transfer matrix and a control input matrix;

[0025] The vehicle outline and trajectory extracted from the image are integrated with the prediction results of the dynamic model, and the prediction error is adjusted through the Kalman gain to output the vehicle position and posture prediction for the next 3-5 seconds.

[0026] Based on the processing results of 5 consecutive frames, the vehicle's speed change rate and steering angle change rate are recursively calculated to provide accurate motion parameters for risk assessment.

[0027] Preferably, in the intelligent traffic intervention system, the deep learning target detection algorithm is used to identify the vehicle in each frame of the image, the corresponding relationship between the feature points of the same vehicle in 5 consecutive frames is established by the feature point matching technology, and the RANSAC algorithm is used to filter the mismatched points to obtain the vehicle motion trajectory, including:

[0028] Using an improved YOLOv5 object detection algorithm, after image preprocessing, the deep neural network is fed into the image. Multi-scale features are fused through the Feature Pyramid Network (FPN). This algorithm detects vehicles within a 300×300 pixel image region with a mean average precision (mAP) of 98.5%. The system also outputs the vehicle's location coordinates (x, y), width (w), height (h), and confidence score.

[0029] For the same vehicle in five consecutive frames, the SIFT feature extraction algorithm is used to generate a 128-dimensional feature vector. The nearest neighbor matching (NN) is used to initially establish the correspondence between feature points. The RANSAC algorithm is then used to eliminate mismatched points, increasing the matching accuracy to no less than 99.2%.

[0030] Based on the matched feature point coordinates, the Kalman filter algorithm is used to recursively estimate the vehicle position, calculate the displacement Δx, Δy and velocity v between adjacent frames, and generate a smooth motion trajectory curve.

[0031] The present invention has at least the following beneficial effects:

[0032] The present invention provides an intelligent driving intervention system, comprising: a road blocking device, which is arranged on the side of a highway, the road blocking device comprising a camera, an image processing device, a flexible intercepting device and a control device; the flexible intercepting device comprises a mounting bracket, a flexible intercepting belt and a driving motor, the mounting bracket comprises a left bracket and a right bracket respectively arranged on both sides of the highway, the two ends of the flexible intercepting belt are respectively connected to the left bracket and the right bracket through two rotating shafts, the driving motor is connected to one of the rotating shafts, the flexible intercepting belt is parallel and abuts against the highway surface in an initial state, the flexible intercepting belt is perpendicular to the highway surface in an intercepting state, the angle of the flexible intercepting belt relative to the highway surface in a pre-lifting state is 30 degrees, and the driving motor drives one of the rotating shafts The rotating shaft rotates so that the flexible intercepting belt switches between the initial state, the pre-lifting state and the intercepting state; the camera is used to collect image data of the monitored road section, and the image processing device processes the image data to identify whether there is an accident image in the image data. If so, the control device is used to control the operation of the drive motor; wherein, the image processing device integrates a vehicle trajectory prediction algorithm, and recursively calculates the vehicle motion parameters in 5 consecutive frames of images, including speed, acceleration and steering angle, through the Kalman filter model to predict the vehicle position and posture in the next 3-5 seconds; when the probability of the predicted trajectory overlapping with the accident area is greater than 90%, the pre-interception instruction is triggered; and the control device dynamically adjusts the triggering condition of the flexible intercepting belt according to the predicted vehicle arrival time: if the predicted arrival time is greater than 8 If the predicted arrival time is ≤8 seconds and >3 seconds, only the sound and light warning is triggered; if the predicted arrival time is ≤8 seconds and >3 seconds, the control device is used to control the operation of the drive motor to switch the flexible interception belt from the initial state to the pre-raised state; if the predicted arrival time is ≤3 seconds, a full vertical interception is performed, and the control device is used to control the operation of the drive motor to switch the flexible interception belt from the initial state to the interception state. The present invention accurately and automatically identifies accident images through real-time acquisition and processing of images of monitored road sections, and promptly controls the flexible interception device to switch from the initial state to the interception state, efficiently and reliably blocking the accident section and minimizing the risk of secondary accidents.

[0033] Other advantages, objectives and features of the present invention will be reflected in part from the following description and will be understood by those skilled in the art through study and practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 This is a structural diagram of the intelligent driving intervention system provided by the present invention.

[0035] Figure 2 This is a schematic structural diagram of the flexible intercepting device provided by the present invention. DETAILED DESCRIPTION

[0036] The present invention will be described in further detail below in conjunction with the accompanying drawings so that those skilled in the art can implement the invention with reference to the description.

[0037] like Figures 1 to 2 As shown, the present invention provides an intelligent driving intervention system, comprising: a road blocking device, which is arranged on the side of the highway, the road blocking device comprising a camera, an image processing device, a flexible intercepting device and a control device; the flexible intercepting device comprises a mounting bracket, a flexible intercepting belt 1 and a driving motor 5, the mounting bracket comprises a left bracket 2 and a right bracket 3 respectively arranged on both sides of the highway, the two ends of the flexible intercepting belt 1 are respectively connected to the left bracket 2 and the right bracket 3 through two rotating shafts, the driving motor 5 is connected to one of the rotating shafts, the flexible intercepting belt 1 is parallel and abuts against the highway surface in the initial state, the flexible intercepting belt 1 is perpendicular to the highway surface in the intercepting state, the angle of the flexible intercepting belt relative to the highway surface in the pre-lifting state is 30°, and the driving motor 5 drives One of the rotating shafts rotates to switch the flexible intercepting belt 1 between the initial state, the pre-lifting state and the intercepting state; the camera is used to collect image data of the monitored road section, and the image processing device processes the image data to identify whether there is an accident image in the image data. If so, the control device is used to control the operation of the drive motor; wherein, the image processing device integrates a vehicle trajectory prediction algorithm, and recursively calculates the vehicle motion parameters in 5 consecutive frames of images, including speed, acceleration and steering angle, through a Kalman filter model to predict the vehicle position and posture in the next 3-5 seconds; when the probability of the predicted trajectory overlapping with the accident area is greater than 90%, the pre-interception instruction is triggered; and the control device dynamically adjusts the triggering condition of the flexible intercepting belt according to the predicted vehicle arrival time: if the predicted arrival time is greater than 8 seconds, only the sound and light warning is triggered; if the predicted arrival time is ≤8 seconds and >3 seconds, the control device is used to control the operation of the drive motor to switch the flexible interception belt from the initial state to the pre-lifting state; if the predicted arrival time is ≤3 seconds, a full vertical interception is performed, and the control device is used to control the operation of the drive motor to switch the flexible interception belt from the initial state to the interception state.

[0038] This road blocking device is installed on the side of the highway and consists of four core components: a camera, an image processing device, a flexible interception device, and a control device. These components work together to automatically monitor and block accident sections.

[0039] The device includes a mounting bracket, a flexible intercepting belt 1 and a drive motor 5. The mounting bracket is divided into a left bracket 2 and a right bracket 3 arranged on both sides of the road. The two ends of the flexible intercepting belt 1 are respectively connected to the left bracket 2 and the right bracket 3 by means of two rotating shafts. The drive motor 5 is connected to one of the rotating shafts. In the initial state, the flexible intercepting belt 1 is parallel and tightly attached to the road surface, and will not cause any obstruction to normal traffic; in the intercepting state, the flexible intercepting belt 1 is perpendicular to the road surface, forming an effective blocking barrier; in the pre-lifting state, the flexible intercepting belt 1 is at an angle of 30° relative to the road surface, making preparations for interception.

[0040] When the drive motor 5 receives a start signal from the control device, it begins to operate, driving the connected rotating shaft to rotate. This rotation of the rotating shaft further drives the flexible intercepting belt 1 to rotate about its axis, gradually rotating the flexible intercepting belt 1 from its initial position against the road surface to a position perpendicular to the road surface, thus completing the road blocking operation.

[0041] The camera is mounted above the road, in front of the flexible interceptor, to continuously collect image data of the monitored road section. The camera is electrically connected to the image processing device, enabling real-time and rapid transmission of the captured image data to the device.

[0042] The image processing device uses an image recognition algorithm to perform in-depth processing on the image data transmitted by the camera. By analyzing features such as vehicle posture, road obstacles, and smoke, it accurately identifies whether the image data contains accident images. For example, if an abnormal posture such as a rollover or rear-end collision is detected, or if there are obvious signs of an accident such as unidentified obstacles on the road or smoke, the image is identified as an accident image. Furthermore, the image processing device integrates a vehicle trajectory prediction algorithm, and recursively calculates the vehicle motion parameters, including speed, acceleration and steering angle, in 5 consecutive frames of images through a Kalman filter model to predict the vehicle position and posture in the next 3-5 seconds; when the probability of the predicted trajectory overlapping with the accident area is greater than 90%, a pre-interception instruction is triggered; and the control device dynamically adjusts the triggering conditions of the flexible interception belt according to the predicted vehicle arrival time: if the predicted arrival time is greater than 8 seconds, only the sound and light warning is triggered; if the predicted arrival time is ≤8 seconds and greater than 3 seconds, the control device is used to control the drive motor to switch the flexible interception belt from the initial state to the pre-lifting state; if the predicted arrival time is ≤3 seconds, a full vertical interception is performed, and the control device is used to control the drive motor to switch the flexible interception belt from the initial state to the interception state.

[0043] The control device is electrically connected to the image processing device and the drive motor 5. When the image processing device detects an accident image in the image data, it immediately sends a signal to the control device. Upon receiving this signal, the control device quickly responds by controlling the drive motor 5, switching the flexible intercepting strip 1 from its initial state to its intercepting state, thereby automatically and rapidly blocking the accident section.

[0044] Leveraging real-time image acquisition and processing technology, this device can identify accidents in a very short time and immediately activate the flexible interception device. Compared with traditional manual blocking methods, this greatly shortens reaction time and significantly reduces the likelihood of secondary accidents. Using image recognition algorithms, the image processing device can accurately identify accident images, effectively reducing misjudgments and missed detections, greatly improving the accuracy of road blocking and providing reliable protection for road safety. The entire road blocking device is highly automated and requires no on-site manual operation. This not only reduces labor costs but also avoids the risks faced by humans working in dangerous environments, improving the safety and convenience of road blocking operations.

[0045] In a preferred embodiment, in the intelligent driving intervention system, the road blocking device includes a pressure sensor, which is arranged on the highway surface and located behind the flexible intercepting device, and is used to generate a vehicle detection signal when the vehicle travels above the pressure sensor, and the control device is electrically connected to the pressure sensor; when the image processing device detects that there is an accident image in the image data, and the pressure sensor detects and generates the vehicle detection signal, the control device is used to control the drive motor 5 to work, and switch the flexible intercepting belt 1 from the initial state to the intercepting state.

[0046] The pressure sensor is installed on the road surface, in a suitable position behind the flexible interceptor. Accurately bury the pressure sensor according to the pressure sensor installation requirements. Ensure that the pressure sensor is flush with the road surface and the surrounding road surface is smooth so as not to affect normal vehicle operation. Ensure that the electrical connection between the pressure sensor and the control device is stable and reliable.

[0047] The pressure sensor is installed on the road surface, behind the flexible interceptor. When a vehicle passes over the pressure sensor, it generates a vehicle detection signal and transmits it to the control unit, providing the control unit with real-time vehicle location information to assist in determining whether to activate the flexible interceptor.

[0048] In a preferred embodiment, in the intelligent driving intervention system, the two ends of the flexible intercepting belt 1 are respectively connected to the two rotating shafts through two buffer components 4, and the buffer component 4 includes a plurality of springs and dampers. One end of the spring is connected to one end of the flexible intercepting belt 1, and the other end is connected to one end of the rotating shaft. The damper is arranged between one end of the flexible intercepting belt 1 and one end of the rotating shaft.

[0049] The buffer assembly 4 provided between the flexible intercepting belt 1 and the rotating shaft can effectively absorb and consume the impact energy when a vehicle collides, reducing the damage to the vehicle and personnel, and further improving the safety of the road blocking equipment in practical applications.

[0050] When a vehicle strikes the flexible intercepting strip 1, the buffer assembly 4 comes into play. The spring first acts as a buffer, absorbing some of the impact force and allowing the flexible intercepting strip 1 to move freely during the impact. Simultaneously, the damper dissipates the impact energy through its own damping action, mitigating the sway and rebound of the flexible intercepting strip 1. This prevents secondary damage to the vehicle caused by excessive rebound, ensuring safety and stability during the interception process.

[0051] In a preferred embodiment, the intelligent driving intervention system further includes a central control system, which is communicatively connected to the traffic monitoring system and the control device. After obtaining the accident notification from the traffic monitoring system, the central control system sends a preliminary interception notification to the control device. When the control device receives the preliminary interception notification and the vehicle detection signal, it controls the drive motor 5 to operate and switches the flexible interception belt 1 from the initial state to the interception state.

[0052] The central control system communicates with the traffic monitoring system and control devices. Upon receiving an accident notification from the traffic monitoring system, the central control system sends a preliminary interception notification to the control device. The traffic monitoring system can obtain accident information through various means, such as feedback from other road sensors and reports from traffic police. The central control system integrates this information and, upon confirming an accident, sends a preliminary interception notification to the control device in advance, enabling the equipment to prepare for interception and further improving the timeliness and accuracy of interception.

[0053] Select a suitable computer room or control center to install the central control system and ensure a stable operating environment. Establish a communication connection between the central control system and the traffic monitoring system, either using a wired network or a wireless network (such as 4G / 5G) to ensure timely and accurate information transmission. Also, set the communication parameters between the central control system and the control device to ensure that pre-interception notifications are accurately delivered.

[0054] In a preferred embodiment, the intelligent vehicle intervention system further comprises a roadside sound and light alarm terminal, which is arranged on the roadside of the road and in front of the road blocking device; the central control system is in communication connection with the roadside sound and light alarm terminal and the traffic monitoring system, after obtaining the accident occurrence notification from the traffic monitoring system, controls the roadside sound and light alarm terminal to issue sound and light alarm, and sends a preliminary interception notification to the control device, and the control device controls the driving motor 5 to work when receiving the preliminary interception notification and the vehicle detection signal, and switches the flexible interception belt 1 from the initial state to the interception state.

[0055] The roadside sound and light alarm terminal is arranged on the roadside at a certain distance in front of the disaster point, and when a disaster accident is detected, it immediately issues strong light flashing and high-decibel alarm sound to attract the attention of the driver.

[0056] In a preferred embodiment, the intelligent vehicle intervention system further comprises a roadside directional high-pitched horn, which is arranged on the roadside of the road, and the roadside directional high-pitched horn is arranged between the road blocking device and the roadside sound and light alarm terminal; the central control system is in communication connection with the roadside directional high-pitched horn, and the central control system controls the roadside directional high-pitched horn to issue a voice alarm after obtaining the accident occurrence notification from the traffic monitoring system.

[0057] The roadside directional high-pitched horn sends a voice alarm to the vehicle to inform the driver of the disaster situation in front of him and guide him to slow down.

[0058] These devices are reasonably combined and deployed at different distance intervals in front of the disaster point. For example, at a position far from the disaster point, the roadside sound and light alarm terminal and the vehicle induction system are used for preliminary indication and guidance; at a slightly closer position, the roadside directional high-pitched horn and the lane indication device are used for speed reduction induction and lane guidance; and when closer to the disaster point, the visual-based road blocking device is used for flexible interception.

[0059] In a preferred embodiment, the intelligent vehicle intervention system further comprises a vehicle induction system, which is arranged near the roadside sound and light alarm terminal and is in communication connection with the traffic monitoring system, and is used to obtain the road condition, recommended driving speed and detour route of the accident section from the traffic monitoring system, and display the road condition, recommended driving speed and detour route of the accident section through an electronic display screen.

[0060] The vehicle induction system displays the information such as the road condition in front, the recommended driving speed, and the detour route through the electronic display screen to guide the orderly driving of the vehicle.

[0061] In a preferred embodiment, the intelligent driving intervention system further includes a lane indication device, which is arranged near the roadside sound and light alarm terminal and is communicatively connected to the traffic monitoring system, and is used to obtain and display the open and closed status of the lanes on the highway from the traffic monitoring system.

[0062] Lane indicator equipment dynamically displays the open or closed status of the lane, reasonably guiding vehicles to change lanes and avoid congestion.

[0063] In a preferred embodiment, in the intelligent driving intervention system, the vehicle trajectory prediction algorithm of the image processing device includes the following steps: gray-scaling five consecutive frames of images captured by the camera in sequence, removing image noise using Gaussian filtering, and extracting the vehicle contour through an edge detection algorithm; using a deep learning target detection algorithm to identify the vehicle in each frame of the image, establishing a correspondence between feature points of the same vehicle in five consecutive frames through feature point matching technology, and filtering out mismatched points using the RANSAC algorithm to obtain the vehicle motion trajectory; defining four motion parameters of the vehicle's position, speed, acceleration, and steering angle as state variables, and constructing a dynamic model including a state transfer matrix and a control input matrix; fusing the vehicle contour and vehicle motion trajectory extracted from the image with the prediction results of the dynamic model, adjusting the prediction error through the Kalman gain, and outputting a prediction of the vehicle position and posture in the next 3-5 seconds; based on the processing results of five consecutive frames, recursively calculating the vehicle's speed change rate and steering angle change rate to provide accurate motion parameters for risk assessment.

[0064] In a preferred embodiment, the intelligent traffic intervention system uses a deep learning target detection algorithm to identify vehicles in each frame, establishes feature point correspondences for the same vehicle in five consecutive frames using feature point matching, and simultaneously uses a RANSAC algorithm to filter out mismatched points to obtain the vehicle's motion trajectory. The system includes: using an improved YOLOv5 target detection algorithm, inputting the image into a deep neural network after image preprocessing, fusing multi-scale features using a feature pyramid network (FPN), identifying vehicle targets with a mean average precision (mAP) of 98.5% within a 300×300 pixel image region, and outputting the vehicle's position coordinates x, y, width w, height h, and confidence score; for the same vehicle in five consecutive frames, using a SIFT feature extraction algorithm to generate a 128-dimensional feature vector, using nearest neighbor matching (NN) to preliminarily establish feature point correspondences, and then using a RANSAC algorithm to eliminate mismatched points, thereby improving the matching accuracy to no less than 99.2%; and recursively estimating the vehicle's position based on the matched feature point coordinates using a Kalman filter algorithm, calculating the displacement Δx, Δy, and velocity v between adjacent frames, and generating a smooth motion trajectory curve.

[0065] Although the embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and exemplary embodiments. They can be applied to a variety of fields suitable for the present invention. Further modifications will be readily apparent to those skilled in the art. Therefore, the present invention is not limited to the specific details and illustrations shown and described herein without departing from the general concept defined by the claims and their equivalents.

Claims

1. An intelligent driving intervention system, characterized in that: include: A road blocking device is provided on the side of a highway, and includes a camera, an image processing device, a flexible intercepting device, and a control device; The flexible interception device includes a mounting bracket, a flexible interception belt and a drive motor, the mounting bracket includes a left bracket and a right bracket respectively arranged on both sides of the road, the two ends of the flexible interception belt are respectively connected to the left bracket and the right bracket through two rotating shafts, the drive motor is connected to one of the rotating shafts, the flexible interception belt is parallel to and abuts against the road surface in an initial state, the flexible interception belt is perpendicular to the road surface in an intercepting state, the angle of the flexible interception belt relative to the road surface in a pre-lifting state is 30°, and the drive motor drives one of the rotating shafts to rotate so that the flexible interception belt switches between the initial state, the pre-lifting state and the intercepting state; The camera is used to collect image data of the monitored road section, and the image processing device processes the image data to identify whether the image data contains an accident image. If so, the control device is used to control the operation of the drive motor; The image processing device integrates a vehicle trajectory prediction algorithm and uses a Kalman filter model to recursively calculate vehicle motion parameters, including speed, acceleration, and steering angle, in five consecutive frames of imagery to predict the vehicle's position and posture within the next 3-5 seconds. When the probability of the predicted trajectory coinciding with the accident area is greater than 90%, a pre-interception command is triggered. And the control device dynamically adjusts the triggering conditions of the flexible interception belt according to the predicted vehicle arrival time: If the predicted arrival time is greater than 8 seconds, only the sound and light warning will be triggered; If the predicted arrival time is ≤8 seconds and >3 seconds, the control device is used to control the drive motor to switch the flexible intercepting belt from the initial state to the pre-lifting state; If the predicted arrival time is ≤ 3 seconds, a full vertical interception is performed, and the control device is used to control the operation of the drive motor to switch the flexible interception belt from the initial state to the interception state.

2. The intelligent driving intervention system according to claim 1, characterized in that: The road blocking device includes a pressure sensor, which is arranged on the road surface and located behind the flexible intercepting device, and is used to generate a vehicle detection signal when a vehicle travels above the pressure sensor, and the control device is electrically connected to the pressure sensor; When the image processing device detects that the image data contains an accident image and the pressure sensor detects and generates the vehicle detection signal, the control device is used to control the operation of the drive motor to switch the flexible interception belt from the initial state to the interception state.

3. The intelligent driving intervention system according to claim 1, characterized in that: The two ends of the flexible intercepting belt are respectively connected to the two rotating shafts through two buffer components, and the buffer components include multiple springs and dampers. One end of the spring is connected to one end of the flexible intercepting belt, and the other end is connected to one end of the rotating shaft. The damper is arranged between one end of the flexible intercepting belt and one end of the rotating shaft.

4. The intelligent driving intervention system according to claim 2, characterized in that: It also includes a central control system, which is communicatively connected to the traffic monitoring system and the control device. After obtaining the accident notification from the traffic monitoring system, the central control system sends a preliminary interception notification to the control device. When the control device receives the preliminary interception notification and the vehicle detection signal, it controls the drive motor to operate and switches the flexible interception belt from the initial state to the interception state.

5. The intelligent driving intervention system according to claim 4, characterized in that: It also includes a roadside sound and light alarm terminal, which is arranged on the side of the highway and in front of the road blocking device; the central control system is communicated with the roadside sound and light alarm terminal and the traffic monitoring system, and after obtaining the accident notification from the traffic monitoring system, controls the roadside sound and light alarm terminal to emit an sound and light alarm, and sends a preparatory interception notification to the control device. When the control device receives the preparatory interception notification and the vehicle detection signal, it controls the drive motor to work and switches the flexible interception belt from the initial state to the interception state.

6. The intelligent driving intervention system according to claim 5, characterized in that: It also includes a roadside directional high-pitched horn, which is arranged on the side of the highway and between the road blocking equipment and the roadside sound and light alarm terminal; the central control system is communicatively connected with the roadside directional high-pitched horn, and after obtaining the accident notification from the traffic monitoring system, the central control system controls the roadside directional high-pitched horn to issue a voice alarm.

7. The intelligent driving intervention system according to claim 6, characterized in that: It also includes a driving induction system, which is arranged near the roadside sound and light alarm terminal and is communicated with the traffic monitoring system. It is used to obtain the road conditions, recommended driving speed and detour route of the accident section from the traffic monitoring system, and display the road conditions, recommended driving speed and detour route of the accident section through an electronic display screen.

8. The intelligent driving intervention system according to claim 7, characterized in that: It also includes a lane indication device, which is arranged near the roadside sound and light alarm terminal and is communicatively connected to the traffic monitoring system, and is used to obtain and display the open and closed status of the lanes on the highway from the traffic monitoring system.

9. The intelligent driving intervention system according to claim 1, characterized in that: The vehicle trajectory prediction algorithm of the image processing device includes the following steps: The five consecutive frames of images captured by the camera are grayscaled in sequence, Gaussian filtering is used to remove image noise, and the vehicle outline is extracted using an edge detection algorithm; A deep learning target detection algorithm is used to identify vehicles in each frame. Feature point matching technology is used to establish the correspondence between feature points of the same vehicle in five consecutive frames. The RANSAC algorithm is used to filter out mismatched points and obtain the vehicle's motion trajectory. Define the vehicle's position, velocity, acceleration, and steering angle as four motion parameters as state variables, and construct a dynamic model including a state transfer matrix and a control input matrix; The vehicle outline and trajectory extracted from the image are integrated with the prediction results of the dynamic model, and the prediction error is adjusted through the Kalman gain to output the vehicle position and posture prediction for the next 3-5 seconds. Based on the processing results of 5 consecutive frames, the vehicle's speed change rate and steering angle change rate are recursively calculated to provide accurate motion parameters for risk assessment.

10. The intelligent driving intervention system according to claim 9, characterized in that: The method uses a deep learning target detection algorithm to identify vehicles in each frame, establishes the corresponding relationship between feature points of the same vehicle in five consecutive frames through feature point matching technology, and uses the RANSAC algorithm to filter out mismatched points to obtain the vehicle's motion trajectory, including: Using an improved YOLOv5 object detection algorithm, after image preprocessing, the deep neural network is fed into the image. Multi-scale features are fused through the Feature Pyramid Network (FPN). This algorithm detects vehicles within a 300×300 pixel image region with a mean average precision (mAP) of 98.5%. The system also outputs the vehicle's location coordinates (x, y), width (w), height (h), and confidence score. For the same vehicle in five consecutive frames, the SIFT feature extraction algorithm is used to generate a 128-dimensional feature vector. The nearest neighbor matching (NN) is used to initially establish the correspondence between feature points. The RANSAC algorithm is then used to eliminate mismatched points, increasing the matching accuracy to no less than 99.2%. Based on the matched feature point coordinates, the Kalman filter algorithm is used to recursively estimate the vehicle position, calculate the displacement Δx, Δy and velocity v between adjacent frames, and generate a smooth motion trajectory curve.

Images