A high-speed information acquisition, analysis and processing system and method based on machine vision

Abstract

This invention discloses a high-speed information acquisition, analysis, and processing system and method based on machine vision, relating to the field of traffic technology. It aims to address the problem that if drivers involved in a traffic accident do not proactively move their vehicles to the emergency lane, prolonged traffic congestion will occur on highways. Subsequent vehicles cannot observe this problem, leading to multi-lane cutting and further disrupting traffic flow. The key technical points of the solution include: a gantry, a high-speed monitoring system mounted on the gantry, an electronic indicator mounted on the gantry, and a storage medium. The storage medium stores computer-executable instructions, which are used to cause a computer to execute the steps included in the method. This invention can analyze the situation on highways based on data captured by the high-speed monitoring system, thereby making advance predictions and enabling timely lane-changing guidance.

Description

Technical Field

[0001] This invention relates to the field of transportation technology, and more specifically, to a high-speed information acquisition, analysis and processing system and method based on machine vision. Background Technology

[0002] Traffic accidents are common on highways. Usually, after an accident occurs, the parties involved report it to the police, and traffic police direct traffic at the scene to prevent excessive congestion. The incident is then assessed based on surveillance footage, or, in some cases, by personnel monitoring the surveillance footage in real time and discovering an accident, they immediately dispatch traffic control personnel to handle the situation. However, on most highways, accidents are not detected immediately. If the drivers involved in the accident do not move their vehicles to the emergency lane, a prolonged traffic jam will occur on the highway. Subsequent vehicles may not notice the problem, leading to multiple lanes cutting in, further disrupting the overall traffic flow.

[0003] Therefore, a new solution is needed to address this problem. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the purpose of this invention is to provide a high-speed information acquisition, analysis and processing system and method based on machine vision.

[0005] The above-mentioned technical objective of the present invention is achieved through the following technical solution: a high-speed information acquisition, analysis and processing system based on machine vision, including a gantry, a high-speed monitoring system installed on the gantry, an electronic indicator installed on the gantry, and a storage medium, wherein the storage medium stores computer-executable instructions, which are used to cause a computer to execute the steps included in the method.

[0006] This invention provides a high-speed information acquisition, analysis, and processing method based on machine vision, comprising the following steps:

[0007] Step 1: Collect highway information based on two adjacent sets of highway monitoring equipment on the same road segment;

[0008] Step 2: Determine whether a car accident has occurred based on the highway information, and whether the car accident occurred within the range of the highway monitoring camera. If a car accident has occurred and is within the range of the highway monitoring camera, proceed to Step 3. If a car accident has occurred but is not within the range of the highway monitoring camera, proceed to Steps 4-5.

[0009] Step 3: Confirm the lane through the footage captured by the highway monitoring system corresponding to the accident, and set a time standard for the accident vehicle to move into the emergency lane. Compare the real-time time of the accident vehicle moving into the emergency lane in the footage. If it is less than the time standard, the electronic sign will not light up. If it is greater than or equal to the time standard, the electronic sign on the accident lane will light up.

[0010] Step 4: The two sets of highway monitoring cameras collect the passing speed, passing time, and distance between the two sets of highway monitoring cameras to estimate the location of the accident.

[0011] Step 5: Set the traffic congestion standard. If the speed of vehicles passing through the previous set of highway monitoring is consistently lower than the traffic congestion standard, the electronic indicator at the accident location will light up. If the speed of vehicles passing through the previous set of highway monitoring is not consistently lower than the traffic congestion standard, the electronic indicator will not light up.

[0012] The present invention is further configured such that: the method in step 2 for determining whether a car accident has occurred and whether the car accident occurred within the shooting range of the highway monitoring system based on the highway information is as follows:

[0013] Within the time frame, the images captured by the two sets of highway monitoring cameras are used to determine whether there are stationary vehicles. If there are, it is determined that a car accident has occurred and the accident is within the range of the highway monitoring camera. If there are no stationary vehicles, the passing speed under the previous set of highway monitoring cameras is compared with the traffic jam standard. If the passing speed under the previous set of highway monitoring cameras is consistently lower than the traffic jam standard, it is determined that a car accident has occurred and the accident is not within the range of the highway monitoring camera.

[0014] The present invention is further configured such that the time standard is 10 minutes.

[0015] The present invention is further configured such that the traffic jam standard is 50 km / h.

[0016] The present invention is further configured such that: the method for estimating the location of the car accident in step 4 is as follows:

[0017] The previous set of highway surveillance footage was divided into several lanes, with lane 1, lane 2, ..., lane 3 as the boundaries. n Furthermore, every 10 seconds, the similarity of the implementation images of the same lane is calculated, and the similarity score (sim) of each lane at each time period is obtained. n t n And calculate the average similarity S1, S2, ..., Sn for each lane, where S1 = (∑s im1t) n ) / 60、S2=(∑s im2t n ) / 60、...、Sn=(∑s im n t n ) / 60, average similarity of total lanes It is (∑Sn) / n, and according to The variance is calculated with the average similarity S1, S2, ..., Sn of each lane, and the lane with the smallest deviation is the location of the accident.

[0018] The present invention is further configured such that the electronic sign displays a "No Entry" message.

[0019] In summary, the present invention has the following beneficial effects:

[0020] The aforementioned system and method enable the system to assess situations such as traffic accidents or vehicles traveling below the minimum speed limit on highways, allowing for advance lane changes for following vehicles. This prevents vehicles from being stuck in the accident lane or the lane where the slow-moving vehicle is located, potentially causing further traffic congestion by trying to change lanes or cut in after encountering an accident or slow-moving vehicle. Furthermore, the system and method utilize data currently monitored by highway surveillance systems, allowing for further analysis to arrive at conclusions. This reduces the difficulty of data acquisition and accelerates the efficiency of providing final conclusions. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the structure of the present invention. Detailed Implementation

[0022] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0023] Example:

[0024] A high-speed information acquisition, analysis and processing system based on machine vision includes a gantry, a high-speed monitor installed on the gantry, an electronic sign installed on the gantry, and a storage medium. The electronic sign displays a "No Entry" message and is actually an LED screen. The storage medium stores computer-executable instructions that are used to cause a computer to execute the steps included in the method.

[0025] A high-speed information acquisition, analysis, and processing method based on machine vision, such as... Figure 1 As shown, it includes the following steps:

[0026] Step 1: Collect highway information from two adjacent sets of highway monitoring systems on the same road segment. This information includes captured images, the real-time time it takes for a vehicle involved in an accident to move into the emergency lane, the speed and duration of vehicles passing through the same group, and the distance between the two sets of highway monitoring systems. This information is crucial for determining whether traffic congestion has occurred on the highway. Since this information primarily reflects factors that change during traffic jams, using it to determine congestion is more reasonable and accurate. While existing highway monitoring technologies can statistically analyze this information, they do not perform calculations or analyses. Therefore, this approach makes the entire system more efficient, eliminating the need to collect additional information.

[0027] Step 2: Determine whether a car accident has occurred based on the highway information, and whether the car accident occurred within the range of the highway monitoring camera. If a car accident has occurred and is within the range of the highway monitoring camera, proceed to Step 3. If a car accident has occurred but is not within the range of the highway monitoring camera, proceed to Steps 4-5.

[0028] In step 2, the method for determining whether a car accident has occurred and whether the accident occurred within the range of the highway surveillance camera based on highway information is as follows:

[0029] Within the time frame for moving the vehicle involved in the accident to the emergency lane, the images captured by the two sets of highway surveillance cameras are analyzed using image recognition algorithms to determine whether a stationary vehicle is consistently present. The specific time frame is 10 minutes. If a stationary vehicle is present, it is determined that a car accident has occurred and is within the range of the highway surveillance cameras. If not, the passing speed under the previous set of highway surveillance cameras is compared with the traffic congestion standard. If the passing speed under the previous set of highway surveillance cameras is consistently lower than the traffic congestion standard, it is determined that a car accident has occurred and is not within the range of the highway surveillance cameras.

[0030] The specific image recognition algorithm is as follows: It acquires several high-speed monitoring images t11-t1n from the previous group and several high-speed monitoring images t21-t2n from the next group. Both t11-t1n and t21-t2n are divided into 20-second time units. Specifically, it extracts t11, t1n, and 3-5 intermediate values ​​from t11-t1n; similarly, it extracts t21, t2n, and 3-5 intermediate values ​​from t21-t2n (specifically, 4 intermediate values). Then, it extracts t11, t1n, and t2n. n and 4 intermediate values ​​or t21, t2n and 4 intermediate values ​​are used for subsequent steps. Step 1.0 Calculate the grayscale histogram of the six captured images, divide the pixel grayscale values ​​into several grayscale levels (usually 256 levels), and count the number of pixels at each grayscale level. Step 2.0 For each grayscale level in the source image, calculate its cumulative probability value in the cumulative distribution function (CDF). Step 3.0 For each grayscale level in the target image, calculate its cumulative probability value in the cumulative distribution function (CDF). Step 4.0 For each pixel in the source image, find the grayscale level in the target image that is closest to its cumulative probability, and replace the grayscale value of the pixel in the source image with the grayscale level of the target image. Step 5.0 Repeat step 4.0 until all pixels in the source image have been processed. Step 6.0 Calculate the grayscale histogram of the six captured images after processing. Step 7.0 Binarize each image. The image is processed by converting it into a binary image with only black and white pixel values. Step 8.0 performs connected component analysis on each image to find all connected components. Step 9.0 extracts features from each connected component, using common features such as area, perimeter, and shape descriptors. Step 10.0 compares the shape features of corresponding connected components in two images. Various distance metrics (such as Euclidean distance, Manhattan distance, and cosine similarity) can be used to evaluate their similarity or difference. If the final similarity is high, that is, the difference is small, it is determined that there is a stationary vehicle in the captured image. If the stationary vehicle has been stationary for more than a certain period of time, it is determined that a car accident has occurred and the accident is within the shooting range of the highway surveillance camera. Otherwise, it is determined that no accident has occurred. Through the above method, people can obtain information about whether a car accident has occurred through relatively intuitive image comparison without having to observe with the naked eye.

[0031] The specific code for steps 1.0 through 5.0 is shown below.

[0032]

[0033]

[0034] Step 3: Confirm the lane via the highway surveillance footage corresponding to the accident, and set a time standard for the accident vehicle to move into the emergency lane. Compare the real-time time of the accident vehicle moving into the emergency lane with the footage. If it is less than the time standard, the electronic sign will not light up; if it is greater than or equal to the time standard, the electronic sign in the accident lane will light up. At this point, simply perform steps 1.0-10.0 in Step 2, and extend the time unit of the imported footage from 10 minutes to more than 10 minutes, and increase the number of imported footage to determine whether an accident has occurred. If the time limit is exceeded, it means that the vehicle involved in the accident has not been moved to the emergency lane and is occupying the normal driving lane. If this time exceeds 10 minutes, it will inevitably cause traffic jams on the highway. Therefore, the electronic indicator lights on the accident lane will be turned on in advance to allow following vehicles to switch to lanes without accidents, ensuring smooth traffic flow and reducing the occurrence of vehicles driving in the accident lane and blocking the way in front of the accident vehicle. It also reduces the problem of vehicles cutting into non-accident lanes, causing congestion or secondary accidents in non-accident lanes.

[0035] Step 4: The two sets of highway monitoring cameras collect data on the speed and duration of the same group of vehicles passing through, as well as the distance between the two sets of monitoring cameras, to estimate the location of the accident.

[0036] The method for estimating the location of the car accident in step 4 is as follows:

[0037] The previous set of highway surveillance footage was divided into several lanes, with lane 1, lane 2, ..., lane 3 as the boundaries. n Furthermore, every 10 seconds, the similarity of the implementation images of the same lane is calculated, and the similarity score (sim) of each lane at each time period is obtained. n t n And calculate the average similarity S1, S2, ..., Sn for each lane, where S1 = (∑s im1t) n ) / 60、S2=(∑s im2t n ) / 60、...、Sn=(∑s im n t n ) / 60, average similarity of total lanes It is (∑Sn) / n, and according to The variance is calculated with the average similarity S1, S2, ..., Sn of each lane, and the lane with the smallest deviation is the location of the accident.

[0038] When a traffic accident occurs outside of highway surveillance cameras, the cameras cannot determine whether an accident has occurred based on direct footage. If an accident happens in any lane, the other non-accident lanes should be in a more congested state, meaning the traffic flow in the non-accident lanes should be faster than in the accident lane. Therefore, after dividing the footage by lane, the lane with the fastest traffic flow indicates that there is no traffic jam or that the impact is less. The traffic flow speed of a lane can be determined by observing the changes in the number of vehicles in that lane. If the number of vehicles in a lane changes significantly after a certain period of time, it indicates that the congestion in that lane is less severe and it is a non-accident lane. Therefore, this method can be used to determine which lane experienced an accident when highway surveillance cameras cannot directly capture the accident scene, thereby guiding subsequent vehicles.

[0039] Step 5: Set the congestion standard. The congestion standard is 50 km / h. Since the minimum speed limit on highways is 60 km / h, and no penalty is imposed for driving at speeds within 20% of the minimum speed limit on highways, it is rounded down to 50 km / h. If the speed is lower than this, regardless of whether a car accident has occurred, it can guide subsequent vehicles to avoid the slow-moving vehicle in advance and avoid congestion. If the speed of vehicles passing under the previous set of highway monitoring is consistently lower than the congestion standard, the electronic indicator at the accident location will light up. If the speed of vehicles passing under the previous set of highway monitoring is not consistently lower than the congestion standard, the electronic indicator will not light up.

[0040] The above-mentioned overall solution enables following vehicles to avoid traffic jams in advance, regardless of whether the accident is within the range of highway surveillance cameras or whether an accident has occurred, as long as there is a slow or stationary situation below the congestion standard, the following vehicles can be instructed to avoid the lane ahead and make arrangements in advance. This reduces the situation where vehicles change lanes and cut in front of other vehicles when they reach a slow or stationary area, thus reducing the congestion problem caused by vehicles changing lanes and cutting in front of others.

[0041] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A high-speed information acquisition, analysis, and processing system based on machine vision, characterized in that: It includes a gantry, a high-speed monitoring system installed on the gantry, an electronic indicator installed on the gantry, and a storage medium, wherein the storage medium stores computer-executable instructions, which are used to enable a computer to execute a high-speed information acquisition, analysis and processing method; Specifically, the high-speed information acquisition, analysis, and processing method includes: Step 1: Collect highway information based on two adjacent sets of highway monitoring equipment on the same road segment; Step 2: Determine whether a car accident has occurred based on the highway information, and whether the car accident occurred within the range of the highway monitoring camera. If a car accident has occurred and is within the range of the highway monitoring camera, proceed to Step 3. If a car accident has occurred but is not within the range of the highway monitoring camera, proceed to Steps 4-5. Step 3: Confirm the lane through the footage captured by the highway monitoring system corresponding to the accident, and set a time standard for the accident vehicle to move into the emergency lane. Compare the real-time time of the accident vehicle moving into the emergency lane in the footage. If it is less than the time standard, the electronic sign will not light up. If it is greater than or equal to the time standard, the electronic sign on the accident lane will light up. Step 4: The two sets of highway monitoring cameras collect the passing speed, passing time, and distance between the two sets of highway monitoring cameras to estimate the location of the accident. Step 5: Set the traffic congestion standard. If the speed of vehicles passing through the previous set of highway monitoring is consistently lower than the traffic congestion standard, the electronic indicator at the accident location will light up. If the speed of vehicles passing through the previous set of highway monitoring is not consistently lower than the traffic congestion standard, the electronic indicator will not light up.

2. The high-speed information acquisition, analysis and processing system based on machine vision according to claim 1, characterized in that: Step 2 involves determining whether a car accident has occurred and whether it occurred within the range of the highway surveillance camera based on highway information. Within the time frame, the images captured by the two sets of highway monitoring cameras are used to determine whether there are stationary vehicles. If there are, it is determined that a car accident has occurred and the accident is within the range of the highway monitoring camera. If there are no stationary vehicles, the passing speed under the previous set of highway monitoring cameras is compared with the traffic jam standard. If the passing speed under the previous set of highway monitoring cameras is consistently lower than the traffic jam standard, it is determined that a car accident has occurred and the accident is not within the range of the highway monitoring camera.

3. The high-speed information acquisition, analysis, and processing system based on machine vision according to claim 1, characterized in that: The time standard is 10 minutes.

4. The high-speed information acquisition, analysis and processing system based on machine vision according to claim 1, characterized in that: The traffic jam standard is 50 km / h.

5. The high-speed information acquisition, analysis and processing system based on machine vision according to claim 1, characterized in that: The method for estimating the location of the car accident in step 4 is as follows: The previous set of highway surveillance footage was divided into several lanes, with lane 1, lane 2, ..., lane as the boundaries. n Furthermore, every 10 seconds, the similarity of the implementation images of the same lane is calculated, and the similarity score of each lane at each time period is obtained. n t n And calculate the average similarity S1, S2, ..., Sn for each lane, where S1 = (∑sim1t) n ) / 60、S2=(∑sim2t) n ) / 60、...、Sn=(∑sim n t n The total average similarity of lanes, `Sn, is (∑Sn) / n. The variance is calculated based on `Sn and the average similarity of each lane, S1, S2, ..., Sn. The lane with the smallest deviation is the lane where the accident occurred.

6. The high-speed information acquisition, analysis and processing system based on machine vision according to claim 1, characterized in that: The electronic sign illuminates a message indicating "No Entry".

Images