A three-dimensional black smoke vehicle intelligent capture system

Abstract

The invention relates to capturing a smoky vehicle, and in particular to a three-dimensional intelligent capturing system for a smoky vehicle, comprising a server and an image acquisition module arranged above each driving lane, wherein the server extracts a target vehicle image containing a target vehicle from a driving image acquired by the image acquisition module through a target vehicle extraction module, and extracts continuous frame images of the same target vehicle from the target vehicle image in chronological order through a continuous frame image extraction module, the server marks a vehicle contour line and an exhaust contour line in the continuous frame image through an image processing module, the server analyzes an internal area of ​​the exhaust contour line in the continuous frame image through an image analysis module, and determines the exhaust blackness of the target vehicle based on the analysis result of the internal area of ​​the exhaust contour line using an exhaust blackness evaluation module; the technical solution provided by the invention can effectively overcome the defect of low accuracy of capturing a smoky vehicle in the prior art.

Description

Technical Field

[0001] The invention relates to capturing a smoky vehicle, and in particular to a three-dimensional intelligent capturing system for a smoky vehicle. Background Art

[0002] In recent years, with the continuous improvement of road monitoring and video quality, exhaust gas detection systems based on video monitoring have begun to appear. Based on traditional video monitoring, they use digital image processing technology to extract the tail features of vehicles in video data to detect black smoke vehicles. However, this exhaust gas detection system is easily affected by factors such as environment, weather, and lighting, resulting in a large number of false alarms or missed alarms. Summary of the invention

[0003] 1. Technical issues to be resolved

[0004] In view of the above-mentioned shortcomings of the prior art, the present invention provides a three-dimensional black smoke vehicle intelligent capture system, which can effectively overcome the defect of low black smoke vehicle capture accuracy in the prior art.

[0005] (II) Technical solution

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0007] A three-dimensional black smoke vehicle intelligent capture system includes a server and an image acquisition module arranged above each lane, wherein the server extracts a target vehicle image containing a target vehicle from a driving image acquired by the image acquisition module through a target vehicle extraction module, and extracts continuous frame images of the same target vehicle from the target vehicle image in chronological order through a continuous frame image extraction module;

[0008] The server marks the vehicle contour line and the exhaust contour line in the continuous frame image through the image processing module, analyzes the internal area of ​​the exhaust contour line in the continuous frame image through the image analysis module, and determines the exhaust blackness of the target vehicle based on the analysis result of the internal area of ​​the exhaust contour line by the exhaust blackness evaluation module;

[0009] The server extracts static features corresponding to the inner area of ​​the exhaust contour line from the continuous frame images through the static feature extraction module, and uses the black smoke vehicle recognition module to determine whether the target vehicle is a black smoke vehicle based on the static features of the continuous frame images combined with the exhaust blackness of the target vehicle. The server intercepts the driving video of the black smoke vehicle driving in the lane through the black smoke vehicle video interception module.

[0010] Preferably, the target vehicle extraction module extracts a target vehicle image containing the target vehicle from the driving image acquired by the image acquisition module, comprising:

[0011] In the driving image, moving targets whose circumscribed rectangular area is smaller than a set threshold are removed, and moving targets whose aspect ratio is outside a set range are removed, so as to obtain a target vehicle image containing the target vehicle.

[0012] Preferably, the continuous frame image extraction module extracts continuous frame images of the same target vehicle from the target vehicle image in chronological order, including:

[0013] Obtain the Euclidean distance between the center coordinates of each target vehicle in the next frame of the target vehicle image and the center coordinates of each target vehicle in the previous frame of the target vehicle image;

[0014] The two target vehicles with the smallest Euclidean distance in two consecutive target vehicle images and whose height in the latter target vehicle image is smaller than that in the previous target vehicle image are determined to be the same target vehicle;

[0015] Continuous frame images of the same target vehicle are extracted accordingly.

[0016] Preferably, the image processing module marks the vehicle contour line and the exhaust contour line in the continuous frame image, including:

[0017] Performing image analysis on each single frame image in the continuous frame image to obtain vehicle contour points used to represent the vehicle contour and exhaust contour points used to represent the exhaust smoke contour in the single frame image;

[0018] In each single frame image, the vehicle contour points and the exhaust gas contour points are connected in sequence to complete the marking of the vehicle contour line and the exhaust gas contour line in the continuous frame image.

[0019] Preferably, the image analysis module analyzes the internal area of ​​the exhaust gas contour line in the continuous frame image, including:

[0020] Performing image analysis on the internal area of ​​the exhaust gas contour line of each single frame image in the continuous frame image, and marking the single frame image in which the exhaust gas smoke color depth in the internal area of ​​the exhaust gas contour line is greater than a set threshold;

[0021] The unmarked single-frame images in the continuous frame images are deleted, and the marked single-frame images representing the exhaust smoke in the inner area of ​​the exhaust contour line as black are retained.

[0022] Preferably, the exhaust gas blackness evaluation module determines the exhaust gas blackness of the target vehicle based on the exhaust gas contour line internal area analysis result, including:

[0023] The marked single-frame image in the continuous frame image sent by the image analysis module is received, and the exhaust blackness of the target vehicle is determined based on the marked single-frame image to obtain the exhaust blackness evaluation of the single target vehicle.

[0024] Preferably, the static feature extraction module extracts static features corresponding to the inner area of ​​the exhaust contour line from the continuous frame images, including:

[0025] A convolutional neural network is used to extract static features from the internal area of ​​the exhaust contour line of each single frame image in the continuous frame image;

[0026] The convolutional neural network is trained using the ImageNet dataset, and the 1000-class fully connected layer classifier of the training model is removed.

[0027] Preferably, the smoky vehicle identification module determines whether the target vehicle is a smoky vehicle based on the static features of the continuous frame images combined with the blackness of the target vehicle's exhaust gas, including:

[0028] All static features in the continuous frame images are input into the long short-term memory network in chronological order to obtain the hidden state of the long short-term memory network at each time point;

[0029] Based on the attention mechanism, the attention weight corresponding to the hidden state at each time point is obtained, and the hidden state at each time point is weighted averaged to obtain the background vector;

[0030] The background vector and the target vehicle exhaust blackness are input into the fully connected layer, and based on the output result of the fully connected layer, it is determined whether the target vehicle is a black smoke vehicle.

[0031] Preferably, it also includes a black smoke vehicle video uploading module connected to the server for uploading the driving video to the monitoring center, and the black smoke vehicle video interception module intercepts the driving video of the black smoke vehicle driving in the lane, including:

[0032] For the target vehicle judged as a black smoke vehicle, a driving video including the target vehicle driving in the lane and the license plate of the target vehicle is intercepted from the driving image collected by the image acquisition module, and the driving video is uploaded to the monitoring center through the black smoke vehicle video upload module.

[0033] (III) Beneficial effects

[0034] Compared with the prior art, the three-dimensional black smoke vehicle intelligent capture system provided by the present invention has the following beneficial effects:

[0035] 1) The target vehicle extraction module extracts the target vehicle image containing the target vehicle from the driving image acquired by the image acquisition module, and the continuous frame image extraction module extracts the continuous frame images of the same target vehicle from the target vehicle image in chronological order, providing data support for the subsequent accurate determination of the exhaust blackness of the target vehicle and for determining whether the target vehicle is a black smoke vehicle in combination with the static features of the continuous frame images;

[0036] 2) The image processing module marks the vehicle contour line and the exhaust contour line in the continuous frame image, the image analysis module analyzes the internal area of ​​the exhaust contour line in the continuous frame image, and the exhaust blackness evaluation module determines the exhaust blackness of the target vehicle based on the analysis result of the internal area of ​​the exhaust contour line, so that the exhaust blackness of the target vehicle can be accurately determined based on the continuous frame images of the same target vehicle;

[0037] 3) The static feature extraction module extracts the static features corresponding to the inner area of ​​the exhaust contour line from the continuous frame images. The black smoke vehicle recognition module determines whether the target vehicle is a black smoke vehicle based on the static features of the continuous frame images combined with the exhaust blackness of the target vehicle. The static features of the continuous frame images of the same target vehicle and the exhaust blackness of the target vehicle are used to make a comprehensive judgment on the target vehicle, which can effectively improve the accuracy of capturing black smoke vehicles. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0039] Figure 1 It is a schematic diagram of the system of the present invention;

[0040] Figure 2 It is a schematic diagram of the process of the present invention;

[0041] Figure 3 It is a schematic diagram of a process of judging whether a target vehicle is a black smoke vehicle based on the static features of continuous frame images combined with the blackness of the target vehicle's exhaust gas in the present invention. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0043] A three-dimensional black smoke vehicle intelligent capture system, such as Figure 1 and Figure 2As shown, it includes a server and an image acquisition module arranged above each lane. The server extracts a target vehicle image containing a target vehicle from the driving image acquired by the image acquisition module through a target vehicle extraction module, and uses a continuous frame image extraction module to extract continuous frame images of the same target vehicle from the target vehicle image in chronological order.

[0044] ① The target vehicle extraction module extracts the target vehicle image containing the target vehicle from the driving image acquired by the image acquisition module, including:

[0045] In the driving image, moving targets whose circumscribed rectangular area is smaller than a set threshold are removed, and moving targets whose aspect ratio is outside a set range are removed, so as to obtain a target vehicle image containing the target vehicle.

[0046] ② The continuous frame image extraction module extracts continuous frame images of the same target vehicle from the target vehicle image in chronological order, including:

[0047] Obtain the Euclidean distance between the center coordinates of each target vehicle in the next frame of the target vehicle image and the center coordinates of each target vehicle in the previous frame of the target vehicle image;

[0048] The two target vehicles with the smallest Euclidean distance in two consecutive target vehicle images and whose height in the latter target vehicle image is smaller than that in the previous target vehicle image are determined to be the same target vehicle;

[0049] Continuous frame images of the same target vehicle are extracted accordingly.

[0050] In the above technical solution, the target vehicle extraction module extracts the target vehicle image containing the target vehicle from the driving image acquired by the image acquisition module, and the continuous frame image extraction module extracts the continuous frame images of the same target vehicle from the target vehicle image in chronological order, so as to provide data support for the subsequent accurate determination of the exhaust blackness of the target vehicle, and to determine whether the target vehicle is a black smoke vehicle in combination with the static features of the continuous frame images.

[0051] like Figure 1 and Figure 2 As shown, the server marks the vehicle contour line and the exhaust contour line in the continuous frame image through the image processing module, analyzes the internal area of ​​the exhaust contour line in the continuous frame image through the image analysis module, and uses the exhaust blackness evaluation module to determine the exhaust blackness of the target vehicle based on the analysis results of the internal area of ​​the exhaust contour line.

[0052] ① The image processing module marks the vehicle contour line and exhaust contour line in the continuous frame image, including:

[0053] Performing image analysis on each single frame image in the continuous frame image to obtain vehicle contour points used to represent the vehicle contour and exhaust contour points used to represent the exhaust smoke contour in the single frame image;

[0054] In each single frame image, the vehicle contour points and the exhaust gas contour points are connected in sequence to complete the marking of the vehicle contour line and the exhaust gas contour line in the continuous frame image.

[0055] ②The image analysis module analyzes the internal area of ​​the exhaust contour line in the continuous frame image, including:

[0056] Performing image analysis on the internal area of ​​the exhaust gas contour line of each single frame image in the continuous frame image, and marking the single frame image in which the exhaust gas smoke color depth in the internal area of ​​the exhaust gas contour line is greater than a set threshold;

[0057] The unmarked single-frame images in the continuous frame images are deleted, and the marked single-frame images representing the exhaust smoke in the inner area of ​​the exhaust contour line as black are retained.

[0058] ③ The exhaust gas blackness evaluation module determines the exhaust gas blackness of the target vehicle based on the analysis results of the internal area of ​​the exhaust gas contour line, including:

[0059] The marked single-frame image in the continuous frame image sent by the image analysis module is received, and the exhaust blackness of the target vehicle is determined based on the marked single-frame image to obtain the exhaust blackness evaluation of the single target vehicle.

[0060] In the above technical solution, the image processing module marks the vehicle contour line and the exhaust contour line in the continuous frame image, the image analysis module analyzes the internal area of ​​the exhaust contour line in the continuous frame image, and the exhaust blackness evaluation module determines the exhaust blackness of the target vehicle based on the analysis result of the internal area of ​​the exhaust contour line, thereby being able to accurately determine the exhaust blackness of the target vehicle based on the continuous frame images of the same target vehicle.

[0061] like Figures 1 to 3 As shown, the server extracts static features corresponding to the inner area of ​​the exhaust contour line from the continuous frame images through the static feature extraction module, and uses the black smoke vehicle recognition module to determine whether the target vehicle is a black smoke vehicle based on the static features of the continuous frame images combined with the exhaust blackness of the target vehicle. The server intercepts the driving video of the black smoke vehicle driving in the driving lane through the black smoke vehicle video capture module.

[0062] ① The static feature extraction module extracts the static features corresponding to the internal area of ​​the exhaust contour line from the continuous frame images, including:

[0063] A convolutional neural network is used to extract static features from the internal area of ​​the exhaust contour line of each single frame image in the continuous frame image;

[0064] The convolutional neural network is trained using the ImageNet dataset, and the 1000-class fully connected layer classifier of the training model is removed.

[0065] ② The black smoke vehicle identification module determines whether the target vehicle is a black smoke vehicle based on the static features of the continuous frame images and the blackness of the target vehicle's exhaust gas, including:

[0066] All static features in the continuous frame images are input into the long short-term memory network in chronological order to obtain the hidden state of the long short-term memory network at each time point;

[0067] Based on the attention mechanism, the attention weight corresponding to the hidden state at each time point is obtained, and the hidden state at each time point is weighted averaged to obtain the background vector;

[0068] The background vector and the target vehicle exhaust blackness are input into the fully connected layer, and based on the output result of the fully connected layer, it is determined whether the target vehicle is a black smoke vehicle.

[0069] The technical solution of the present application also includes a black smoke vehicle video uploading module connected to the server for uploading the driving video to the monitoring center, and a black smoke vehicle video interception module intercepts the driving video of the black smoke vehicle driving in the driving lane, including:

[0070] For the target vehicle judged as a black smoke vehicle, a driving video including the target vehicle driving in the lane and the license plate of the target vehicle is intercepted from the driving image collected by the image acquisition module, and the driving video is uploaded to the monitoring center through the black smoke vehicle video upload module.

[0071] In the above technical solution, the static feature extraction module extracts the static features corresponding to the inner area of ​​the exhaust contour line from the continuous frame images, and the black smoke vehicle recognition module determines whether the target vehicle is a black smoke vehicle based on the static features of the continuous frame images combined with the exhaust blackness of the target vehicle. The static features of the continuous frame images of the same target vehicle and the exhaust blackness of the target vehicle are used to make a comprehensive judgment on the target vehicle, which can effectively improve the accuracy of capturing black smoke vehicles.

[0072] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A three-dimensional black smoke vehicle intelligent capture system, characterized by: The server comprises a server and an image acquisition module arranged above each lane, wherein the server extracts a target vehicle image containing a target vehicle from a driving image acquired by the image acquisition module through a target vehicle extraction module, and extracts continuous frame images of the same target vehicle from the target vehicle image in chronological order through a continuous frame image extraction module; The server marks the vehicle contour line and the exhaust contour line in the continuous frame image through the image processing module, analyzes the internal area of ​​the exhaust contour line in the continuous frame image through the image analysis module, and determines the exhaust blackness of the target vehicle based on the analysis result of the internal area of ​​the exhaust contour line by the exhaust blackness evaluation module; The server extracts static features corresponding to the inner area of ​​the exhaust contour line from the continuous frame images through the static feature extraction module, and uses the black smoke vehicle identification module to determine whether the target vehicle is a black smoke vehicle based on the static features of the continuous frame images combined with the blackness of the target vehicle's exhaust gas. The server intercepts the driving video of the black smoke vehicle driving in the lane through the black smoke vehicle video interception module; The target vehicle extraction module extracts a target vehicle image containing the target vehicle from the driving image acquired by the image acquisition module, including: In the driving image, moving targets whose circumscribed rectangular area is smaller than a set threshold are removed, and moving targets whose aspect ratio is outside a set range are removed, so as to obtain a target vehicle image containing the target vehicle; The continuous frame image extraction module extracts continuous frame images of the same target vehicle from the target vehicle image in chronological order, including: Obtain the Euclidean distance between the center coordinates of each target vehicle in the next frame of the target vehicle image and the center coordinates of each target vehicle in the previous frame of the target vehicle image; The two target vehicles with the smallest Euclidean distance in two consecutive target vehicle images and whose height in the latter target vehicle image is smaller than that in the previous target vehicle image are determined to be the same target vehicle; Correspondingly extracting continuous frame images of the same target vehicle; The image processing module marks the vehicle contour line and the exhaust contour line in the continuous frame image, including: Performing image analysis on each single frame image in the continuous frame image to obtain vehicle contour points used to represent the vehicle contour and exhaust contour points used to represent the exhaust smoke contour in the single frame image; Connect the vehicle contour points and exhaust gas contour points in order in each single frame image, and complete the marking of the vehicle contour line and exhaust gas contour line in the continuous frame image; The image analysis module analyzes the internal area of ​​the exhaust gas contour line in the continuous frame image, including: Performing image analysis on the internal area of ​​the exhaust gas contour line of each single frame image in the continuous frame image, and marking the single frame image in which the exhaust gas smoke color depth in the internal area of ​​the exhaust gas contour line is greater than a set threshold; The unmarked single-frame images in the continuous frame images are deleted, and the marked single-frame images representing the exhaust smoke in the inner area of ​​the exhaust contour line as black are retained.

2. The three-dimensional black smoke vehicle intelligent capture system according to claim 1 is characterized by: The exhaust gas blackness evaluation module determines the exhaust gas blackness of the target vehicle based on the exhaust gas contour line internal area analysis result, including: The marked single-frame image in the continuous frame image sent by the image analysis module is received, and the exhaust blackness of the target vehicle is determined based on the marked single-frame image to obtain the exhaust blackness evaluation of the single target vehicle.

3. The three-dimensional black smoke vehicle intelligent capture system according to claim 1 is characterized by: The static feature extraction module extracts static features corresponding to the inner area of ​​the exhaust contour line from the continuous frame images, including: A convolutional neural network is used to extract static features from the internal area of ​​the exhaust contour line of each single frame image in the continuous frame image; The convolutional neural network is trained using the ImageNet dataset, and the 1000-class fully connected layer classifier of the training model is removed.

4. The three-dimensional black smoke vehicle intelligent capture system according to claim 3 is characterized by: The black smoke vehicle identification module determines whether the target vehicle is a black smoke vehicle based on the static features of the continuous frame images combined with the blackness of the target vehicle's exhaust gas, including: All static features in the continuous frame images are input into the long short-term memory network in chronological order to obtain the hidden state of the long short-term memory network at each time point; Based on the attention mechanism, the attention weight corresponding to the hidden state at each time point is obtained, and the hidden state at each time point is weighted averaged to obtain the background vector; The background vector and the target vehicle exhaust blackness are input into the fully connected layer, and based on the output result of the fully connected layer, it is determined whether the target vehicle is a black smoke vehicle.

5. The three-dimensional black smoke vehicle intelligent capture system according to claim 4 is characterized in that: It also includes a black smoke vehicle video uploading module connected to the server for uploading the driving video to the monitoring center, and the black smoke vehicle video interception module intercepts the driving video of the black smoke vehicle driving on the driving lane, including: For the target vehicle judged as a black smoke vehicle, a driving video including the target vehicle driving in the lane and the license plate of the target vehicle is intercepted from the driving image collected by the image acquisition module, and the driving video is uploaded to the monitoring center through the black smoke vehicle video upload module.

Images