Illegal occupation system for fire fighting access

Abstract

The invention belongs to the technical field of fire-fighting access, and particularly relates to a fire-fighting access illegal occupation system, which comprises a fire-fighting access image shooting module, a fire-fighting access information acquisition module, a fire-fighting access information processing module, a fire-fighting access information processing module, a fire-fighting access information processing module and a fire-fighting access information processing module, and is characterized in that the fire-fighting access image shooting module is used for shooting fire-fighting access information; the occupied object distinguishing module is used for distinguishing occupied objects, and distinguishing types comprise a vehicle type and a sundry type; the occupancy early warning module is used for judging whether the occupancy object affects the use of the fire fighting access or not; and the vehicle invasion and occupation state analysis module is used for analyzing the state of the vehicle in the fire fighting access, wherein the state comprises an invasion and occupation state and a tailgating state.

Description

Technical Field

[0001] The invention belongs to the technical field of fire passages, in particular to a fire passage illegal occupation system. Background Art

[0002] Fire escapes refer to spaces and passages inside or outside a building that are specifically provided for firefighters to carry out firefighting and rescue and evacuate people when a fire occurs. They are "life passages" when a fire occurs and are crucial to ensuring the safety of personnel and quickly extinguishing fires.

[0003] However, if the fire passage is occupied by vehicles or other difficult-to-move objects, it may delay fire rescue, causing greater economic losses and casualties. At present, most units only set up warning signs in front of the fire passage, and generally rely on people to comply voluntarily. However, in most cases, people are lucky enough to park their vehicles in the fire passage, and the guards often fail to find them in time, which can easily cause safety accidents.

[0004] The current monitoring system is difficult to accurately identify objects occupying fire passages, and may generate false alarms when vehicles only stay in the fire passage for a period of time and do not actually occupy the fire passage.

[0005] To this end, the present invention provides a fire passage illegal occupation system. Summary of the invention

[0006] In order to make up for the deficiencies of the prior art, at least one technical problem raised in the background technology is solved.

[0007] The technical solution adopted by the present invention to solve the technical problem is: a fire passage illegal occupation system described in the present invention comprises a fire passage image shooting module: the fire passage image shooting module is used to shoot fire passage information in real time, and the fire passage information includes whether there is an occupied object in the fire passage; Occupancy object identification module: The occupation object identification module is used to identify the occupied objects, and the identification types include vehicle type and sundries type; Occupancy warning module: The occupancy warning module is used to determine whether the occupied object affects the use of the fire escape; Vehicle occupation status analysis module: The vehicle occupation status analysis module is used to analyze the status of the vehicle in the fire passage, which includes occupation status and borrowing status; Specifically: if the occupied object classification module classifies the occupied object as a vehicle, and the current state of the vehicle is an occupied state; The area of ​​fire passage occupied by vehicles is divided into risk levels; The risk levels include primary risk and secondary risk: Get the occupied area of ​​the vehicle in the fire channel image , pre-set the maximum threshold for vehicles occupying the fire passage area ; like , then the vehicle encroaching on the fire passage area is judged as a first-level risk and marked as a first-level risk that does not affect the passage of fire trucks; like , then the vehicle encroaching on the fire passage area is judged as a second-level risk, and marked as a second-level risk that seriously affects the passage of fire trucks; Based on the first-level risk, the fire passage image capture module captures the moment when the vehicle first stops at the fire passage. , and from the moment Start recording, pre-set the maximum vehicle dwell time threshold to If the detected vehicle stay time is greater than the maximum stay time , then a signal is generated and transmitted to the occupancy time warning module, the vehicle license plate information is extracted through the fire channel image shooting module, the license plate information is reported to the traffic management department, the vehicle information database is queried, the owner information is queried, and the owner is notified by SMS; Based on the second level risk, a signal is generated and transmitted to the occupancy warning module, and an alarm is directly issued.

[0008] Preferably, the occupied object is identified based on the occupied object identification module as follows: Establish a convolutional neural network (CNN) model, collect a large amount of fire passage image data, the fire passage image data includes various possible occupancy conditions in the fire passage, and preprocess the image data, the preprocessing includes image scaling, normalization and data enhancement; Annotate objects such as vehicles at different locations, objects of different shapes, and objects at different locations in the image data, including bounding boxes and categories; Use labeled image data to train the model so that it can learn how to recognize and classify different objects. The training model learns to distinguish the features of objects such as vehicles in different positions and debris of different shapes through training; The trained model is deployed to the fire passage image shooting module, and the fire passage image shooting module is used to capture the channel information of the fire passage in real time.

[0009] Preferably, the model includes YOLO and Faster R-CNN models; When the fire passage image capturing module captures information about an occupied object in the fire passage; The YOLO and Faster R-CNN models are trained to learn the features of objects such as vehicles in different positions and debris of different shapes, and can directly determine whether the occupied object information is a vehicle or debris; If the occupied object is a vehicle, the occupancy warning module will be used to determine whether the vehicle affects the use of the fire passage; If the occupied object is debris, a signal is generated and transmitted to the occupancy warning module, reminding the staff to come and deal with it.

[0010] Preferably, the occupied area of ​​the vehicle in the fire passage image The specific calculation method is: First, the occupancy information is determined as the vehicle category through the YOLO and Faster R-CNN models; And according to the bounding box of the vehicle in the image data, get four coordinate values, the upper left corner x and y coordinates and the x and y coordinates of the lower right corner ; Calculate the occupied area .

[0011] Preferably, the vehicle occupation state analysis module analyzes whether the vehicle is in an occupation state or a borrowing state specifically as follows: Analyze the lighting status of the vehicle through the vehicle information captured by the fire channel shooting module, where the lighting status includes whether the headlights or taillights are on; If the headlights or taillights are on, a signal is generated and transmitted to the vehicle encroachment status analysis module, and the vehicle is determined to be in a state of borrowing lanes; Otherwise it is in occupied state.

[0012] Preferably, the fire passage image capturing module is also used to process images in bad weather or poor lighting conditions; Specifically: image enhancement techniques, such as histogram equalization, contrast enhancement, gamma correction, etc., to improve the visibility of the image. For example, histogram equalization can adjust the contrast of the image to make the image visually clearer; Filter denoising, applying filtering algorithms, such as Gaussian filtering, median filtering or bilateral filtering, to reduce noise in the image. These filters can help improve image quality, especially in low light conditions.

[0013] Preferably, the vehicle encroachment status analysis module is also used to analyze the vehicle's lane borrowing status; If the vehicle occupation status analysis module determines that the vehicle is in a state of borrowing the lane, and the fire passage shooting module detects that the vehicle is not turned off in the fire passage, the YOLO and Faster R-CNN models are used to determine whether the driver has left: If the driver leaves the vehicle, and the time the driver has left the vehicle is less than the maximum time threshold, the vehicle is determined to have borrowed the lane for a short period of time; If the driver leaves the vehicle, and the time the driver has been away from the vehicle is greater than the maximum time threshold, it is determined that the vehicle has been borrowing the lane for a long time, and a signal is generated and transmitted to the occupancy warning module to issue a warning message to the owner.

[0014] Preferably, based on the fact that the vehicle is not turned off in the fire passage, the specific calculation method of the driver's position from the vehicle is: When the fire channel camera module detects the driver's departure time, it is marked as , and from the departure time Start timing, and pre-set the maximum time threshold for leaving , and will be from The time of timing and Compare.

[0015] Preferably, the vehicle license plate information is extracted based on the fire passage image shooting module as follows: Use image processing technology to determine the potential license plate area in the image, and use thresholding, edge detection, and morphological operations to achieve the shape and position of the license plate; The characters on the license plate are segmented into separate images by horizontal projection histogram and other methods, and features are extracted from the segmented character images. Commonly used features include Huff transform features, shape features, histogram features, etc. Use the labeled license plate character dataset to train the SVM model. For the province abbreviation and alphanumeric combination, create two datasets and train two models respectively; The extracted features are input into the trained SVM model, and the model will output the predicted category of the character; The recognized characters are stitched together to form a complete license plate number, and the complete license plate number is reported to the traffic control department.

[0016] Preferably, the fire passage image capturing module extracting vehicle license plate information further includes: If the complete license plate number is not extracted, or the location of the license plate is not identified through image processing technology, a signal is generated and transmitted to the occupancy warning module, and the occupancy warning module plays a warning message, which includes a horn broadcast or towing away with a trailer.

[0017] The beneficial effects of the present invention are as follows: 1. A fire passage illegal occupation system described in the present invention can timely determine the occupation status of vehicles and other debris at the fire passage through the fire passage image shooting module, the occupied object identification module, the occupancy warning module and the vehicle occupation status analysis module. Various warning information can be selected according to whether the vehicle is borrowing the passage or encroaching, and the occupation status of the fire passage can be accurately monitored. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The present invention will be further described below in conjunction with the accompanying drawings.

[0019] Figure 1 It is a system diagram of the present invention. DETAILED DESCRIPTION

[0020] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods.

[0021] like Figure 1 As shown, a fire passage illegal occupation system according to an embodiment of the present invention includes a fire passage image shooting module, an occupied object identification module, an occupancy warning module and a vehicle occupation status analysis module; Fire passage image shooting module: the fire passage image shooting module is used to shoot fire passage information in real time, the fire passage information includes whether there are occupied objects in the fire passage; Occupancy object identification module: The occupation object identification module is used to identify the occupied objects, and the identification types include vehicle type and sundries type; Occupancy warning module: The occupancy warning module is used to determine whether the occupied object affects the use of the fire escape; Vehicle occupation status analysis module: The vehicle occupation status analysis module is used to analyze the status of the vehicle in the fire passage, which includes occupation status and borrowing status; Specifically: if the occupied object classification module classifies the occupied object as a vehicle, and the current state of the vehicle is an occupied state; The area of ​​fire passage occupied by vehicles is divided into risk levels; The risk levels include primary risk and secondary risk: Get the occupied area of ​​the vehicle in the fire channel image , pre-set the maximum threshold for vehicles occupying the fire passage area ; like , then the vehicle encroaching on the fire passage area is judged as a first-level risk and marked as a first-level risk that does not affect the passage of fire trucks; like , then the vehicle encroaching on the fire passage area is judged as a second-level risk, and marked as a second-level risk that seriously affects the passage of fire trucks; Based on the first-level risk, the fire passage image capture module captures the moment when the vehicle first stops at the fire passage. , and from the moment Start recording, pre-set the maximum vehicle dwell time threshold to If the detected vehicle stay time is greater than the maximum stay time , then a signal is generated and transmitted to the occupancy time warning module, the vehicle license plate information is extracted through the fire channel image shooting module, the license plate information is reported to the traffic management department, the vehicle information database is queried, the owner information is queried, and the owner is notified by SMS; Based on the second level risk, a signal is generated and transmitted to the occupancy warning module, and an alarm is directly issued.

[0022] For example, if the stay time is set to 10 minutes, when a vehicle stays in the fire passage for longer than this time, an early warning message will be generated and illegal images will be extracted. If it does not exceed 10 minutes, it will be automatically filtered and no early warning message will be generated. Image preprocessing module: The image preprocessing module is used to grayscale, filter and remove noise on the image of the fire channel image shooting module to improve the subsequent image segmentation effect; Image processing module: The image processing module is used to perform standardization processing on the captured images; Identify occupied objects in video images and determine whether the occupied objects meet the warning rules. If they do, generate warning information and hidden danger images, and upload the warning information and hidden danger images to the monitoring center platform; Determine the vehicle's residence time and pre-set the maximum residence time threshold as , if the detected vehicle stay time is greater than the maximum stay time, a signal is generated and transmitted to , and the vehicle license plate information is extracted, and the information is transmitted to the monitoring center platform; Occupancy object recognition includes category recognition. Different warning information is used according to the category of the object. The category of the occupied object can be a vehicle or miscellaneous items.

[0023] The occupied object identification module is used to identify the occupied object as follows: Establish a convolutional neural network (CNN) model, collect a large amount of fire passage image data, the fire passage image data includes various possible occupancy conditions in the fire passage, and preprocess the image data, the preprocessing includes image scaling, normalization and data enhancement; Annotate objects such as vehicles at different locations, objects of different shapes, and objects at different locations in the image data, including bounding boxes and categories; Use labeled image data to train the model so that it can learn how to recognize and classify different objects. The training model learns to distinguish the features of objects such as vehicles in different positions and debris of different shapes through training; Deploy the trained model to the fire passage image shooting module, and use the fire passage image shooting module to capture the passage information of the fire passage in real time; Convolutional neural network (CNN) collects a large amount of image data, including various possible occupancy conditions in the fire channel. Data preprocessing includes image scaling, normalization, and data enhancement. Image segmentation, uses image segmentation algorithms such as K-means clustering, watershed algorithm, edge detection, etc. to separate the vehicle in the image from the background.

[0024] The models include YOLO and Faster R-CNN models; When the fire passage image capturing module captures information about an occupied object in the fire passage; The YOLO and Faster R-CNN models are trained to learn the features of objects such as vehicles in different positions and debris of different shapes, and can directly determine whether the occupied object information is a vehicle or debris; If the occupied object is a vehicle, the occupancy warning module will be used to determine whether the vehicle affects the use of the fire passage; If the occupied object is debris, a signal is generated and transmitted to the occupancy warning module, reminding the staff to come and deal with it.

[0025] YOLO is a fast object detection system that treats object detection as a regression problem. It divides the image into grids, and each grid is responsible for predicting the bounding box and category probability of the object in the area. The specific operation process is: divide the input image into SxS grids, predict B bounding boxes and C conditional probabilities of each grid, each bounding box contains 5 parameters, and each bounding box also predicts a conditional probability, indicating the probability of belonging to each category.

[0026] Faster R-CNN uses a region proposal network to quickly and efficiently generate candidate regions, and then uses FastR-CNN to classify and regress bounding boxes for these regions; The specific workflow is as follows: Feature extraction: Use pre-trained CNN to extract image features; Region Proposal Network: RPN slides the window on the feature map to generate candidate regions for each position; ROI pooling: Pool the candidate regions generated by RPN to obtain a feature map of fixed size; Classification and regression: Classify and regress the pooled feature map; Convolutional neural network (CNN) is a deep learning model that is mainly used to process data with obvious grid-like topological structures. The core feature of CNN is that it can automatically and layer by layer extract local features of images through convolutional layers; convolutional layer: by sliding the convolution kernel (or filter) on the input image, calculating the dot product between the convolution kernel and the local area of ​​the image, generating a feature map, it can extract features such as edges and textures in the image, and then use nonlinear activation functions such as ReLU to increase the nonlinear expression ability of the network.

[0027] The occupied area of ​​the vehicle in the fire passage image The specific calculation method is: First, the occupancy information is determined as the vehicle category through the YOLO and Faster R-CNN models; And according to the bounding box of the vehicle in the image data, get four coordinate values, the upper left corner x and y coordinates and the x and y coordinates of the lower right corner ; Calculate the occupied area .

[0028] The width of the fire passage, the net width and net height of the fire truck passage should not be less than 4 meters to ensure that fire trucks can pass smoothly; the turning radius of the fire passage must meet the turning requirements of fire trucks; If a detected vehicle occupies a fire lane, a high-priority warning will be issued, as this will seriously hinder the passage of fire trucks; if debris is detected, the system will issue different levels of warnings depending on its size and location; To calculate the area of ​​the fire escape occupied by an occupied object, first determine the bounding box of the occupied object in the image, which is usually provided by the object detection algorithm YOLO or Faster R-CNN. The bounding box is usually represented by four coordinate values: If the bounding box gives the coordinates of the center of the object and the width and height of the object, then the area calculation formula is the product of the width and height.

[0029] The vehicle occupation state analysis module analyzes whether the vehicle is in an occupation state or a borrowing state specifically as follows: Analyze the lighting status of the vehicle through the vehicle information captured by the fire channel shooting module, where the lighting status includes whether the headlights or taillights are on; If the headlights or taillights are on, a signal is generated and transmitted to the vehicle encroachment status analysis module, and the vehicle is determined to be in a state of borrowing lanes; Otherwise it is in occupied state.

[0030] Specifically, if a vehicle encounters a traffic jam during driving, and the traffic jam is located at a fire passage, the vehicle information captured by the fire passage shooting module is used to analyze whether the vehicle lights are on. If the vehicle lights are on, it indicates that the vehicle is still in driving or temporarily stopped, that is, the engine is not turned off. At this time, it can be preliminarily determined that the vehicle is in a state of borrowing the road; if the headlights or taillights are not lit, it indicates that the vehicle in the fire passage is in a state of being turned off, and it can be determined as an encroachment state. After that, the above steps can be used to monitor whether the vehicle illegally occupies the fire passage.

[0031] The fire passage image capturing module is also used to process images in bad weather or poor lighting conditions; Specifically: image enhancement techniques, such as histogram equalization, contrast enhancement, gamma correction, etc., to improve the visibility of the image. For example, histogram equalization can adjust the contrast of the image to make the image visually clearer; Filter denoising, applying filtering algorithms, such as Gaussian filtering, median filtering or bilateral filtering, to reduce noise in the image. These filters can help improve image quality, especially in low light conditions; Deep learning models, such as convolutional neural networks (CNNs), can extract useful features from images of severe weather conditions, such as feature extraction and image enhancement; Multi-scale processing is used in object detection algorithms to detect objects at different sizes and resolutions.

[0032] The vehicle encroachment status analysis module is also used to analyze the vehicle's lane borrowing status; If the vehicle occupation status analysis module determines that the vehicle is in a state of borrowing the lane, and the fire passage shooting module detects that the vehicle is not turned off in the fire passage, the YOLO and Faster R-CNN models are used to determine whether the driver has left: If the driver leaves the vehicle, and the time the driver has left the vehicle is less than the maximum time threshold, the vehicle is determined to have borrowed the lane for a short period of time; If the driver leaves the vehicle, and the time the driver has been away from the vehicle is greater than the maximum time threshold, it is determined that the vehicle has been borrowing the lane for a long time, and a signal is generated and transmitted to the occupancy warning module to issue a warning message to the owner.

[0033] Based on the fact that the vehicle is not turned off in the fire passage, the specific calculation method of the driver's position from the vehicle is: When the fire channel camera module detects the driver's departure time, it is marked as , and from the departure time Start timing, and pre-set the maximum time threshold for leaving , and will be from The time of the timing and Compare.

[0034] The specific information of vehicle license plate extracted based on the fire channel image shooting module is as follows: Use image processing technology to determine the potential license plate area in the image, and use thresholding, edge detection, and morphological operations to achieve the shape and position of the license plate; The characters on the license plate are segmented into separate images by horizontal projection histogram and other methods, and features are extracted from the segmented character images. Commonly used features include Huff transform features, shape features, histogram features, etc. Use the labeled license plate character dataset to train the SVM model. For the province abbreviation and alphanumeric combination, create two datasets and train two models respectively; The extracted features are input into the trained SVM model, and the model will output the predicted category of the character; The recognized characters are stitched together to form a complete license plate number, and the complete license plate number is reported to the traffic control department.

[0035] SVM is a supervised learning algorithm that finds an optimal segmentation hyperplane in the feature space so that the interval between sample points of different categories is maximized. For license plate recognition, two different SVM models can be trained to solve two types of problems: one model is used to recognize the abbreviation of the province, and the other model is used to recognize the combination of letters and numbers.

[0036] The fire passage image capturing module extracts vehicle license plate information and further comprises: If the complete license plate number is not extracted, or the location of the license plate is not identified through image processing technology, a signal is generated and transmitted to the occupancy warning module, and the occupancy warning module plays a warning message, which includes a horn broadcast or towing away with a trailer.

[0037] In summary, the system can timely determine the occupation status of vehicles and other debris in the fire passage, can select various warning information according to whether the vehicle is borrowing the passage or encroaching, and can accurately monitor the occupation status of the fire passage.

[0038] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A fire passage illegal occupation system, characterized by: include Fire passage image shooting module: the fire passage image shooting module is used to shoot fire passage information in real time, the fire passage information includes whether there are occupied objects in the fire passage; Occupancy object identification module: The occupation object identification module is used to identify the occupied objects, and the identification types include vehicle type and sundries type; Occupancy warning module: The occupancy warning module is used to determine whether the occupied object affects the use of the fire escape; Vehicle occupation status analysis module: The vehicle occupation status analysis module is used to analyze the status of the vehicle in the fire passage, which includes occupation status and borrowing status; Specifically: if the occupied object classification module classifies the occupied object as a vehicle, and the current state of the vehicle is an occupied state; The area of ​​fire passage occupied by vehicles is divided into risk levels; The level of risk includes level one risk and level two risk: Get the occupied area of ​​the vehicle in the fire channel image , pre-set the maximum threshold for vehicles occupying the fire passage area ; like , then the vehicle encroaching on the fire passage area is judged as a first-level risk and marked as a first-level risk that does not affect the passage of fire trucks; like , then the vehicle encroaching on the fire passage area is judged as a second-level risk, and marked as a second-level risk that seriously affects the passage of fire trucks; Based on the first-level risk, the fire passage image capture module captures the moment when the vehicle first stops at the fire passage. , and from the moment Start recording, pre-set the maximum vehicle dwell time threshold to If the detected vehicle stay time is greater than the maximum stay time , then a signal is generated and transmitted to the occupancy time warning module, the vehicle license plate information is extracted through the fire channel image shooting module, the license plate information is reported to the traffic management department, the vehicle information database is queried, the owner information is queried, and the owner is notified by SMS; Based on the second level risk, a signal is generated and transmitted to the occupancy warning module, and an alarm is directly issued.

2. A fire passage illegal occupation system according to claim 1, characterized in that: The occupied object identification module is used to identify the occupied object as follows: Establish a convolutional neural network (CNN) model and collect a large amount of fire passage image data, which includes various possible occupancy conditions in the fire passage; Annotate objects such as vehicles at different locations, objects of different shapes, and objects at different locations in the image data, including bounding boxes and categories; Use labeled image data to train the model so that it can learn how to recognize and classify different objects. The training model learns to distinguish the features of objects such as vehicles in different positions and debris of different shapes through training; The trained model is deployed to the fire passage image shooting module, and the fire passage image shooting module is used to capture the channel information of the fire passage in real time.

3. A fire passage illegal occupation system according to claim 2, characterized in that: The models include YOLO and Faster R-CNN models; When the fire passage image capturing module captures information about an occupied object in the fire passage; The YOLO and Faster R-CNN models are trained to learn the features of objects such as vehicles in different positions and debris of different shapes, and can directly determine whether the occupied object information is a vehicle or debris; If the occupied object is a vehicle, the occupancy warning module will be used to determine whether the vehicle affects the use of the fire passage; If the occupied object is debris, a signal is generated and transmitted to the occupancy warning module, reminding the staff to come and deal with it.

4. A fire passage illegal occupation system according to claim 3, characterized in that: The occupied area of ​​the vehicle in the fire passage image The specific calculation method is: First, the occupancy information is determined as the vehicle category through the YOLO and Faster R-CNN models; And according to the bounding box of the vehicle in the image data, get four coordinate values, the upper left corner x and y coordinates and the x and y coordinates of the lower right corner ; Calculate the occupied area .

5. A fire passage illegal occupation system according to claim 3, characterized in that: The vehicle occupation state analysis module analyzes whether the vehicle is in an occupation state or a borrowing state specifically as follows: Analyze the lighting status of the vehicle through the vehicle information captured by the fire channel shooting module, where the lighting status includes whether the headlights or taillights are on; If the headlights or taillights are on, a signal is generated and transmitted to the vehicle encroachment status analysis module, and the vehicle is determined to be in a state of borrowing lanes; Otherwise it is in occupied state.

6. A fire passage illegal occupation system according to claim 5, characterized in that: The fire passage image capturing module is also used to process images in bad weather or poor lighting conditions; Specifically: image enhancement techniques, such as histogram equalization, contrast enhancement, gamma correction, etc., to improve image visibility; Filtering denoising, applying filtering algorithms, such as Gaussian filtering, median filtering or bilateral filtering, to reduce noise in the image.

7. A fire passage illegal occupation system according to claim 6, characterized in that: The vehicle encroachment status analysis module is also used to analyze the vehicle's lane borrowing status; If the vehicle occupation status analysis module determines that the vehicle is in a state of borrowing the lane, and the fire passage shooting module detects that the vehicle is not turned off in the fire passage, the YOLO and Faster R-CNN models are used to determine whether the driver has left: If the driver leaves the vehicle, and the time the driver has left the vehicle is less than the maximum time threshold, the vehicle is determined to have borrowed the lane for a short period of time; If the driver leaves the vehicle, and the time the driver has been away from the vehicle is greater than the maximum time threshold, it is determined that the vehicle has been borrowing the lane for a long time, and a signal is generated and transmitted to the occupancy warning module to issue a warning message to the vehicle owner.

8. A fire passage illegal occupation system according to claim 7, characterized in that: Based on the fact that the vehicle is not turned off in the fire passage, the specific calculation method of the driver's position from the vehicle is: When the fire channel camera module detects the driver's departure time, it is marked as , and from the departure time Start timing, and pre-set the maximum time threshold for leaving , and will be from The time of timing and Compare.

9. A fire passage illegal occupation system according to claim 6, characterized in that: The specific information of vehicle license plate extracted based on the fire channel image shooting module is as follows: Use image processing technology to determine the potential license plate area in the image, and use thresholding, edge detection, and morphological operations to achieve the shape and position of the license plate; The characters on the license plate are segmented into separate images by horizontal projection histogram and other methods, and features are extracted from the segmented character images. Commonly used features include Huff transform features, shape features, histogram features, etc. Use the labeled license plate character dataset to train the SVM model. For the province abbreviation and alphanumeric combination, create two datasets and train two models respectively; The extracted features are input into the trained SVM model, and the model will output the predicted category of the character; The recognized characters are stitched together to form a complete license plate number, and the complete license plate number is reported to the traffic control department.

10. A fire passage illegal occupation system according to claim 9, characterized in that: The fire passage image capturing module extracts vehicle license plate information and further comprises: If the complete license plate number is not extracted, or the location of the license plate is not identified through image processing technology, a signal is generated and transmitted to the occupancy warning module, and the occupancy warning module plays a warning message, which includes a horn broadcast or towing away with a trailer.

Images