CN111476224B - Safety belt detection method and device, electronic equipment and system - Google Patents
Safety belt detection method and device, electronic equipment and system Download PDFInfo
- Publication number
- CN111476224B CN111476224B CN202010593874.8A CN202010593874A CN111476224B CN 111476224 B CN111476224 B CN 111476224B CN 202010593874 A CN202010593874 A CN 202010593874A CN 111476224 B CN111476224 B CN 111476224B
- Authority
- CN
- China
- Prior art keywords
- safety belt
- detection
- driver
- safety
- belt
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/59—Context or environment of the image inside of a vehicle, e.g. relating to seat occupancy, driver state or inner lighting conditions
- G06V20/597—Recognising the driver's state or behaviour, e.g. attention or drowsiness
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
Abstract
The embodiment of the invention discloses a safety belt detection method, a safety belt detection device, electronic equipment and a safety belt detection system, wherein the safety belt detection method comprises the following steps: acquiring a driver driving state image acquired by a vehicle-mounted camera; carrying out safety belt detection on the driving state image of the driver by utilizing a safety belt detection deep learning network to obtain a safety belt detection result; wherein the seat belt detection result comprises that the driver has fastened a seat belt or the driver has not fastened the seat belt; and outputting the safety belt detection result. Therefore, the embodiment of the invention realizes the accurate and real-time detection of the wearing condition of the safety belt of the driver, and improves the reliability and the efficiency of the safety belt detection.
Description
Technical Field
The invention relates to the field of deep learning and machine vision, in particular to a safety belt detection method, a safety belt detection device, electronic equipment and a safety belt detection system.
Background
With the economic development, the speed of city construction is accelerated, and the number of automobiles is increasing year by year. The continuous and rapid growth of automobiles brings huge burden to urban traffic systems, illegal driving and poor driving behaviors still exist, and drivers do not wear safety belts and directly relate to the life safety of the drivers and other people.
The traditional safety belt detection is triggered by a contact switch in a lock catch, when a driver enters a cab and turns on an ignition key controller to start working, and when the safety belt is not inserted into the lock catch, a loop of a sound prompting device of the automobile safety belt is conducted to prompt the driver to fasten the safety belt; after the safety belt is fastened, the contact switch in the safety belt lock catch is in a closed state, and the power supply of the safety belt reminding system is disconnected, so that the reminding sound stops.
However, since the detection system is only directed to the inside of the buckle, there is a limitation in that the driver can drive without fastening the seat belt by shielding the seat belt prompt by using a seat belt buckle or by directly inserting the seat belt into the buckle from the back.
Disclosure of Invention
Because the existing method has the problems, the embodiment of the invention provides a safety belt detection method, a safety belt detection device, electronic equipment and a safety belt detection system, so that the wearing condition of a safety belt of a driver can be accurately detected in real time in a cab, a method for giving an alarm to the driver when the driver does not wear the safety belt is provided, and the problem that the existing bayonet safety belt detection system is easily shielded by the driver is solved.
In a first aspect, an embodiment of the present invention provides a seat belt detection method, including:
acquiring a driver driving state image acquired by a vehicle-mounted camera;
carrying out safety belt detection on the driving state image of the driver by utilizing a safety belt detection deep learning network to obtain a safety belt detection result; wherein the seat belt detection result comprises that the driver has fastened a seat belt or the driver has not fastened the seat belt;
and outputting the safety belt detection result.
Optionally, the performing seat belt detection on the driving state image of the driver by using a seat belt detection deep learning network to obtain a seat belt detection result includes:
inputting the driving state image of the driver into the safety belt detection deep learning network to obtain at least one safety belt pixel point;
clustering each safety belt pixel point to obtain the area of the safety belt;
and determining the safety belt detection result according to the number of the safety belt pixel points in the area of the safety belt.
Optionally, the seatbelt detection deep learning network is a semantic segmentation network using a deep separable convolution module of a mobile terminal neural network MobileNet in a U-type network structure.
Optionally, the clustering the pixels of the safety belts to obtain the area where the safety belts are located includes:
and clustering each safety belt pixel point by using a density-based noise application space clustering DBSCAN mode, only reserving the largest safety belt pixel point cluster after clustering, removing the safety belt pixel points in other areas as noise, and taking the largest pixel point cluster as the area where the safety belt is located.
Optionally, the determining the seat belt detection result according to the number of the seat belt pixel points in the area where the seat belt is located includes:
judging whether the safety belt is detected or not according to the number of safety belt pixel points in the area of the safety belt and a set number threshold value according to a first set rule to obtain a judgment result;
calculating the current accumulated detection number corresponding to the judgment result according to a second set rule;
if the current accumulated detection number is larger than 0, determining that the driver fastens the safety belt;
and if the current accumulated detection number is less than or equal to 0, determining that the driver does not wear the safety belt.
Optionally, the first setting rule includes: if the number of the safety belt pixel points is larger than or equal to the set number threshold, the safety belt is determined to be detected; if the number of the safety belt pixel points is smaller than the set number threshold, determining that the safety belt is not detected;
the second setting rule includes: the initial state of the current accumulated detection number is 0, and if the safety belt is detected by the current frame image, the value for representing the current accumulated detection number is added with 1; if the safety belt is not detected in the current frame image, subtracting 1 from the value representing the current accumulated detection number; and if the current accumulative detection number exceeds a preset accumulative detection upper limit frame number, setting the absolute value of the current accumulative detection number as the preset accumulative detection upper limit frame number.
Optionally, the outputting the seat belt detection result includes:
sending the safety belt detection result to a multimedia central control screen to enable the multimedia central control screen to display the safety belt detection result, and if the safety belt detection result indicates that the driver does not fasten the safety belt, outputting voice information of the safety belt not fastened by the multimedia central control screen; and/or
And if the safety belt detection result indicates that the accumulated continuous times of the unbelted safety belts of the driver exceed a set time threshold, sending unbelted safety belt alarm information to a server so that the server outputs the unbelted safety belt alarm information.
In a second aspect, an embodiment of the present invention provides a seat belt detection apparatus, including:
the acquisition module is used for acquiring a driver driving state image acquired by the vehicle-mounted camera;
the detection module is used for carrying out safety belt detection on the driving state image of the driver by utilizing a safety belt detection deep learning network to obtain a safety belt detection result; wherein the seat belt detection result comprises that the driver has fastened a seat belt or the driver has not fastened the seat belt;
and the output module is used for outputting the safety belt detection result.
In a third aspect, an embodiment of the present invention provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor implements the steps of the seat belt detection method according to the first aspect when executing the program.
In a fourth aspect, an embodiment of the present invention provides a seat belt detection system, including: the system comprises a vehicle-mounted camera, a vehicle-mounted terminal, a multimedia central control screen and/or a server;
the vehicle-mounted camera comprises an infrared camera, and the infrared camera is positioned above a vehicle instrument panel and on the left side or the right side of a vehicle steering wheel;
the vehicle-mounted terminal is used for realizing the steps of the safety belt detection method in the first aspect.
According to the technical scheme, the driving state image of the driver acquired by the vehicle-mounted camera is acquired, the safety belt detection is carried out on the driving state image of the driver by utilizing the safety belt detection depth learning network, and the safety belt detection result is obtained, wherein the safety belt detection result comprises that the driver fastens the safety belt or the driver does not fasten the safety belt, and the safety belt detection result is output, so that the safety belt wearing condition of the driver is accurately detected in real time, and the reliability and the efficiency of safety belt detection are improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
Fig. 1 is a schematic flow chart of a seat belt detection method according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of a seat belt detection deep learning network according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of a seat belt detection device according to an embodiment of the present invention;
fig. 4 is a schematic physical structure diagram of an electronic device according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of a seat belt detection system according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
For the convenience of clearly describing the technical solutions of the embodiments of the present invention, in each embodiment of the present invention, if words such as "first" and "second" are used to distinguish the same items or similar items with basically the same functions and actions, those skilled in the art can understand that the words such as "first" and "second" do not limit the quantity and execution order.
Fig. 1 is a schematic flow chart of a seat belt detection method according to an embodiment of the present invention; the method can be used for electronic equipment for realizing the safety belt detection function, such as: a vehicle-mounted terminal; as shown in fig. 1, the seat belt detection method may include:
s101, obtaining a driver driving state image collected by a vehicle-mounted camera.
Specifically, the vehicle-mounted camera may be an infrared camera, which may be located above a vehicle dashboard, and on the left or right side of a vehicle steering wheel.
Wherein, the installation requirement of infrared camera is in order not hindering the position that the driver normally drives the sight to, the image that this infrared camera was gathered requires that the safety belt has certain overlapping with driver's chest region, and the proportion of vehicle steering wheel in the image is as few as possible, avoids the driver to take the direction to influence the safety belt and detects.
S102, carrying out safety belt detection on the driving state image of the driver by using a safety belt detection deep learning network to obtain a safety belt detection result. Wherein the seat belt detection result comprises that the driver has fastened a seat belt or the driver has not fastened the seat belt.
Specifically, after acquiring a frame of driving state image of the driver, the vehicle-mounted terminal scales the image to 128 × 128 pixels according to the configured detection frequency, inputs the image into the seat belt detection deep learning network, and performs seat belt detection on the scaled image by using the seat belt detection deep learning network, thereby determining that the driver has fastened a seat belt or the driver has not fastened the seat belt.
And S103, outputting a safety belt detection result.
According to the embodiment, the driving state image of the driver collected by the vehicle-mounted camera is obtained, the safety belt detection is carried out on the driving state image of the driver by utilizing the safety belt detection depth learning network, and the safety belt detection result is obtained, wherein the safety belt detection result comprises that the driver wears the safety belt or the driver does not wear the safety belt, and the safety belt detection result is output, so that the safety belt wearing condition of the driver is accurately detected in real time, and the reliability and the efficiency of safety belt detection are improved.
Further, based on the above method, performing seat belt detection on the driving state image of the driver by using a deep seat belt detection learning network in S102 to obtain a seat belt detection result may include:
and S1021, inputting the driving state image of the driver into the safety belt detection deep learning network to obtain at least one safety belt pixel point.
Specifically, after acquiring a frame of driver driving state image, the vehicle-mounted terminal scales the image to 128 × 128 pixels according to the configured detection frequency, and inputs the image into the safety belt detection deep learning network to obtain at least one safety belt pixel point.
Alternatively, the seatbelt detection deep learning network in S102 or S1021 may be: in the U-type network structure, a semantic segmentation network of a deep separable convolution module of a mobile terminal neural network MobileNet, namely a mobileunet semantic segmentation network, is used.
The mobilenet is a combination of mobilenet and unet, that is, in a U-type network structure, a depth separable convolution module in the mobilenet is used. The mobilenet is a network structure which applies deep separable convolution and is suitable for reasoning on mobile equipment; the unet is because the network has two parts, the first part, feature extraction, the second part, and upsampling. Moreover, the network structure is U-shaped, so the network is called a Unet network.
The network structure of the safety belt detection deep learning network in S102 or S1021 is shown in fig. 2, and the mobilenet combines the deep separable convolution of the mobilenet and the U-shaped structure of feature extraction and upsampling in the inet, so that the inference speed is reduced while the edge is refined. The safety belt detection deep learning network inputs images of 128 x 3, and finally deduces results of 128 x n, wherein the results represent the probability that each pixel point is deduced to be a corresponding category.
Where 128 × 128 × 3 (i.e., length × width × channel) in fig. 2 represents the size of the input image. Wherein, the channel can be RGB three channel. N in 128 × 128 × n in fig. 2 represents the number of categories, and if n =2, represents 2 categories of the background and the seat belt.
And S1022, clustering the pixels of the safety belts to obtain the areas of the safety belts.
Specifically, each of the pixels of the seat belt may refer to a pixel having a maximum probability of the seat belt category. The clustering can be carried out on each safety belt pixel point, the largest cluster, namely the largest communication area, is reserved, and the other clusters are used as noise removal to determine the safety belt area.
Optionally, clustering the pixels of the safety belts in step S1022 to obtain the areas where the safety belts are located may include:
and clustering each safety belt pixel point by using a DBSCAN mode, only keeping the largest safety belt pixel point cluster after clustering, removing the safety belt pixel points in other areas as noise, and taking the largest pixel point cluster as the area where the safety belt is located.
The DBSCAN method is a Density-Based Noise application space clustering (Applications with Noise) method.
S1023, determining the safety belt detection result according to the number of the safety belt pixel points in the area where the safety belt is located.
Specifically, the safety belt detection result may be determined according to the number of safety belt pixel points in the area where the safety belt is located according to a certain rule.
Such as: if the number of the safety belt pixel points is larger than or equal to the set number threshold value, determining that the driver fastens the safety belt or the driver; and if the number of the safety belt pixel points is less than the set number threshold value, determining that the driver does not fasten the safety belt.
According to the embodiment, the driving state image of the driver is input into the safety belt detection deep learning network to obtain at least one safety belt pixel point, the safety belt pixel points are clustered to obtain the area where the safety belt is located, and the safety belt detection result is determined according to the number of the safety belt pixel points in the area where the safety belt is located, so that the safety belt detection accuracy is improved, and the problem that an existing bayonet safety belt detection system is easily shielded by the driver is solved.
Further, based on the foregoing method, the step of determining the seat belt detection result according to the number of seat belt pixel points in the area where the seat belt is located in S1023 may include:
and S1024, judging whether the safety belt is detected according to the number of the safety belt pixel points in the area where the safety belt is located and a set number threshold value according to a first set rule, and obtaining a judgment result.
Specifically, the first set rule may be a rule set in advance according to the actual situation for determining whether or not the seat belt is detected.
Alternatively, the first setting rule may include: if the number of the safety belt pixel points is larger than or equal to the set number threshold, the safety belt is determined to be detected; and if the number of the safety belt pixel points is smaller than the set number threshold, determining that the safety belt is not detected.
Specifically, if the number P of the safety belt pixel points is greater than or equal to a set number threshold value P, it is determined that the safety belt is detected by the current frame image; and if the number P of the safety belt pixel points is less than the set number threshold value P, determining that the safety belt is not detected in the current frame image. The set number threshold P may be a value set in advance according to actual conditions, such as: 65.
and S1025, calculating the current accumulated detection number corresponding to the judgment result according to a second set rule.
Specifically, the first setting rule may be a rule for calculating the current accumulated detection number set in advance according to the actual situation.
Optionally, the second setting rule may include: the initial state of the current accumulated detection number is 0, and if the safety belt is detected by the current frame image, the value for representing the current accumulated detection number is added with 1; if the safety belt is not detected in the current frame image, subtracting 1 from the value representing the current accumulated detection number; and if the current accumulative detection number exceeds a preset accumulative detection upper limit frame number, setting the absolute value of the current accumulative detection number as the preset accumulative detection upper limit frame number.
Such as: setting the accumulative detection upper limit frame number as S, the current accumulative detection number as F, setting the F initial state as 0, if a safety belt is detected in the current frame image, F = F +1, if the safety belt is not detected in the current frame image, F = F-1, and if the accumulative detection number exceeds the accumulative detection upper limit frame number, namely | F | > S, setting F as +/-S. The accumulated detection upper limit frame number is used for avoiding accidental false detection, or false alarm caused by that a safety belt is not detected due to shielding when a driver drives a steering wheel or the situation outside an observation window under the real situation.
And S1026, if the current accumulated detection number is larger than 0, determining that the driver fastens the safety belt.
S1027, if the current accumulated detection number is less than or equal to 0, determining that the driver does not wear the safety belt.
Specifically, setting the current accumulated detection number as F, and if F is greater than 0, determining that the driver fastens the safety belt; if F < =0, it is determined that the driver has fastened the seat belt.
As can be seen from the above embodiments, according to a first setting rule, whether a seat belt is detected or not can be determined according to the number of seat belt pixel points in the area where the seat belt is located and a set number threshold, a determination result is obtained, and a current accumulated detection number corresponding to the determination result is calculated according to a second setting rule, if the current accumulated detection number is greater than 0, it is determined that a driver has fastened the seat belt, and if the current accumulated detection number is less than or equal to 0, it is determined that the driver has not fastened the seat belt, so that the accuracy of seat belt detection is improved, especially, the accumulated detection upper limit number is used for avoiding accidental false detection, or false alarm caused by that the seat belt is not detected due to shielding when the driver is driving a steering wheel or outside an observation window under a real condition is further improved, and the reliability of seat belt detection.
Further, based on the above method, executing the step of outputting the seat belt detection result in S103 may include:
s1031, sending the safety belt detection result to a multimedia central control screen to enable the multimedia central control screen to display the safety belt detection result, and outputting the voice information of the safety belt not fastened if the safety belt detection result indicates that the driver does not fasten the safety belt; and/or
S1032, if the safety belt detection result indicates that the accumulated continuous times of the unbelted safety belts of the driver exceed a set time threshold, sending unbelted safety belt alarm information to a server so that the server outputs the unbelted safety belt alarm information.
Specifically, when the safety belt detection result is sent to a multimedia central control screen, when F is greater than 0, safety belt wearing information of a driver wearing a safety belt can be sent; when F < =0, the safety belt wearing information of a driver who does not wear a safety belt can be sent, therefore, after the safety belt wearing information is received, the safety belt detection result can be overlaid in the video to be displayed to the driver through the multimedia central control screen, and particularly when the multimedia central control screen receives the safety belt wearing information of the driver who does not wear the safety belt, the voice information of 'the driver does not wear the safety belt' can be broadcasted to prompt the driver to wear the safety belt.
In addition, when the unbelted belts are not fastened for 10 cumulative consecutive times (namely, the set number threshold), the vehicle-mounted terminal uploads the alarm information to the server, and the vehicle-mounted terminal firstly prompts the driver and uploads the alarm information when the driver does not always feel that the unbelted belts are not fastened), the uploading interval is 5 minutes for the continuous unbelted belt alarm. The alarm information uploaded to the server comprises: alarm time (year, month, day, hour, minute and second), location (GPS longitude and latitude), video picture for triggering alarm time, and video clips 5 seconds before and after the alarm time.
As can be seen from the above embodiments, when a seat belt detection result is output, the seat belt detection result may be sent to a multimedia central control screen, so that the multimedia central control screen displays the seat belt detection result, and if the seat belt detection result indicates that the driver does not fasten the seat belt, a seat belt unfastening voice message is output; and when the safety belt detection result indicates that the accumulated continuous times of unbelted safety belts of the driver exceed the set time threshold, sending unbelted safety belt alarm information to the server to enable the server to output the unbelted safety belt alarm information, so that voice prompt of unbelted safety belts and unbelted safety belt alarm are realized, and driving of the driver under the condition of not belted safety belts is avoided to the greatest extent.
Fig. 3 is a schematic structural diagram of a seat belt detection apparatus according to an embodiment of the present invention, where the seat belt detection apparatus may be used in an electronic device that implements a seat belt detection function, for example: a vehicle-mounted terminal; as shown in fig. 3, the seat belt detecting device may include:
the acquisition module 31 is used for acquiring a driver driving state image acquired by the vehicle-mounted camera;
the detection module 32 is configured to perform seat belt detection on the driving state image of the driver by using a seat belt detection deep learning network to obtain a seat belt detection result; wherein the seat belt detection result comprises that the driver has fastened a seat belt or the driver has not fastened the seat belt;
and an output module 33, configured to output the seat belt detection result.
Further, based on the above-mentioned device, the detecting module 32 includes:
the input submodule is used for inputting the driving state image of the driver into the safety belt detection deep learning network to obtain at least one safety belt pixel point;
the clustering submodule is used for clustering the pixels of the safety belts to obtain the areas of the safety belts;
and the determining submodule is used for determining the safety belt detection result according to the number of the safety belt pixel points in the area where the safety belt is located.
Further, based on the device, the deep learning network for seat belt detection is a semantic segmentation network which uses a deep separable convolution module of a mobile terminal neural network MobileNet in a U-shaped network structure.
Further, based on the above-mentioned apparatus, the clustering submodule may include:
and the clustering unit is used for clustering the safety belt pixel points by utilizing a density-based noise application space clustering DBSCAN mode, only the largest safety belt pixel point cluster is reserved after clustering, the safety belt pixel points in other areas are taken as noise to be removed, and the largest pixel point cluster is the area where the safety belt is located.
Further, based on the above-mentioned apparatus, the determining unit may include:
the judging unit is used for judging whether the safety belt is detected according to the number of the safety belt pixel points in the area where the safety belt is located and a set number threshold value according to a first set rule to obtain a judgment result;
the calculating unit is used for calculating the current accumulated detection number corresponding to the judgment result according to a second set rule;
the first determination unit is used for determining that the driver fastens the safety belt if the current accumulated detection number is larger than 0;
and the second determination unit is used for determining that the driver does not wear the safety belt if the current accumulated detection number is less than or equal to 0.
Further, based on the above-mentioned apparatus, the first setting rule includes: if the number of the safety belt pixel points is larger than or equal to the set number threshold, the safety belt is determined to be detected; if the number of the safety belt pixel points is smaller than the set number threshold, determining that the safety belt is not detected;
the second setting rule includes: the initial state of the current accumulated detection number is 0, and if the safety belt is detected by the current frame image, the value for representing the current accumulated detection number is added with 1; if the safety belt is not detected in the current frame image, subtracting 1 from the value representing the current accumulated detection number; and if the current accumulative detection number exceeds a preset accumulative detection upper limit frame number, setting the absolute value of the current accumulative detection number as the preset accumulative detection upper limit frame number.
Further, based on the above-mentioned apparatus, the output module may include:
the first output sub-module is used for sending the safety belt detection result to a multimedia central control screen so that the multimedia central control screen can display the safety belt detection result, and if the safety belt detection result indicates that the driver does not fasten the safety belt, the multimedia central control screen outputs the voice information of the safety belt not fastened; and/or
And the second output submodule is used for sending the warning information of the unbuckled safety belt to a server if the safety belt detection result indicates that the accumulated continuous times of unbuckled safety belt of the driver exceeds a set time threshold value, so that the server outputs the warning information of the unbuckled safety belt.
The safety belt detection device provided by the embodiment of the invention can be used for executing the method embodiment, the principle and the technical effect are similar, and the details are not repeated here.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the embodiment of the present invention. One of ordinary skill in the art can understand and implement it without inventive effort.
Fig. 4 is a schematic entity structure diagram of an electronic device according to an embodiment of the present invention, and as shown in fig. 4, the electronic device may include: a processor (processor)401, a communication Interface (communication Interface)402, a memory (memory)403 and a communication bus 404, wherein the processor 401, the communication Interface 402 and the memory 403 complete communication with each other through the communication bus 404. Processor 401 may call logic instructions in memory 403 to perform the following method:
acquiring a driver driving state image acquired by a vehicle-mounted camera;
carrying out safety belt detection on the driving state image of the driver by utilizing a safety belt detection deep learning network to obtain a safety belt detection result; wherein the seat belt detection result comprises that the driver has fastened a seat belt or the driver has not fastened the seat belt;
and outputting the safety belt detection result.
In addition, the logic instructions in the memory 403 may be implemented in the form of software functional units and stored in a computer readable storage medium when the software functional units are sold or used as independent products. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
Further, embodiments of the present invention disclose a computer program product comprising a computer program stored on a non-transitory computer-readable storage medium, the computer program comprising program instructions, which when executed by a computer, the computer is capable of performing the methods provided by the above-mentioned method embodiments, for example, comprising:
acquiring a driver driving state image acquired by a vehicle-mounted camera;
carrying out safety belt detection on the driving state image of the driver by utilizing a safety belt detection deep learning network to obtain a safety belt detection result; wherein the seat belt detection result comprises that the driver has fastened a seat belt or the driver has not fastened the seat belt;
and outputting the safety belt detection result.
In another aspect, an embodiment of the present invention further provides a non-transitory computer-readable storage medium, on which a computer program is stored, where the computer program is implemented by a processor to perform the method provided by the foregoing embodiments, for example, including:
acquiring a driver driving state image acquired by a vehicle-mounted camera;
carrying out safety belt detection on the driving state image of the driver by utilizing a safety belt detection deep learning network to obtain a safety belt detection result; wherein the seat belt detection result comprises that the driver has fastened a seat belt or the driver has not fastened the seat belt;
and outputting the safety belt detection result.
Further, as shown in fig. 5, an embodiment of the present invention further discloses a seat belt detection system, including: the system comprises a vehicle-mounted camera 501, a vehicle-mounted terminal 502, a multimedia center control screen 503 and/or a server 504;
the vehicle-mounted camera 501 comprises an infrared camera, and the infrared camera is located above a vehicle instrument panel and on the left side or the right side of a vehicle steering wheel;
the vehicle-mounted terminal 502 is used for implementing the steps of the safety belt detection method.
Specifically, the server 504 may be placed within a machine room server 504 provided by a user to deploy a vehicle management system; the vehicle-mounted camera 501, the vehicle-mounted terminal 502 and the multimedia center control screen 503 can be installed in a vehicle cab. The installation positions of the vehicle-mounted terminal 502 and the multimedia center control screen 503 have no strict requirement, and the vehicle-mounted terminal and the multimedia center control screen are convenient for a driver to operate and watch and are installed at the position which does not obstruct the normal driving sight of the driver.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (7)
1. A seat belt detection method, comprising:
acquiring a driver driving state image acquired by a vehicle-mounted camera;
carrying out safety belt detection on the driving state image of the driver by utilizing a safety belt detection deep learning network to obtain a safety belt detection result; wherein the seat belt detection result comprises that the driver has fastened a seat belt or the driver has not fastened the seat belt;
outputting the safety belt detection result;
the safety belt detection is carried out on the driving state image of the driver by utilizing the safety belt detection deep learning network to obtain a safety belt detection result, and the safety belt detection method comprises the following steps:
inputting the driving state image of the driver into the safety belt detection deep learning network to obtain at least one safety belt pixel point;
clustering each safety belt pixel point to obtain the area of the safety belt;
determining the safety belt detection result according to the number of safety belt pixel points in the area of the safety belt;
the determining the safety belt detection result according to the number of the safety belt pixel points in the area where the safety belt is located includes:
judging whether the safety belt is detected or not according to the number of safety belt pixel points in the area of the safety belt and a set number threshold value according to a first set rule to obtain a judgment result;
calculating the current accumulated detection number corresponding to the judgment result according to a second set rule;
if the current accumulated detection number is larger than 0, determining that the driver fastens the safety belt;
if the current accumulated detection number is less than or equal to 0, determining that the driver does not fasten the safety belt;
wherein the first setting rule includes: if the number of the safety belt pixel points is larger than or equal to the set number threshold, the safety belt is determined to be detected; if the number of the safety belt pixel points is smaller than the set number threshold, determining that the safety belt is not detected;
the second setting rule includes: the initial state of the current accumulated detection number is 0, and if the safety belt is detected by the current frame image, the value for representing the current accumulated detection number is added with 1; if the safety belt is not detected in the current frame image, subtracting 1 from the value representing the current accumulated detection number; and if the current accumulative detection number exceeds a preset accumulative detection upper limit frame number, setting the absolute value of the current accumulative detection number as the preset accumulative detection upper limit frame number.
2. The method according to claim 1, wherein the seatbelt detection deep learning network is a semantic segmentation network using a deep separable convolution module of a mobile terminal neural network MobileNet in a U-type network configuration.
3. The method for detecting the seat belt according to claim 1, wherein the clustering the pixels of the seat belt to obtain the area where the seat belt is located comprises:
and clustering each safety belt pixel point by using a density-based noise application space clustering DBSCAN mode, only reserving the largest safety belt pixel point cluster after clustering, removing the safety belt pixel points in other areas as noise, and taking the largest pixel point cluster as the area where the safety belt is located.
4. The seat belt detection method according to claim 1, wherein the outputting the seat belt detection result includes:
sending the safety belt detection result to a multimedia central control screen to enable the multimedia central control screen to display the safety belt detection result, and if the safety belt detection result indicates that the driver does not fasten the safety belt, outputting voice information of the safety belt not fastened by the multimedia central control screen; and/or
And if the safety belt detection result indicates that the accumulated continuous times of the unbelted safety belts of the driver exceed a set time threshold, sending unbelted safety belt alarm information to a server so that the server outputs the unbelted safety belt alarm information.
5. A seat belt detection apparatus, comprising:
the acquisition module is used for acquiring a driver driving state image acquired by the vehicle-mounted camera;
the detection module is used for carrying out safety belt detection on the driving state image of the driver by utilizing a safety belt detection deep learning network to obtain a safety belt detection result; wherein the seat belt detection result comprises that the driver has fastened a seat belt or the driver has not fastened the seat belt;
the output module is used for outputting the safety belt detection result;
the detection module comprises:
the input submodule is used for inputting the driving state image of the driver into the safety belt detection deep learning network to obtain at least one safety belt pixel point;
the clustering submodule is used for clustering the pixels of the safety belts to obtain the areas of the safety belts;
the determining submodule is used for determining the safety belt detection result according to the number of the safety belt pixel points in the area where the safety belt is located;
the determination sub-module includes:
the judging unit is used for judging whether the safety belt is detected according to the number of the safety belt pixel points in the area where the safety belt is located and a set number threshold value according to a first set rule to obtain a judgment result;
the calculating unit is used for calculating the current accumulated detection number corresponding to the judgment result according to a second set rule;
the first determination unit is used for determining that the driver fastens the safety belt if the current accumulated detection number is larger than 0;
the second determining unit is used for determining that the driver does not wear the safety belt if the current accumulated detection number is less than or equal to 0;
wherein the first setting rule includes: if the number of the safety belt pixel points is larger than or equal to the set number threshold, the safety belt is determined to be detected; if the number of the safety belt pixel points is smaller than the set number threshold, determining that the safety belt is not detected;
the second setting rule includes: the initial state of the current accumulated detection number is 0, and if the safety belt is detected by the current frame image, the value for representing the current accumulated detection number is added with 1; if the safety belt is not detected in the current frame image, subtracting 1 from the value representing the current accumulated detection number; and if the current accumulative detection number exceeds a preset accumulative detection upper limit frame number, setting the absolute value of the current accumulative detection number as the preset accumulative detection upper limit frame number.
6. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the steps of the seat belt detection method according to any of claims 1 to 4 are implemented when the processor executes the program.
7. A seat belt detection system, comprising: the system comprises a vehicle-mounted camera, a vehicle-mounted terminal, a multimedia central control screen and/or a server;
the vehicle-mounted camera comprises an infrared camera, and the infrared camera is positioned above a vehicle instrument panel and on the left side or the right side of a vehicle steering wheel;
the vehicle-mounted terminal is used for realizing the steps of the safety belt detection method according to any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010593874.8A CN111476224B (en) | 2020-06-28 | 2020-06-28 | Safety belt detection method and device, electronic equipment and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010593874.8A CN111476224B (en) | 2020-06-28 | 2020-06-28 | Safety belt detection method and device, electronic equipment and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111476224A CN111476224A (en) | 2020-07-31 |
| CN111476224B true CN111476224B (en) | 2020-10-09 |
Family
ID=71765359
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010593874.8A Active CN111476224B (en) | 2020-06-28 | 2020-06-28 | Safety belt detection method and device, electronic equipment and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111476224B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111931642A (en) * | 2020-08-07 | 2020-11-13 | 上海商汤临港智能科技有限公司 | Safety belt wearing detection method and device, electronic equipment and storage medium |
| US20220203930A1 (en) * | 2020-12-29 | 2022-06-30 | Nvidia Corporation | Restraint device localization |
| CN113298000B (en) * | 2021-06-02 | 2022-10-25 | 上海大学 | Safety belt detection method and device based on infrared camera |
| CN115187967B (en) * | 2022-09-13 | 2023-02-17 | 苏州魔视智能科技有限公司 | Detection method, training method, electronic device and readable storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108647630A (en) * | 2018-05-08 | 2018-10-12 | 北京优创新港科技股份有限公司 | A kind of dangerous driving behavior measure of supervision and device based on video identification |
| CN109460699A (en) * | 2018-09-03 | 2019-03-12 | 厦门瑞为信息技术有限公司 | A kind of pilot harness's wearing recognition methods based on deep learning |
| CN110598560A (en) * | 2019-08-15 | 2019-12-20 | 重庆特斯联智慧科技股份有限公司 | Night monitoring and identifying method and system based on neural network enhancement |
-
2020
- 2020-06-28 CN CN202010593874.8A patent/CN111476224B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108647630A (en) * | 2018-05-08 | 2018-10-12 | 北京优创新港科技股份有限公司 | A kind of dangerous driving behavior measure of supervision and device based on video identification |
| CN109460699A (en) * | 2018-09-03 | 2019-03-12 | 厦门瑞为信息技术有限公司 | A kind of pilot harness's wearing recognition methods based on deep learning |
| CN110598560A (en) * | 2019-08-15 | 2019-12-20 | 重庆特斯联智慧科技股份有限公司 | Night monitoring and identifying method and system based on neural network enhancement |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111476224A (en) | 2020-07-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111476224B (en) | Safety belt detection method and device, electronic equipment and system | |
| US20220292956A1 (en) | Method and system for vehicular-related communications | |
| US20240109514A1 (en) | Recording video of an operator and a surrounding visual field | |
| US10235768B2 (en) | Image processing device, in-vehicle display system, display device, image processing method, and computer readable medium | |
| US9707971B2 (en) | Driving characteristics diagnosis device, driving characteristics diagnosis system, driving characteristics diagnosis method, information output device, and information output method | |
| CN111469802A (en) | Seat belt state determination system and method | |
| EP3825981A1 (en) | Warning apparatus, driving tendency analysis device, driving tendency analysis method, and program | |
| JP6950432B2 (en) | Driving support device, information processing device, driving support system, driving support method | |
| JP5521893B2 (en) | Driving support system, in-vehicle device | |
| CN108928294A (en) | Driving dangerous based reminding method, device, terminal and computer readable storage medium | |
| CN103578293B (en) | The warning system of vehicle and method | |
| US8688321B2 (en) | Traffic density estimation | |
| CN111210620B (en) | Method, device and equipment for generating driver portrait and storage medium | |
| JP2014081947A (en) | Information distribution device | |
| CN111661059B (en) | Method and system for monitoring distracted driving and electronic equipment | |
| US10741076B2 (en) | Cognitively filtered and recipient-actualized vehicle horn activation | |
| CN110869262A (en) | Method, mobile user equipment, computer program for generating visual information for at least one occupant of a vehicle | |
| US10891496B2 (en) | Information presentation method | |
| CN113436464B (en) | Vehicle danger early warning method, device, equipment and storage medium | |
| CN114332941A (en) | Alarm prompting method and device based on riding object detection and electronic equipment | |
| CN115520196B (en) | Weight determination method and device for driver, electronic equipment and storage medium | |
| GB2597092A (en) | A method for determining a state of mind of a passenger, as well as an assistance system | |
| CN114390254B (en) | Rear-row cockpit monitoring method and device and vehicle | |
| CN115019242B (en) | Abnormal event detection method and device for traffic scene and processing equipment | |
| CN111583714A (en) | Vehicle driving early warning method and device, computer readable medium and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |