KR102517546B1 - System and method for predicting danger during driving of vehicle - Google Patents

Abstract

A vehicle driving risk prediction system and method capable of preventing a secondary accident by quickly predicting the risk of a collision between obstacles (surrounding vehicles or people) around the vehicle are disclosed. A vehicle driving risk prediction system according to an embodiment of the present invention is a vehicle driving risk prediction system for predicting the driving risk of a running vehicle, which is configured to sequentially acquire images of objects located around the vehicle. image acquisition unit; a moving object extraction unit configured to extract moving objects by analyzing the images; a motion vector calculation unit configured to calculate motion vector profiles corresponding to motion vectors according to time for each of the moving objects; an accident risk prediction unit configured to predict an accident risk of the vehicle by an artificial neural network based on motion vector profiles calculated for the moving objects; and a danger notification unit configured to inform the driver of the accident risk when the accident risk is predicted.

Description

Vehicle driving risk prediction system and method {System and method for predicting danger during driving of vehicle}

The present invention is for predicting vehicle driving risk, and more particularly, a vehicle driving risk prediction system and method capable of preventing secondary accidents by quickly predicting the risk of collision between obstacles (surrounding vehicles or people) around the vehicle and its method It is about.

A technology for detecting obstacles by mounting sensors such as a stereo camera and lidar sensor on a vehicle to recognize and analyze surrounding objects is being researched. Conventional technology is mainly focused on predicting the risk of collision between a vehicle and a nearby vehicle, and research on the risk prediction of a secondary accident caused by a collision between nearby vehicles or between a nearby vehicle and a person is lacking.

In addition, in order to predict the risk of a traffic accident in real time while driving a vehicle, it is necessary to quickly predict the risk of an accident from data collected through an image sensor such as a stereo camera or lidar sensor. There is a limitation that it is difficult to apply to the actual driving environment because the time required to predict the risk of an accident is long.

An object of the present invention is to provide a vehicle driving risk prediction system and method capable of preventing a secondary accident by quickly predicting the risk of a collision between obstacles (surrounding vehicles or people) around the vehicle and its method.

A vehicle driving risk prediction system according to an embodiment of the present invention is a vehicle driving risk prediction system for predicting the driving risk of a running vehicle, which is configured to sequentially acquire images of objects located around the vehicle. image acquisition unit; a moving object extraction unit configured to extract moving objects by analyzing the images; a motion vector calculation unit configured to calculate motion vector profiles corresponding to motion vectors according to time for each of the moving objects; an accident risk prediction unit configured to predict an accident risk of the vehicle by an artificial neural network based on motion vector profiles calculated for the moving objects; and a risk notification unit configured to inform the driver of the accident risk when the accident risk is predicted.

The accident risk prediction unit: Differential motion that is a difference value between a first motion vector profile calculated for a first moving object among the moving objects and a second motion vector profile calculated for a second moving object among the moving objects. calculate a vector profile; Based on the first motion vector profile, the second motion vector profile, and the differential motion vector profile, a collision probability between the first moving object and the second moving object, a predicted collision position, and a collision by the artificial neural network Predicting collision risk information including a post-movement path; And based on the predicted collision risk information and the driving information of the vehicle, configured to determine a countermeasure for preventing an accident of the vehicle, including braking control-related information of the vehicle and direction change-related information of the vehicle. It can be.

The accident risk prediction unit: collects information related to the road on which the vehicle is driving, and surrounding environment information including information related to weather at the time the vehicle is driving at the time the vehicle is driving; processing the images to analyze moving object related information including a type of the first moving object, a type of the second moving object, and an estimated weight or vehicle type of the first and second moving objects; processing the images to calculate a first direction profile representing direction information of the first moving object over time and a second direction profile representing direction information of the second moving object over time; The collision risk information may be predicted by the artificial neural network based on the surrounding environment information, the moving object related information, the first direction profile, and the second direction profile.

A vehicle driving risk prediction method according to an embodiment of the present invention is a vehicle driving risk prediction method for predicting the driving risk of a vehicle in motion, and sequentially detects objects located around the vehicle by an image acquisition unit. acquiring images; extracting moving objects by analyzing the images by a moving object extraction unit; calculating, by a motion vector calculation unit, motion vector profiles corresponding to motion vectors of each of the moving objects according to time; predicting, by an accident risk predictor, an accident risk of the vehicle based on motion vector profiles calculated for the moving objects by an artificial neural network; and notifying the driver of the vehicle of the accident risk when the accident risk is predicted by the risk notification unit.

The step of predicting the accident risk may include: a difference between a first motion vector profile calculated for a first moving object among the moving objects and a second motion vector profile calculated for a second moving object among the moving objects. Calculating a differential motion vector profile of Based on the first motion vector profile, the second motion vector profile, and the differential motion vector profile, a collision probability between the first moving object and the second moving object, a predicted collision position, and a collision by the artificial neural network Predicting collision risk information including a post-movement path; and determining a countermeasure for preventing an accident of the vehicle, including braking control-related information of the vehicle and direction change-related information of the vehicle, based on the predicted collision risk information and driving information of the vehicle. can include

The predicting the risk of an accident may include: collecting surrounding environment information including information related to a road on which the vehicle is driving and information related to weather at a point in time when the vehicle is driving; Processing the images to analyze moving object-related information including the type of the first moving object, the type of the second moving object, and estimated weights or vehicle types of the first and second moving objects; ; processing the images to calculate a first direction profile representing direction information of the first moving object over time and a second direction profile representing direction information of the second moving object over time; and predicting the collision risk information by the artificial neural network based on the surrounding environment information, the moving object related information, the first direction profile, and the second direction profile.

According to an embodiment of the present invention, a computer-readable recording medium on which a program for executing the vehicle driving risk prediction method is recorded is provided.

According to an embodiment of the present invention, a vehicle driving risk prediction system and method capable of preventing a secondary accident by quickly predicting the risk of a collision between obstacles (surrounding vehicles or people) around the vehicle are provided.

1 is a block diagram of a vehicle driving risk prediction system according to an embodiment of the present invention.
2 is a configuration diagram of an accident risk prediction unit constituting a vehicle driving risk prediction system according to an embodiment of the present invention.
3 is a flowchart of a vehicle driving risk prediction method according to an embodiment of the present invention.
4 is a conceptual diagram for explaining a vehicle driving risk prediction method according to an embodiment of the present invention.
5 is a flowchart illustrating step S140 of FIG. 3 according to an embodiment of the present invention.
6 is a flowchart illustrating step S140 of FIG. 3 according to another embodiment of the present invention.
FIG. 7 is a conceptual diagram for explaining step S147 of FIG. 6 .

Advantages and features of the present invention, and methods of achieving them, will become clear with reference to the detailed description of the following embodiments taken in conjunction with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but may be implemented in various different forms, and only these embodiments make the disclosure of the present invention complete, and common knowledge in the art to which the present invention belongs. It is provided to completely inform the person who has the scope of the invention, and the present invention is only defined by the scope of the claims. Like reference numbers designate like elements throughout the specification.

In this specification, when a certain component is said to "include", it means that it may further include other components, not excluding other components unless otherwise stated. '~ unit' used in this specification is a unit that processes at least one function or operation, and may mean, for example, software, an FPGA, or a hardware component. Functions provided by '~unit' may be performed separately by a plurality of components or may be integrated with other additional components. '~unit' in this specification is not necessarily limited to software or hardware, and may be configured to be in an addressable storage medium or configured to reproduce one or more processors. Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

1 is a block diagram of a vehicle driving risk prediction system according to an embodiment of the present invention. Referring to FIG. 1 , the vehicle driving risk prediction system 100 according to an embodiment of the present invention predicts the driving risk of a vehicle while driving, detects and informs obstacles in the vehicle's blind spots or surrounding situations, and obtains an image. unit 110, moving object extraction unit 120, motion vector calculation unit 130, accident risk prediction unit 140, artificial neural network 150, artificial neural network learning unit 160, and risk notification unit 170 can include

The image acquisition unit 110 may be mounted on a vehicle. The image acquisition unit 110 may sequentially acquire images of objects (eg, a car, a person, a surrounding structure, etc.) located around a vehicle in motion. The image acquisition unit 110 may include a device for obtaining a 3D image, for example, a 3D camera such as a stereo camera, a lidar sensor, and the like.

The moving object extractor 120 may analyze the images acquired by the image acquirer 110 and extract moving objects such as nearby cars and people. In one embodiment, the moving object extractor 120 may extract moving objects using a deep learning algorithm and/or an algorithm for extracting objects from an image.

The motion vector calculator 130 may calculate a motion vector profile corresponding to a motion vector according to time for each of the moving objects extracted by the moving object extractor 120 . The motion vector profile may indicate change information of sequential motion vectors at regular time intervals. The motion vector may include movement direction information and movement speed (speed) information of a moving object moving around the vehicle.

The accident risk prediction unit 140 is configured according to a collision accident between moving objects by an artificial neural network 150 based on motion vector profiles calculated for moving objects located around a vehicle in motion. The accident risk (secondary accident risk) of the vehicle can be predicted.

The artificial neural network learning unit 160 may learn an artificial neural network for predicting an accident risk of a vehicle using learning data. Data for learning of the artificial neural network 150 may include, for example, data corresponding to a simulation of a collision between vehicles and an actual traffic accident. The artificial neural network learning unit 160 may learn the artificial neural network 150 before and/or after driving of the vehicle based on the collected learning data.

Learning data used for learning of the artificial neural network 150 includes input variables including road-related information, weather-related information, moving object-related information, etc. ) may include output variables including the movement path. Such learning data may be collected from information about past traffic accidents or generated by a simulation device simulating traffic accidents.

In an embodiment, the artificial neural network learning unit 160 includes road-related information (eg, road slope, road type, speed limit, etc.) and weather-related information (rainy weather information, snowfall information, snowfall amount, etc.) information, moving object-related information including the type of moving object (car, truck, bus, motorcycle, person, etc.), the weight of the moving object, or the type of vehicle (eg, make, vehicle model, etc.), the time of the moving object It includes input variables including a direction profile corresponding to directional information, a motion vector profile of a moving object, and a differential motion vector profile that is a difference between motion vector profiles between moving objects, a collision predicted position, and a movement path after collision. The artificial neural network 150 may be learned by learning weights between the output variables.

The artificial neural network 150 is, for example, a fully connected layer artificial neural network in which an input layer, one or more hidden layers, and an output layer are connected with weights, and a convolutional method for extracting feature maps by convolution processing. ) may be implemented with an artificial neural network, etc., but is not limited thereto.

The danger notification unit 170 may inform the driver of the accident risk when the accident risk is predicted by the accident risk prediction unit 140 . The danger notification unit 170 may notify the driver of the danger of an accident through a warning sound, a warning voice, or a warning light, or notify the driver of the danger of an accident through a vehicle navigation system or a display screen.

In addition, the danger notification unit 170 monitors the surroundings (eg, the rear) of the moving objects according to the collision probability of the surrounding moving objects (surrounding vehicles or people, etc.), the predicted location of the collision, and the predicted movement path of the moving objects after the collision. The vehicle may be automatically controlled to prevent a secondary collision, such as having the vehicle in motion change lanes, or the driver may be guided to control the vehicle.

2 is a configuration diagram of an accident risk prediction unit constituting a vehicle driving risk prediction system according to an embodiment of the present invention. Referring to FIGS. 1 and 2 , the accident risk prediction unit 140 includes a differential motion vector calculation unit 141, a surrounding environment information collection unit 142, a moving object information analysis unit 143, and a direction profile calculation unit 144. ), and a collision risk information prediction unit 145.

The differential motion vector calculation unit 141 includes a first motion vector profile calculated for a first moving object among moving objects moving around the vehicle and a second motion vector calculated for a second moving object among the moving objects. A differential motion vector profile corresponding to the difference between the profiles (vector difference operation result) may be calculated.

The surrounding environment information collection unit 142 includes information related to the road on which the vehicle is driving (eg, the slope of the road (uphill road, downhill road, flat land, slope angle, etc.), tunnel, road type (straight road, curved road, intersection, Highway, etc.), speed limit, etc., surrounding environment information including weather related information (rainy weather information, snowfall information, snowfall information, etc.) at the time the vehicle is driving may be collected. The unit 142 may collect surrounding environment information from one or more servers (meteorological office server, electronic map server, etc.) or a navigation system providing information related to the driving environment of the vehicle.

The moving object information analyzer 143 processes the images acquired by the image acquisition unit 110 to determine the types of first and second moving objects (large car, mid-size car, slope, truck, bus, motorcycle, person, etc.) , moving object-related information including estimated weights of the first and second moving objects or vehicle types (eg, manufacturer, vehicle model, etc.) may be analyzed. The moving object information analyzer 143 may analyze information related to the moving object by extracting the size, manufacturer's symbol, etc. of the moving object from the image.

The surrounding environment information and information related to the moving object may be analyzed in advance several seconds or more than tens of seconds before a risk of collision between the moving objects occurs while the vehicle is driving. Therefore, at the time of occurrence of risk of collision of real moving objects, the motion vector profile in the risk of collision situation is added to the artificial neural network 150 along with the previously analyzed surrounding environment information and information related to the moving object. Accident risk prediction results may be output in real time from the artificial neural network 150 within a few seconds.

The direction profile calculation unit 144 processes the images acquired by the image acquisition unit 110 to obtain a first direction profile representing direction information over time of the first moving object and direction information over time of the second moving object. A second direction profile representing can be calculated.

The collision risk information prediction unit 145 includes the motion vector profiles of the first and second moving objects calculated by the motion vector calculator 130, the differential motion vector profiles calculated by the differential motion vector calculator 141, and the surroundings. Surrounding environment information collected by the environment information collection unit 142, moving object related information analyzed by the moving object information analysis unit 143, and first and second moving objects calculated by the direction profile calculation unit 144 Based on the direction profile of the artificial neural network 150, collision risk information including a collision probability between the first moving object and the second moving object, an expected collision position, and a movement path after collision may be predicted.

In addition, the collision risk information prediction unit 145 considers the predicted braking distance of the vehicle based on the predicted collision risk information (probability of occurrence of a collision) and the driving information of the vehicle, and provides vehicle-related information and vehicle direction change-related information. It is possible to determine countermeasures for vehicle accident prevention, including whether to change lanes and which lanes to change.

The vehicle driving risk prediction system according to an embodiment of the present invention does not perform calculations for vehicle collision simulation while driving, and based on only information that can be quickly collected through image processing, collision probability, collision probability, It is possible to predict the predicted location of the collision, the movement path after the collision, and predict the risk of an accident in real time while driving the vehicle.

3 is a flowchart of a vehicle driving risk prediction method according to an embodiment of the present invention. 4 is a conceptual diagram for explaining a vehicle driving risk prediction method according to an embodiment of the present invention. 1 to 4 , in the vehicle driving risk prediction method according to an embodiment of the present invention, an object located around the vehicle 10 while driving is first performed by the image acquisition unit 110 to predict the vehicle driving risk. Images (eg, cars, people, surrounding structures, etc.) may be sequentially obtained (S110).

The moving object extractor 120 may analyze the images acquired by the image acquirer 110 and extract moving objects 20 and 30 such as nearby cars and people (S120). In one embodiment, the moving object extractor 120 may extract the moving objects 20 and 30 using a deep learning algorithm or an image object extraction algorithm.

The motion vector calculator 130 may calculate motion vector profiles corresponding to motion vectors 24 and 34 according to time, respectively, for the moving objects 20 and 30 extracted by the moving object extractor 120. Yes (S130). The motion vectors 24 and 34 may include movement direction information and movement speed information of the moving objects 20 and 30 moving around the vehicle.

In the embodiment, the motion vectors 24 and 34 of the moving objects 20 and 30 are the motion vectors 22 and 32 of the moving objects 20 and 30 measured by the vehicle 10, respectively. ) may be a vector obtained by subtracting the motion vector 12 of The motion vector 12 of the vehicle 10 may be a vector calculated based on the moving direction and speed of the vehicle, and may be collected by various sensors or image sensors of the vehicle 10 .

The accident risk prediction unit 140 is configured according to a collision accident between moving objects by an artificial neural network 150 based on motion vector profiles calculated for moving objects located around a vehicle in motion. The accident risk of the vehicle (secondary accident risk) and the predicted collision position 40 may be predicted (S140).

The artificial neural network learning unit 160 may learn an artificial neural network for predicting an accident risk of a vehicle using learning data. Data for learning of the artificial neural network 150 may include, for example, data corresponding to a simulation of a collision between vehicles and an actual traffic accident.

The danger notification unit 170 may inform the driver of the accident risk when the accident risk is predicted by the accident risk prediction unit 140 (S150). The danger notification unit 170 may notify the danger of an accident through a warning sound, a warning voice, or a warning light, or through a vehicle navigation system or a display screen.

5 is a flowchart illustrating step S140 of FIG. 3 according to an embodiment of the present invention. Referring to FIGS. 1 to 5 , the differential motion vector calculation unit 141 calculates a first motion vector profile calculated for a first moving object among moving objects moving around the vehicle and a second movement among the moving objects. A differential motion vector profile, which is a difference value between the second motion vector profiles calculated for the object, may be calculated (S141).

The collision risk information prediction unit 145 is based on the motion vector profiles of the first and second moving objects calculated by the motion vector calculator 130 and the differential motion vector profiles calculated by the differential motion vector calculator 141. As a result, it is possible to predict collision risk information including the probability of collision between the first moving object 20 and the second moving object 30, the expected collision position 40, and the moving path after the collision by the artificial neural network 150. (S142).

The collision risk information prediction unit 145 includes information related to braking control of the vehicle and information related to direction change of the vehicle (whether lane is changed and which lane to change) based on the predicted collision risk information and driving information of the vehicle, A countermeasure for preventing a vehicle accident may be determined (S143).

6 is a flowchart illustrating step S140 of FIG. 3 according to another embodiment of the present invention. FIG. 7 is a conceptual diagram for explaining step S147 of FIG. 6 . 1 to 7 , the surrounding environment information collection unit 142 provides information related to the road on which the vehicle is driving (eg, road slope (uphill, downhill, flat, slope angle, etc.), tunnel, and road type). (straight roads, curved roads, intersections, highways, etc.), speed limit, etc., and surrounding environment information including information related to the weather (rain information, snowfall information, snowfall, etc.) It can be collected (S144).

The moving object information analyzer 143 processes the images acquired by the image acquisition unit 110 to determine the types of first and second moving objects (large car, mid-size car, slope, truck, bus, motorcycle, person, etc.) , moving object related information including estimated weights of the first and second moving objects or vehicle types (eg, manufacturer, vehicle model, etc.) may be analyzed (S145).

The direction profile calculation unit 144 processes the images acquired by the image acquisition unit 110 to obtain a first direction profile representing direction information over time of the first moving object and direction information over time of the second moving object. It is possible to calculate a second direction profile representing (S146).

The collision risk information prediction unit 145 includes the motion vector profiles of the first and second moving objects calculated by the motion vector calculator 130, the differential motion vector profiles calculated by the differential motion vector calculator 141, and the surroundings. Surrounding environment information collected by the environment information collection unit 142, moving object related information analyzed by the moving object information analysis unit 143, and first and second moving objects calculated by the direction profile calculation unit 144 Based on the direction profile of the artificial neural network 150, collision risk information including a collision probability between the first moving object and the second moving object, an expected collision position, and a moving path after the collision may be predicted (S147).

According to the vehicle driving risk prediction system and method according to an embodiment of the present invention as described above, it is possible to quickly predict the risk of a collision accident between obstacles (surrounding cars or people, etc.) around the vehicle in real time while the vehicle is driving. A secondary accident of a vehicle following a collision between obstacles can be effectively prevented.

At least some of the configurations of the embodiments described above may be implemented as hardware components, software components, and/or a combination of hardware components and software components. For example, the devices, methods and components described in the embodiments may include, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate (FPGA) array), programmable logic units (PLUs), microprocessors, or any other device capable of executing and responding to instructions.

A processing device may run an operating system and one or more software applications running on the operating system. A processing device may also access, store, manipulate, process, and generate data in response to execution of software. For convenience of understanding, there are cases in which one processing device is used, but those skilled in the art know that a processing device includes a plurality of processing elements and/or a plurality of types of processing elements. It will be understood that it can include

For example, a processing device may include a plurality of processors or a processor and a controller. Also, other processing configurations are possible, such as a parallel processor. Software may include a computer program, code, instructions, or a combination of one or more of the foregoing, which configures a processing device to operate as desired or processes independently or collectively. The device can be commanded.

Software and/or data may be any tangible machine, component, physical device, virtual equipment, computer storage medium or device, intended to be interpreted by or provide instructions or data to a processing device. , or may be permanently or temporarily embodied in a transmitted signal wave. Software may be distributed on networked computer systems and stored or executed in a distributed manner. Software and data may be stored on one or more computer readable media.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer readable medium. Computer readable media may include program instructions, data files, data structures, etc. alone or in combination. Program commands recorded on the medium may be specially designed and configured for the embodiment or may be known and usable to those skilled in computer software.

Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CDROMs and DVDs, and ROMs, RAMs, and flash memories. hardware devices specially configured to store and execute program instructions, such as; Examples of program instructions include high-level language codes that can be executed by a computer using an interpreter, as well as machine language codes such as those produced by a compiler. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

As described above, although the embodiments have been described with limited examples and drawings, those skilled in the art can make various modifications and variations from the above description. For example, the described techniques may be performed in an order different from the method described, and/or components of the described system, structure, device, circuit, etc. may be combined or combined in a different form than the method described, or other components may be used. Or even if it is replaced or substituted by equivalents, appropriate results can be achieved. Therefore, other implementations, other embodiments, and equivalents of the claims are within the scope of the following claims.

10: vehicle
20, 30: moving object
100: vehicle driving risk prediction system
110: image acquisition unit
120: moving object extraction unit
130: motion vector calculator
140: accident risk prediction unit
141: differential motion vector calculator
142: surrounding environment information collection unit
143: moving object information analysis unit
144: direction profile calculator
145: collision risk information prediction unit
150: artificial neural network
160: artificial neural network learning unit
170: danger notification unit

Claims (7)

A vehicle driving risk prediction system for predicting the driving risk of a vehicle in motion,
an image acquisition unit configured to sequentially acquire images of objects located around the vehicle;
a moving object extraction unit configured to extract moving objects by analyzing the images;
a motion vector calculation unit configured to calculate motion vector profiles corresponding to motion vectors according to time for each of the moving objects;
an accident risk prediction unit configured to predict an accident risk of the vehicle by an artificial neural network based on motion vector profiles calculated for the moving objects; and
Including a risk notification unit configured to notify the driver of the vehicle of the risk of an accident when predicting the risk of an accident;
The accident risk prediction unit:
A differential motion vector profile that is a difference value between a first motion vector profile calculated for a first one of the moving objects and a second motion vector profile calculated for a second one among the moving objects is configured to be calculated. a differential motion vector calculation unit that is; and
Based on the first motion vector profile, the second motion vector profile, and the differential motion vector profile, a collision probability between the first moving object and the second moving object, a predicted collision position, and a collision by the artificial neural network Predicting collision risk information including a post-movement path; Based on the predicted collision risk information and the driving information of the vehicle, determining a countermeasure for preventing an accident of the vehicle, including braking control-related information of the vehicle and direction change-related information of the vehicle Including a collision risk information prediction unit,
The accident risk prediction unit:
a surrounding environment information collecting unit configured to collect surrounding environment information including information related to a road on which the vehicle is driving and information related to weather at a time when the vehicle is driving;
To process the images to analyze moving object related information including the type of the first moving object, the type of the second moving object, and the estimated weight or vehicle type of the first and second moving objects a moving object information analysis unit; and
A direction profile calculator configured to process the images to calculate a first direction profile representing direction information of the first moving object over time and a second direction profile representing direction information of the second moving object over time. include more,
The collision risk information prediction unit is configured to predict the collision risk information by the artificial neural network based on the surrounding environment information, the moving object related information, the first direction profile, and the second direction profile,
The surrounding environment information collection unit may include information related to a road on which the vehicle is driving, the information related to the road including whether the road is uphill, downhill, or flat and inclination information including an angle of inclination; tunnel or not; road types corresponding to straight roads, curved roads, intersections or highways; road speed limit; and information related to weather at the time the vehicle is driving, wherein the information related to weather includes rain information, snowfall information, or snowfall information.
The moving object information analyzer extracts and processes the size and manufacturer symbol of the moving object from the images acquired by the image acquisition unit, so that the types of the first moving object and the second moving object - the first moving object and the type of the second moving object includes a large car, a medium car, a ramp, a truck, a bus, a motorcycle, and a person; estimated weights of the first moving object and the second moving object; Analyzing the information related to the moving object including vehicle types of the first moving object and the second moving object - the vehicle types include manufacturers and vehicle models of the first moving object and the second moving object - configured to
The surrounding environment information and the moving object-related information are pre-analyzed before the collision risk of the moving objects occurs while the vehicle is driving, and the surrounding environment information and the moving object-related information analyzed in advance at the time of occurrence of the collision risk of the moving objects Together, a motion vector profile in a collision risk situation is additionally input to the artificial neural network, and an accident risk prediction result is output in real time from the artificial neural network,
The collision risk information prediction unit considers the predicted braking distance of the vehicle based on the predicted collision risk information and the driving information of the vehicle, and takes countermeasures for preventing accidents of the vehicle - the countermeasures include information related to braking control of the vehicle and lane A vehicle driving risk prediction system, configured to determine whether to change and information related to turning of the vehicle including a lane to change. delete delete A vehicle driving risk prediction method for predicting the driving risk of a vehicle in motion,
sequentially acquiring images of objects located around the vehicle by an image acquiring unit;
extracting moving objects by analyzing the images by a moving object extraction unit;
calculating, by a motion vector calculation unit, motion vector profiles corresponding to motion vectors of each of the moving objects according to time;
predicting, by an accident risk predictor, an accident risk of the vehicle based on motion vector profiles calculated for the moving objects by an artificial neural network; and
Informing the driver of the vehicle of the accident risk when the accident risk is predicted by the risk notification unit,
The step of predicting the accident risk is:
A difference value between a first motion vector profile calculated for a first moving object among the moving objects and a second motion vector profile calculated for a second moving object among the moving objects by the differential motion vector calculator, calculating differential motion vector profiles;
Collision between the first moving object and the second moving object is performed by the artificial neural network based on the first motion vector profile, the second motion vector profile, and the differential motion vector profile by a collision risk information prediction unit. predicting collision risk information including a probability, an expected location of a collision, and a movement path after a collision; and
Based on the collision risk information predicted by the collision risk information prediction unit and the driving information of the vehicle, accident prevention of the vehicle includes braking control related information of the vehicle and direction change related information of the vehicle. Determining countermeasures for
The step of predicting the accident risk is:
Collecting, by a surrounding environment information collection unit, surrounding environment information including information related to a road on which the vehicle is driving and information related to weather at a time point when the vehicle is driving;
The moving object information analysis unit processes the images to include the type of the first moving object, the type of the second moving object, and the estimated weight or vehicle type of the first moving object and the second moving object. Analyzing moving object related information;
A direction profile calculation unit processes the images to calculate a first direction profile representing direction information of the first moving object over time and a second direction profile representing direction information of the second moving object over time. doing; and
Predicting the collision risk information by the artificial neural network based on the surrounding environment information, the moving object related information, the first direction profile, and the second direction profile by the collision risk information prediction unit do,
The collecting of the surrounding environment information may include information related to a road on which the vehicle is driving, the information related to the road including whether the road is uphill, downhill, or flat and inclination information including an angle of inclination; tunnel or not; road types corresponding to straight roads, curved roads, intersections or highways; road speed limit; Collecting the surrounding environment information including information related to weather at the time the vehicle is driving, wherein the information related to weather includes rain information, snowfall information, or amount of snow,
Analyzing the moving object-related information includes extracting and processing the size and manufacturer symbol of the moving object from the images acquired by the image acquisition unit, so that the type of the first moving object and the second moving object - the first moving object. 1 type of the moving object and the second moving object include a large car, a mid-sized car, a ramp, a truck, a bus, a motorcycle, and a person; estimated weights of the first moving object and the second moving object; Analyzing the information related to the moving object including vehicle types of the first moving object and the second moving object - the vehicle types include manufacturers and vehicle models of the first moving object and the second moving object - including the steps of
The surrounding environment information and the moving object-related information are pre-analyzed before the collision risk of the moving objects occurs while the vehicle is driving, and the surrounding environment information and the moving object-related information analyzed in advance at the time of occurrence of the collision risk of the moving objects Together, a motion vector profile in a collision risk situation is additionally input to the artificial neural network, and an accident risk prediction result is output in real time from the artificial neural network,
The step of estimating the collision risk information may include a response measure for preventing an accident of the vehicle in consideration of a predicted braking distance of the vehicle based on the predicted collision risk information and driving information of the vehicle - the response action is information related to braking control of the vehicle. , and including information related to direction change of the vehicle including whether or not to change a lane and the lane to be changed. delete delete A computer-readable recording medium on which a program for executing the vehicle driving risk prediction method of claim 4 is recorded.

Images