KR102025354B1 - Danger vehicle warning device and operation method thereof - Google Patents

Abstract

According to an embodiment of the present invention, a danger vehicle warning device comprises: a camera unit for photographing a front area of a host vehicle; a detection unit for detecting at least one vehicle object and at least one lane object from an image photographed through the camera unit; an extraction unit for extracting a target vehicle object and a pair of lane objects located in a predetermined interest area in the image from the vehicle object and the lane object; and a diagnosis learning unit diagnosing a danger with respect to the target vehicle object based on a separation distance between the target vehicle object and the pair of lane objects.

Description

Dangerous vehicle warning device and its operation method {DANGER VEHICLE WARNING DEVICE AND OPERATION METHOD THEREOF}

The present application relates to a dangerous vehicle warning device, and more particularly, to a dangerous vehicle warning device and an operation method thereof for objectively diagnosing an abnormal condition of a preceding vehicle.

Recently, researches are being actively conducted to prevent an accident risk of a driving vehicle. In one of these attempts, devices for transmitting a radar signal or a laser signal to the front and receiving a reflected signal to determine a collision situation with the front vehicle have been developed.

However, these methods differ in the accuracy of collision determination when driving on straight roads or curved roads, and can only determine distance information between vehicles, and are limited in risk of accidents. .

In addition, when driving on a road (for example, a highway), the driver of the vehicle may not accurately recognize the driving speed of the front vehicle. A driver driving on a road in a situation such as nighttime has difficulty in directly recognizing a driving state (eg, a high speed driving state, a low speed driving state, or a stopping state) of a front vehicle.

Accordingly, there is a need for an apparatus capable of determining an operation state of a front vehicle more objectively and faster, and providing information on an accident risk to a surrounding vehicle and a corresponding front vehicle.

An object of the present application, a dangerous vehicle warning device and its operation that can learn the driving of the preceding vehicle, based on the image, and more quickly and objectively diagnose the abnormal state of the preceding vehicle, to warn one's own vehicle and surrounding vehicles It is an object to provide a method.

Dangerous vehicle warning apparatus according to an embodiment of the present application, the camera unit for photographing the front area of the own vehicle, the detection unit for detecting at least one vehicle object and at least one lane object from the image photographed through the camera unit, An extraction unit for extracting a target vehicle object and a pair of lane objects located in a region of interest preset in the image from the at least one vehicle object and the at least one lane object, and a separation between the target vehicle object and the pair of lane objects Based on the distance, the diagnostic learning unit for diagnosing the danger to the target vehicle object.

In an embodiment, the diagnostic learning unit accumulates the separation distance for a predetermined time and learns a reference deviation based on the accumulated separation distance.

In an embodiment, the diagnosis learning unit diagnoses the target vehicle object as a dangerous vehicle when the change in length per unit time with respect to the separation distance exceeds the reference deviation.

In example embodiments, when the length change per unit time with respect to the separation distance is less than or equal to the reference deviation, the target vehicle object according to the on-off state of the direction indicator light of the target vehicle object and the size of the separation distance. Reassess the risk for

In an embodiment, the diagnostic learning unit learns a reduction rate average of the area reduction rate of the remaining vehicle objects based on the speed information of the own vehicle, and when the area reduction rate of the target vehicle object is greater than or equal to the reduction rate average, The target vehicle object is diagnosed as a dangerous vehicle.

The driving apparatus and the non-driving region may be included in the image based on the peripheral distance between the corresponding area of the pair of lane objects and the remaining vehicle objects based on the speed information of the own vehicle. Set a value, and diagnose the front area as a dangerous area according to the number of entry and exit of the at least one vehicle object with respect to the non-driving area.

In an embodiment, a warning alarm unit which receives vehicle information on the dangerous vehicle and warns a driver of the vehicle through at least one of audio, visual and tactile information, and transmits the vehicle information through a network. And a vehicle auxiliary controller configured to generate a horn sound through the wireless communication unit and the electronic device (ECU) of the own vehicle, and to flash the lighting device at a predetermined flashing cycle.

The vehicle assistance control unit may receive steering information of the subject vehicle through a self-diagnostic apparatus and adjust the horizontal photographing angle of the camera in the direction of the steering information based on the steering information of the subject vehicle. .

The vehicle auxiliary control unit may receive the position, altitude, and inclination information of the own vehicle through a GPS module unit, an altitude sensor unit, and a gyro sensor unit, based on the position, altitude, and inclination information of the own vehicle. Adjust the vertical shooting angle of the camera.

A method of operating a dangerous vehicle warning apparatus according to an embodiment of the present application, the method comprising: photographing a front area of a subject vehicle, detecting at least one vehicle object and at least one lane object from an image of the front area, Extracting, from at least one vehicle object and the at least one lane object, a target vehicle object and a pair of lane objects located in a region of interest preset in the image; and a separation distance between the target vehicle object and the pair of lane objects. Based on the step of diagnosing whether the target vehicle object is dangerous.

In an embodiment, the step of diagnosing whether or not the risk may include accumulating one or more separation distances between the target vehicle object and the pair of lane objects for a predetermined time period for each frame. Learning the length change of each frame per unit time for the one separation distance based on the maximum separation distance of accumulated distances per frame, learning a reference deviation by averaging the length change of each frame, Determining whether the length change of the frame per unit time for any one separation distance exceeds the reference deviation after a certain time, and if the length change of the frame per unit time exceeds the reference deviation, risk the target vehicle object Diagnosing with the vehicle.

The method may include: diagnosing the target vehicle object as a normal vehicle when the length change of each frame per frame time is less than or equal to the reference deviation, and when the target vehicle object is diagnosed as a normal vehicle, a direction indicator of the target vehicle object. And re-diagnosing risk for the target vehicle object based on the on-off state and the magnitude of the one separation distance.

Dangerous vehicle warning apparatus and its operation method according to an embodiment of the present application has the effect of preventing the accident risk of the preceding vehicle in advance, by objectively and more quickly diagnose the risk of accident of the preceding vehicle.

1 is a block diagram of a dangerous vehicle warning apparatus according to an embodiment.
2 is a view illustrating an operation of the dangerous vehicle warning apparatus of FIG. 1.
3 is an embodiment illustrating a learning operation of the diagnostic learning unit of FIG. 1.
4 is another exemplary embodiment of the dangerous vehicle warning apparatus of FIG. 2.
5 is an exemplary embodiment illustrating an overlap image of the dangerous vehicle warning apparatus of FIG. 4.
6 is a block diagram of a dangerous vehicle warning apparatus according to another embodiment.
7 is a block diagram of a dangerous vehicle warning apparatus according to another embodiment.
8 is a block diagram of a dangerous vehicle warning system.
9 is a flowchart illustrating the operation of the dangerous vehicle warning apparatus according to the embodiment.
FIG. 10 is a diagnostic flowchart illustrating deviation learning of the diagnostic learning unit of FIG. 3.

Specific structural or functional descriptions of the embodiments according to the concept of the present application disclosed in the present application are only illustrated for the purpose of describing embodiments according to the concept of the present application, and the embodiments according to the concept of the present application It may be embodied in various forms and is not limited to the embodiments described in this application.

Embodiments according to the concept of the present application may be various modifications and may have various forms will be illustrated in the drawings and described in detail in the present application. However, this is not intended to limit the embodiments in accordance with the concepts of the present application to specific disclosure forms, and includes all changes, equivalents, or substitutes included in the spirit and scope of the present application.

Terms such as first or second may be used to describe various components, but the components should not be limited by the terms. The terms are only for the purpose of distinguishing one component from another component, for example, without departing from the scope of rights in accordance with the concepts of the present application, the first component may be called a second component, Similarly, the second component may also be referred to as the first component.

When a component is referred to as being "connected" or "connected" to another component, it may be directly connected to or connected to that other component, but it may be understood that other components may be present in between. Should be. On the other hand, when a component is said to be "directly connected" or "directly connected" to another component, it should be understood that there is no other component in between. Other expressions describing the relationship between components, such as "between" and "immediately between," or "neighboring to," and "directly neighboring to" should be interpreted as well.

The terminology used herein is for the purpose of describing particular example embodiments only and is not intended to be limiting of the present application. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this application, the terms "comprise" or "have" are intended to indicate that there is a feature, number, step, action, component, part, or combination thereof that is practiced, and that one or more other features or numbers are present. It should be understood that it does not exclude in advance the possibility of the presence or addition of steps, actions, components, parts or combinations thereof.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. Terms such as those defined in the commonly used dictionaries should be construed as having meanings consistent with the meanings in the context of the related art, and shall not be construed in ideal or excessively formal meanings unless expressly defined in this application. Do not.

Hereinafter, exemplary embodiments of the present application will be described in detail with reference to the accompanying drawings.

1 is a block diagram of a dangerous vehicle warning device 10 according to an embodiment, Figure 2 is an embodiment according to the operation of the dangerous vehicle warning device 10 of FIG.

1 and 2, the dangerous vehicle warning device 10 may include a camera unit 100, a detection unit 200, an extraction unit 300, and a diagnostic learning unit 400.

First, the camera unit 100 may capture an image 110 of the front region of the own vehicle 1. For example, the camera unit 100 is a fixed camera that captures a predetermined front area according to a preset angle of view, and may capture a general angle of view (about 80 degrees), more than 360 degrees. At this time, the camera unit 100 is an IP camera, an HD-SDI camera, an analog camera, a fire detection color camera, a thermal imaging camera, an HD (1920x1080, HD1080p) camera, or an IP zoom speed camera at an SD (720x486, NTSC) level resolution. And CCTV cameras, including at least one, can take a variety of forms of the image 110 or image of the vehicle (1). In addition, the camera unit 100 may further include an IR camera that can shoot in a dark environment, such as night driving.

In addition, the camera unit 100 is a PTZ (PAN, TILT, ZOOM) camera in which the angle of view is adjusted up, down, left and right or the focal length of the lens is adjusted according to the control of the vehicle auxiliary control unit 530 or the driver's operation to be described later, A function such as pan / tilt / zoom may be provided. That is, the camera unit 100 may capture an image 110 of the front region of the own vehicle 1 to photograph at least one preceding vehicle 2_1 ˜ 2_N running in the front region of the own vehicle 1. have.

According to an embodiment, the camera unit 100 is a camera manufacturer, camera model, rotation direction, shooting date and time, color space, focal length, flash, ISO speed, aperture, shutter speed, gps when taking a still image or image The data on which the back is recorded can be further obtained.

Next, the detector 200 may detect at least one vehicle object 210_1 to 210_N and at least one lane object 220_1 to 220_N from the image 110 captured by the camera unit 100. . Here, the at least one vehicle object 210_1 to 210_N and the at least one lane object 220_1 to 220_N may include coordinate information X and Y in the image. For example, the detector 200 may detect respective coordinate information of at least one vehicle and lane objects 210_1 to 210_N and 220_1 to 220_N from the image 110.

More specifically, the detector 200 may detect at least one vehicle object 210_1 to 210_N from a plurality of image objects (not shown) included in the image 110 based on the first object information stored in advance. Can be. Here, the first object information includes a vehicle object including any one of an inter-axle distance, a vehicle width, a tire width, a number of tires, an axle distance, a weight, a height, an on-off state such as a brake, and an on-off state such as a direction indicator according to a vehicle model. Information.

In addition, the detection unit 200 may detect at least one lane object 220_1 to 220_N from a plurality of image objects (not shown) included in the image 110 based on the second object information stored in advance. . Here, the second object information may be road object information including at least one of a lane, a gate, a gantry, a curb, and a central separator.

For example, the detector 200 may previously use at least one of a Haar-Like Feature HOG (Histogram of oriented Gradient), a Center-Symmetric Local Binary Pattern (CS-LBP), and a Gabor Filter Response extraction algorithm. Based on the stored first and second object information, at least one vehicle object 210_1 to 210_N and at least one lane object 220_1 to 220_N may be detected from the image 110 of the front region.

According to an exemplary embodiment, the detector 200 may detect vehicle information on the vehicle object with respect to at least one vehicle object 210_1 to 210_N diagnosed through the diagnosis learning unit 400. The vehicle information may include any one of a vehicle number, a vehicle type, and a vehicle color.

Next, the extractor 300 may include the target vehicle object 210_1 positioned in the ROI preset in the image 110 among the at least one vehicle object 210_1 to 210_N and the at least one lane object 220_1 to 220_N. A pair of lane objects 220_1 and 220_2 may be extracted. Here, the predetermined region of interest may be coordinate section information located at the center of the screen region of the image 110 captured by the camera unit 100. That is, the extractor 300 extracts coordinate information of the target vehicle object 210_1 from coordinate information of at least one vehicle object 210_1 to 210_N, and coordinate information about at least one lane object 220_1 to 220_N. Coordinate information of the remaining vehicle objects 210_2 to 210_N may be identified from.

According to an exemplary embodiment, the extractor 300 classifies the target vehicle object 210_1 and the remaining vehicle objects 210_2 to 210_N from at least one vehicle object 210_1 to 210_N based on a predetermined ROI. Can be extracted.

According to another exemplary embodiment, the extractor 300 may select and extract any one of the target vehicle object 210_1 and the remaining vehicle objects 210_2 to 210_N based on the speed information of the own vehicle 1. . In this case, the extraction unit 300 may receive the speed information of the own vehicle 1 from the self-diagnostic apparatus OBD 700 of the own vehicle 1 through the diagnostic learning unit 400.

For example, when the speed of the own vehicle 1 exceeds a preset speed, the extractor 300 may move the target vehicle object 210_1 among the target vehicle object 210_1 and the remaining vehicle objects 210_2 to 210_N. Can be extracted by selection. In addition, when the speed of the own vehicle 1 is less than the preset speed, the extractor 300 selects the remaining vehicle objects 210_2 to 210_N from the target vehicle object 210_1 and the remaining vehicle objects 210_2 to 210_N. Can be extracted.

Next, the diagnostic learning unit 400 may diagnose whether the target vehicle object 210_1 is dangerous based on the separation distance between the target vehicle object 210_1 and the pair of lane objects 220_1 and 220_2.

Here, the separation distance is the first separation distance DL between the first lane object 220_1 facing one side of the target vehicle object 210_1 and the second lane object facing the other side of the target vehicle object 210_1 ( 220_2) may include a second separation distance DR. In this case, the size of the first separation distance DL may be changed in inverse proportion to the size of the second separation distance DR.

For example, the first separation distance DL may be a first lane object 220_1 located on one side of the target vehicle object 210_1 and on the left side of the corresponding target vehicle object 210_1 among the pair of lane objects 220_1 and 220_2. It may be a vertically spaced distance between the). In addition, the second separation distance DR is perpendicular between the target vehicle object 210_1 and the second lane object 220_2 located on the right side of the target vehicle object 210_1 among the pair of lane objects 220_1 and 220_2. It may be a distance apart.

More specifically, the diagnostic learning unit 400 may learn the first and second separation distances DL and DR between the target vehicle object 210_1 and the pair of lane objects 220_1 and 220_2. That is, the diagnostic learning unit 400 may include a vertical X-axis coordinate section between one side of the target vehicle object 210_1 and the first lane object 220_1 facing the one of the pair of lane objects 220_1 and 220_2. The first separation distance DL corresponding to the learning may be learned. In addition, the diagnostic learning unit 400 includes a vertical X-axis coordinate section between the other side of the target vehicle object 210_1 and the second lane object 220_2 facing the other side of the pair of lane objects 220_1 and 220_2. The second separation distance DR corresponding to the learning may be learned.

According to one embodiment, the diagnostic learning unit 400 is based on the distance (eg, DL) of any one of the first and second distance (DL, DR), for the target vehicle object 210_1 Diagnose risk

For example, when the separation distance (eg, DL) of any one of the first and second separation distances DL and DR is smaller than the predetermined reference section or exceeds the predetermined reference section, the diagnostic learning unit 400 may diagnose the target vehicle object 210_1 as a dangerous vehicle. In addition, when any one of the first and second separation distances DL and DR is included in the predetermined reference section, the diagnostic learning unit 400 may determine the target vehicle object 210_1. Can be diagnosed as a normal vehicle.

According to another embodiment, the diagnostic learning unit 400 may accumulate the separation distance for a predetermined time and learn the reference deviation V _REF based on the accumulated separation distance. More specifically, the diagnostic learning unit 400 changes the length-per-frame length L and the reference deviation V with respect to any one of the first and second separation distances DL and DR (eg, DL). Based on the difference between _REF ), it is possible to diagnose whether or not the risk to the target vehicle object 210_1.

As illustrated in FIG. 3, the diagnostic learning unit 400 may accumulate the separation distance (eg, DL) of any one of the first and second separation distances DL and DR for each frame for a predetermined time. . The accumulated data for any one distance (eg, DL) may be stored in a memory (not shown). For example, the memory may store the frame-specific separation distances D _L1 to D _LN for any one separation distance (eg, DL).

Then, the diagnostic learning unit 400 based on the maximum separation distance (for example, D _L4 ) of the separation distance (D _L1 ~ D _LN ) for each frame stored in the memory, any one separation distance (for example, It is possible to learn the length change (L) per unit time for DL). Here, the length change L per unit time for any one separation distance (eg, DL) may be stored in the memory. That is, the memory may store the frame separation distance D _L1 to D _LN and the frame length change L ₁ to L _N-1 with respect to any one separation distance (for example, DL).

In this case, the diagnostic learning unit 400 may learn the reference deviation V _REF by averaging the length change L ₁ to L _{N-1 of} each frame per unit time. Here, the reference deviation V _REF may be an average value derived from (L ₁ + L ₂ + L ₃ ... L _N ) / N.

Then, after a certain time, when the length change per unit time (L) for any one separation distance (eg, DL) exceeds the reference deviation (V _REF ), the diagnostic learning unit 400 is a target vehicle object 210_1 may be diagnosed as a dangerous vehicle. In addition, after a certain time, when the length change per unit time (L) for any distance (eg, DL) does not exceed the reference deviation (V _REF ), the diagnostic learning unit 400 is a target vehicle object 210_1 may be diagnosed as a normal vehicle.

According to another embodiment, after the diagnostic learning unit 400 learns the reference deviation V _REF , the on / off state of the direction indicator of the target vehicle object 210_1 and any one separation distance (for example, DL). Based on), the risk of the target vehicle object 210_1 may be re-diagnosed. Here, the on-off state of the direction indicator light may mean a blinking state of the direction indicator light.

More specifically, when the target vehicle object 210_1 is diagnosed as a normal vehicle according to the reference deviation V _REF , the diagnostic learning unit 400 turns off the direction indicator light of the target vehicle object 210_1, When the separation distance (eg, DL) reaches the minimum threshold, the risk of the target vehicle object 210_1 may be re-diagnosed.

For example, when the target vehicle object 210_1 is diagnosed as a normal vehicle according to the reference deviation V _REF , the direction indicator light of the target vehicle object 210_1 is off and one of the target vehicle objects 210_1 is turned off. When the separation distance (eg, DL) is a minimum threshold of 0, the target vehicle object 210_1 may be diagnosed as a dangerous vehicle.

According to another embodiment, the diagnostic learning unit 400 may learn each object area for the target vehicle object 210_1 and the remaining vehicle objects 210_2 to 210_N according to the speed information of the own vehicle 1. have.

More specifically, when the diagnostic learning unit 400 receives the speed information of the own vehicle 1 from the self-diagnostic apparatus OBD 700, the target vehicle object 210_1 and the rest extracted through the extracting unit 300. The vehicle objects 210_2 to 210_N may be transmitted. For example, the diagnostic learning unit 400 may be connected to the self diagnosis device OBD 700 of the own vehicle 1 to receive the speed information of the vehicle 1 from the self diagnosis device OBD 700. have.

Then, the diagnostic learning unit 400 learns each object area for the target vehicle object 210_1 and the remaining vehicle objects 210_2 to 210_N according to the speed information of the own vehicle 1 and stores a memory (not shown). Can be stored in For example, when the speed of the own vehicle 1 is 60 KM / H or less, the diagnostic learning unit 400 may include the first object area corresponding to the target vehicle object 210_1 and the remaining vehicle objects (eg, 210_2 to 210_N). 2nd to Nth object areas corresponding to) can be learned. That is, when the speed of the own vehicle 1 is less than the predetermined speed, the diagnostic learning unit 400 is applied to the target vehicle object 210_1 and the remaining vehicle objects 210_2 to 210_N extracted through the extractor 300. You can learn about each object area.

In addition, the diagnostic learning unit 400 may estimate a reference reduction rate by learning a reduction rate for each object area of the remaining vehicle objects 210_2 to 210_N for a predetermined time. In this case, the diagnostic learning unit 400 may learn a reduction rate of the object area of the target vehicle object 210_1. Then, based on the difference between the reduction rate of the object area of the target vehicle object 210_1 and the reference reduction rate according to the remaining vehicle objects 210_2 to 210_N, it is possible to diagnose whether the risk to the target vehicle object 210_1. .

For example, the diagnostic learning unit 400 risks the target vehicle object 210_1 when the reduction rate of the area of the target vehicle object 210_1 is greater than or equal to a predetermined reduction rate according to the remaining vehicle objects 210_2 to 210_N. Diagnosis can be made by vehicle.

In the present application, the dangerous vehicle warning device 10 may detect a target vehicle object 210_1 and a pair of lane objects 220_1 and 220_2 located in a predetermined region of interest from the image 110 photographed through the camera unit 100. Can be extracted. Then, the dangerous vehicle warning device 10 may diagnose whether or not the danger for the target vehicle object 210_1 is based on the separation distance between the target vehicle object 210_1 and the pair of lane objects 220_1 and 220_2. . That is, the dangerous vehicle warning device 10 may objectively diagnose whether the danger of the preceding vehicle is dangerous, based on the deviation of the driving of the preceding vehicle ahead of the own vehicle 1. Accordingly, it is possible to prevent the accident risk of the preceding vehicle in advance.

4 is another exemplary embodiment according to the operation of the dangerous vehicle warning device 10 of FIG. 2, and FIG. 5 is an exemplary embodiment illustrating an overlap image 111 of the dangerous vehicle warning device 11 of FIG. 4.

4 and 5, the dangerous vehicle warning device 11 may include a camera unit 100, a detection unit 200, an extraction unit 300, and a diagnostic learning unit 400.

Hereinafter, duplicate descriptions of the camera part 100, the detection part 200, the extraction part 300, and the diagnostic learning part 400 of the same member number described with reference to FIGS. 1 to 4 will be omitted.

First, the diagnostic learning unit 400 receives a pair of lane objects 220_1 and 220_2 and the remaining vehicle objects 210_2 to 210_N through the extractor 300 based on the speed information of the own vehicle 1. Can be. Next, the diagnostic learning unit 400 may learn a driving area located in the front area of the own vehicle 1 based on the pair of lane objects 220_1 and 220_2. In addition, the diagnostic learning unit 400 may learn the non-driving area based on the separation distance between the driving area and the remaining vehicle objects 210_2 to 210_N.

In this case, the diagnostic learning unit 400 may set a driving area and a non-driving area on the image 110 captured by the camera unit 100. Here, the driving area and the non-driving area may be images overlapping the image 110. That is, the camera unit 100 may include an image processor (not shown) that overlaps the driving area and the non-driving area in the captured image 100.

Next, as illustrated in FIG. 5, the detector 200 may perform the non-driving area (Non-) for a predetermined time from the image 111 in which the driving area and the non-driving area overlap. At least one vehicle object 210_1 to 210_N entering and leaving the driving area may be detected. In this case, the extractor 300 may count the number of times at least one vehicle object 210_1 to 210_N enters and leaves the non-driving area for a predetermined time.

According to an exemplary embodiment, the diagnostic learning unit 400 may diagnose whether a danger is caused in the front area of the own vehicle 1 based on the number of times of entering and leaving the non-driving area for a predetermined time. have. For example, when the number of times of entering and leaving the non-driving area exceeds a predetermined number of times for a predetermined time, the diagnostic learning unit 400 moves the front area of the subject vehicle 1 to the dangerous area. Diagnosis can be made. In addition, when the number of times of entering and leaving the non-driving area for a predetermined time is less than or equal to a predetermined number of times, the diagnosis learning unit 400 may diagnose the front area of the subject vehicle 1 as a normal area. .

6 is a block diagram of a dangerous vehicle warning device 11 according to another embodiment.

Referring to FIG. 6, the dangerous vehicle warning device 11 includes a camera unit 100, a detection unit 200, an extraction unit 300, a diagnostic learning unit 400, an alarm alarm unit 510, and a wireless communication unit 520. ) And the vehicle assistance control unit 530.

Hereinafter, duplicate descriptions of the camera part 100, the detector 200, and the diagnostic learning unit 400 having the same member numbers described with reference to FIGS. 1 to 5 will be omitted.

First, the warning alarm unit 510 receives the vehicle information on the dangerous vehicle diagnosed through the diagnostic learning unit 400 in the form of a warning alarm through at least one of audio, visual and tactile information. Can be output to the driver.

More specifically, the warning alarm unit 510 receives visual information about a dangerous vehicle received from the diagnostic learning unit 400 or stored in a memory (not shown), for example, a warning screen or a red screen through a display device (not shown). You can print For example, a display device (not shown) may be a liquid crystal display, a thin film transistor-liquid crystal display, an organic light-emitting diode, or a flexible display. It may include at least one of a 3D display. In addition, the warning alarm unit 510 outputs audible information about a dangerous vehicle received from the diagnostic learning unit 400 or stored in a memory (not shown) through a sound output module (not shown). can do. For example, the sound output module (not shown) may include a speaker, a buzzer, a piezo sensor, or the like. In addition, the warning alarm unit 510 may output the tactile information on the dangerous vehicle, for example, a vibration of a certain period through an alarm device (not shown). For example, the alarm device (not shown) may be vibration means for outputting a signal for notifying occurrence of an event in a form other than an audio signal or a video signal. For example, the warning device may output tactile information in the form of a vibration.

Next, the wireless communication unit 520 connects the vehicle information on the dangerous vehicle diagnosed through the diagnostic learning unit 400 through the road management server 50 or the wireless network 30 through the wireless network 30. Can be transmitted to the fields 2_1 to 2_N.

For example, the wireless communication unit 520 may include a wireless LAN (WLAN), a digital living network alliance (DLNA), a wireless broadband (Wibro), a WiMAX (World Interoperability for Microwave Access: Wimax), and an HSDPA (High Speed). Downlink Packet AccesB), High Speed Uplink Packet AccesB (HSUPA), IEEE 802.16, Long Term Evolution (LTE), Long Term Evolution-Advanced (LTE-A), Wireless Mobile Broadband Service: WMBB) may be included.

In addition, the wireless communication unit 520 may include Bluetooth, Radio Frequency Identification (RFID), Infrared Data Association (IrDA), Ultra Wideband (UWB), ZigBee, and Near Field Communication (NFC). ), Ultrasonic communication (Ultra Sound Communication (USC)), visible light communication (Visible Light Communication (VLC), Wi-Fi (Wi-Fi), may include at least one of Wi-Fi Direct (Wi-Fi Direct).

Next, the vehicle auxiliary control unit 530 automatically generates an horn sound to the outside through the electronic device ECU of the own vehicle 1 according to the dangerous vehicle diagnosed through the diagnostic learning unit 400. The lighting device (not shown) of 1) may be blinked at a predetermined cycle. For example, the vehicle assistance control unit 530 may communicate with the electronic device ECU of the vehicle 1 to control an illumination device (not shown) and a horn device (not shown) of the vehicle 1. .

According to an embodiment, the vehicle assistance control unit 530 may perform a vehicle driving assistance function through the electronic device ECU of the own vehicle 1 according to the dangerous vehicle diagnosed by the diagnostic learning unit 400.

Here, the vehicle driving assistance function may include an autonomous emergency braking (AEB), a forward collision avoidance function (FCW: foward collision warning), a lane departure warning function (LDW: lane departure warning), and a lane maintenance assistance function (LKA: Lane Keeping Assist (SAS), Speed Assist System (SAS), Traffic Sign Recognition (TSR), Adaptive High Beam Assist (HBA), Blind Spot Monitoring (BSD: Blind Spot) Detection), Autonomous Emergency Steering (AES), Curve Speed Warning System (CSWS), Adaptive Cruise Control (ACC), Smart Parking Assist System (SPAS) ), At least one of a traffic jam support (TJA) and an around view monitor (AVM).

For example, the vehicle assistance control unit 530 receives an automatic emergency steering function (AES: Autonomous Emergency Steering) and a speed support function (SAS) when receiving vehicle information on a dangerous vehicle diagnosed through the diagnostic learning unit 400. Speed Assist System (1) first, then forward collision avoidance (FCW: Foward Collision Warning), Lane Departure Warning (LDW), Lane Keeping Assist (LKA) Then, the autonomous emergency braking (AEB) can be activated. Thereafter, the vehicle auxiliary control unit 530 may induce the vehicle to stop by operating an autonomous emergency braking (AEB) function.

According to an embodiment, the vehicle assistance control unit 530 may horizontally photograph the camera unit 100 based on steering information of the own vehicle 1 transmitted through the self-diagnostic apparatus OBD 700 of the own vehicle 1. The angle can be adjusted in the direction of the steering information.

For example, the vehicle assistance controller 530 may receive steering information of the own vehicle 1 through the self-diagnostic apparatus OBD 700 of the own vehicle 1. Then, when the steering information of the own vehicle 1 is left turn, the vehicle assistance control unit 530 may adjust the horizontal photographing angle of the camera unit 100 to the left direction. Here, the horizontal photographing angle of the camera unit 100 may correspond to the rotation angle of the steering information of the own vehicle 1.

7 is a block diagram of a dangerous vehicle warning device 12 according to another embodiment.

Referring to FIG. 7, the dangerous vehicle warning device 12 may include a camera unit 100, a detection unit 200, a diagnostic learning unit 400, an alarm alarm unit 510, a wireless communication unit 520, and a vehicle assistance control unit 530. The GPS module unit 610 may further include an altitude sensor unit 620 and a gyro sensor unit 630.

Hereinafter, the camera unit 100, the detector 200, the diagnostic learning unit 400, the alarm alarm unit 510, the wireless communication unit 520, and the vehicle assistance control unit 530 of the same member number described with reference to FIGS. 1 to 6 will be described. Omitted duplicates of) are omitted.

Here, the GPS module unit 610 receives the position information of the own vehicle 1, the altitude sensor unit 620 detects the altitude information of the own vehicle 1, and the gyro sensor unit 630 is the own vehicle ( Tilt information of 1) can be detected.

According to an embodiment, the vehicle assistance control unit 530 may adjust the vertical photographing angle of the camera unit 100 based on the inclination information about the driving road. More specifically, the vehicle assistance control unit 530 may receive the position, altitude, and tilt information of the own vehicle 1 through the GPS module unit 610, the altitude sensor unit 620, and the gyro sensor unit 630. . Then, the vehicle assistance controller 530 may estimate slope information on the driving road based on the position, altitude, and slope information of the own vehicle 1. In this case, the vehicle assistance controller 530 may adjust the vertical photographing angle of the camera unit 100 based on the inclination information about the driving road.

For example, when the inclination information on the driving road is estimated at an inclination angle of 20 degrees down according to the position, altitude, and inclination information of the own vehicle 1, the vehicle assistance control unit 530 is directed downward of 20 degrees. The vertical photographing angle of the camera unit 100 can be adjusted. Here, the vertical photographing angle may correspond to the inclination information on the driving road.

8 is a block diagram of a dangerous vehicle warning system 1000.

1 to 8, the dangerous vehicle warning system 1000 may include a plurality of dangerous vehicle warning devices 10_1 to 10_N and a road management server 50.

First, the plurality of dangerous vehicle warning devices 10_1 to 10_N may communicate with each other through the wireless communication unit 520. More specifically, at least one of the dangerous vehicle warning devices 10_1 ˜ 10_N includes the remaining dangerous vehicle warning devices connected to the network 30 through the wireless communication unit 520. It can communicate with (10_2 ~ 10_N). That is, one dangerous vehicle warning device 10_1 may relay vehicle information diagnosed as a dangerous vehicle by the diagnostic learning unit 400 to the remaining dangerous vehicle warning devices 10_2 to 10_N through the wireless communication unit 520. have.

In addition, the plurality of dangerous vehicle warning devices 10_1 to 10_N may communicate with the road management server 50 through the network 30. For example, the plurality of dangerous vehicle warning devices 10_1 to 10_N may communicate with the road management server 50 by accessing a network 30 connected through each wireless communication unit 520.

Here, the plurality of dangerous vehicle warning devices 10_1 to 10_N may be implemented as a computer, a portable terminal, or a television to enable data transmission / reception through the network 30. For example, a computer includes a laptop, a desktop, a laptop, etc. capable of short-range wireless communication, and a portable terminal is, for example, a wireless communication device that ensures portability and mobility, and is a personal communication system (PCS). ), Global System for Mobile communications (GSM), Personal Digital Cellular (PDC), Personal Handyphone System (PHS), Personal Digital Assistant (PDA), International Mobile Telecommunication (IMT) -2000, Code Division Multiple Access (CDMA) 2000, It may include all kinds of handheld based wireless communication devices such as W-CDMA (W-Code Division Multiple Access), Wibro (Wireless Broadband Internet) terminal, smart phone, and the like. In addition, the television may include an Internet Protocol Television (IPTV), an Internet Television (Internet Television), a terrestrial TV, a cable TV, or the like.

Next, the network 30 includes various services existing in the TCP / IP protocol and its upper layers, namely, Hyper Text Transfer Protocol (HTTP), Telnet, File Transfer Protocol (FTP), Domain Name System (DNS), and SMTP (Simple). A global open computer network architecture that provides Mail Transfer Protocol (SMP), Simple Network Management Protocol (SNMP), Network File Service (NFS), and Network Information Service (NIS). In addition, the network 30 may provide an environment in which at least one or more dangerous vehicle warning devices 10_1 to 10_N may be connected to the road management server 50. Meanwhile, the network 30 may be a wired or wireless internet, or may be a core network integrated with a wired public network, a wireless mobile communication network, or a portable internet.

Next, the road management server 50 receives the vehicle information on the dangerous vehicle from any one of the dangerous vehicle warning device (for example, 10_1), the remaining dangerous vehicle warning devices (for example, 10_2 ~ 10_N) Can be relayed to. That is, any one of the dangerous vehicle warning devices 10_1 to 10_N of the dangerous vehicle warning devices (for example, 10_1) is connected to the remaining dangerous vehicle warning devices (for example, 10_2 to 10_2 through the road management server 50). 10_N) may transmit vehicle information about the dangerous vehicle.

 In addition, the road management server 50 obtains the state information of the road through the linkage with the road traffic situation server managed by the government agency or the associated agency, such as the road traffic situation room for managing road traffic, a plurality of dangerous vehicle warning device Fields 10_1 to 10_N.

For example, the state information of the road may include traffic traffic conditions of the road such as speed, traffic volume, travel time, waiting length, congestion degree, section prediction average speed, section prediction passing time, section speed trend, and the like. Traffic control information indicating the traffic situation, control location, type of control, control object, control time, traffic control information indicating the control situation of the traffic such as the construction of the pre-planned roads, events, emergency situation information, emergency situation handling tips, etc. It may include emergency situation information indicating unforeseen road conditions and road condition information indicating a road condition according to a sudden situation such as weather conditions, road conditions, rainfall / snowfall, and water level.

At this time, the road management server 50 may update the vehicle information on the dangerous vehicle and the state information of the road in the database 51. Here, the database 51 classifies, stores and manages the vehicle information on the dangerous vehicle updated from the road management server 50 and the state information of the road into a database (DB), and stores Oracle and Infomix. , Object-oriented database management system (OODBMS) such as Sybase, DB2, or object-oriented database management system (OODBMS) such as Gemston, Orion, O2, etc. It can be implemented and has the appropriate fields to achieve its function. In the present invention, the road management server 50 and the database 51 is shown as a configuration that includes, but is not limited to this, the road management server 50 and the database 51 may be an independent configuration.

9 is an operation flowchart of the dangerous vehicle warning apparatus 10 according to the embodiment.

1 and 9, in operation S110, the camera unit 100 may capture a front area of the subject vehicle 1.

Next, in operation S120, the detector 200 may detect at least one vehicle object 210_1 to 210_N and at least one lane object 220_1 to 220_N from the image 110 captured by the camera unit 100. Can be detected.

Next, in operation S130, the extractor 300 may include the target vehicle object located in the ROI preset in the image 110 from the at least one vehicle object 210_1 to 210_N and the at least one lane object 220_1 to 220_N. 210_1 and a pair of lane objects 220_1 and 220_2 may be extracted.

Thereafter, the diagnosis learning unit 400 may diagnose whether the target vehicle object 210_1 is dangerous based on the separation distance between the target vehicle object 210_1 and the pair of lane objects 220_1 and 220_2.

FIG. 10 is a diagnostic flowchart illustrating deviation learning of the diagnostic learning unit 400 of FIG. 3.

1, 3, and 10, in step S210, the diagnostic learning unit 400 may include a first and second separation distance DL between a target vehicle object 210_1 and a pair of lane objects 220_1 and 220_2. , DR) may be accumulated for a predetermined time for each frame.

Then, in step S220, the diagnostic learning unit 400, based on the maximum separation distance (for example, D _L4 ) of the accumulated distance separation distance (D _L1 ~ D _LN ) for each frame, any one separation distance ( For example, the length change L per unit time for DL) can be learned.

At this time, in step S230, the diagnostic learning unit 400 may learn the reference deviation (V _REF ) by averaging the length change (L ₁ ~ L _N-1 ) for each frame per unit time.

Then, in step S240, when the length change per unit time (L) for any one distance (for example, DL) after a certain time exceeds the reference deviation (V _REF ), the diagnostic learning unit 400 May diagnose the target vehicle object 210_1 as a dangerous vehicle.

At this time, in step S250, the vehicle assistance control unit 530 automatically generates an horn sound through the electronic device ECU of the own vehicle 1, and presets an illumination device (not shown) of the own vehicle 1. Can blink in cycles. At this time, the warning alarm unit 510 receives the vehicle information on the target vehicle object 210_1 diagnosed as a dangerous vehicle through the diagnostic learning unit 400 through at least one of audio, visual and tactile information. It can output to the driver of the own vehicle 1 in the form. In addition, the wireless communication unit 520 may transmit vehicle information about the target vehicle object 210_1 to the road management server 50 or the surrounding vehicles 2_1 to 2_N through the wireless network 30.

On the other hand, in step S260, when the length change per unit time (L) for any one distance (for example, DL) after a certain time does not exceed the reference deviation (V _REF ), the diagnostic learning unit 400 May diagnose the target vehicle object 210_1 as a normal vehicle.

Then, in step S270 and S280, based on the reference deviation (V _REF ), when the target vehicle object 210_1 is diagnosed as a normal vehicle, the diagnostic learning unit 400 is a direction indicator of the target vehicle object 210_1. Based on the on state and any one separation distance (eg, DL), it is possible to re-diagnose the risk on the target vehicle object 210_1.

Thereafter, when the target vehicle object 210_1 is re-diagnosed as a dangerous vehicle, the vehicle assistance control unit 530 automatically generates an horn sound through the electronic device ECU of the own vehicle 1 in the same manner as in step S250. In addition, the lighting apparatus (not shown) of the own vehicle 1 may be blinked at a predetermined cycle.

Although the present application has been described with reference to one embodiment shown in the drawings, this is merely exemplary, and it will be understood by those skilled in the art that various modifications and equivalent other embodiments are possible. Therefore, the true technical protection scope of the present application will be defined by the technical spirit of the appended claims.

10: Dangerous Vehicle Warning System
100: camera unit
200: detector
300: extraction unit
400: diagnostic learning department
510: warning alarm
520: wireless communication unit
530: vehicle auxiliary control unit
610: GPS module unit
620: altitude sensor unit
630: gyro sensor
1000: Dangerous Vehicle Warning System

Claims (12)

A camera unit for photographing a front area of the own vehicle;
A detector for detecting at least one vehicle object and at least one lane object from an image captured by the camera unit;
An extraction unit configured to extract a pair of lane objects and a target vehicle object located in a region of interest preset in the image among the at least one vehicle object and the at least one lane object; And
A diagnosis learning unit for diagnosing whether the target vehicle object is dangerous based on a distance between the target vehicle object and the pair of lane objects;
The ROI is coordinate section information located at the center of the screen region of the image photographed by the camera unit.
The extractor extracts coordinate information of the at least one vehicle object according to the coordinate section information, checks whether the coordinate information of the at least one vehicle object is located in the ROI, and detects the target information from the at least one vehicle object. Dangerous vehicle warning device for extracting vehicle objects. The method of claim 1,
The diagnostic learning unit accumulates the separation distance for a predetermined time, based on the accumulated separation distance, dangerous vehicle warning device for learning a reference deviation. The method of claim 2,
And the diagnostic learning unit diagnoses the target vehicle object as a dangerous vehicle when the change in length per unit time with respect to the separation distance exceeds the reference deviation. The method of claim 2,
The diagnostic learning unit, when the length change per unit time with respect to the separation distance is less than the reference deviation,
According to the on-off state of the direction indicator light of the target vehicle object and the size of the separation distance, the dangerous vehicle warning device for re-diagnosing the danger to the target vehicle object. The method of claim 1,
The diagnostic learning unit learns the average of the reduction rate of the area reduction rate of the remaining vehicle objects based on the speed information of the own vehicle,
And a dangerous vehicle warning device for diagnosing the target vehicle object as a dangerous vehicle when the area reduction rate of the target vehicle object is greater than or equal to the reduction rate average. The method of claim 1,
The diagnostic learning unit sets a driving area and a non-driving area in the image based on the speed information of the own vehicle.
And a vehicle warning device for diagnosing the front area as a danger area according to the number of entry and exit of the at least one vehicle object to the non-driving area. The method of claim 1,
A warning alarm unit receiving vehicle information on the dangerous vehicle and warning the driver of the vehicle through at least one of audio, visual and tactile information;
A wireless communication unit transmitting the vehicle information through a network; And
And a vehicle auxiliary controller configured to generate a horn sound through the electronic device (ECU) of the own vehicle, and flash the lighting device at a predetermined flashing cycle. The method of claim 7, wherein
The vehicle auxiliary control unit receives steering information of the own vehicle through a self-diagnostic apparatus,
And a vehicle warning device for adjusting the horizontal photographing angle of the camera in the direction of the steering information based on the steering information of the own vehicle. The method of claim 7, wherein
The vehicle auxiliary control unit receives the position, altitude and inclination information of the vehicle through a GPS module unit, an altitude sensor unit, and a gyro sensor unit.
Dangerous vehicle warning device for adjusting the vertical shooting angle of the camera based on the position, altitude and inclination information of the own vehicle. As a dangerous vehicle warning device,
Photographing a front area of the subject vehicle;
Detecting at least one vehicle object and at least one lane object from the image of the front region;
Extracting, from the at least one vehicle object and the at least one lane object, a target vehicle object and a pair of lane objects located in a region of interest preset in the image; And
Diagnosing whether the target vehicle object is dangerous based on a separation distance between the target vehicle object and the pair of lane objects;
The ROI is coordinate section information located at the center of the screen region of the image photographed through the camera unit.
An extracting unit extracts coordinate information of the at least one vehicle object according to the coordinate section information, checks whether the coordinate information of the at least one vehicle object is located in the ROI, and detects the target vehicle from the at least one vehicle object. Operation method of dangerous vehicle warning device for extracting objects. The method of claim 10,
Diagnosing whether the risk is,
Accumulating one of the first and second separation distances between the target vehicle object and the pair of lane objects for a predetermined time;
Learning a length change per unit time for any one separation distance based on the maximum separation distance among the accumulated separation distances;
Learning a reference deviation by averaging the change in length per unit time;
Determining whether a change in length per unit time with respect to any one separation distance exceeds the reference deviation after a predetermined time; And
And diagnosing the target vehicle object as a dangerous vehicle when the change in length per unit time exceeds the reference deviation. The method of claim 11,
Diagnosing the target vehicle object as a normal vehicle when the length change per unit time is less than or equal to the reference deviation; And
If the target vehicle object is diagnosed as a normal vehicle, re-diagnosing whether the target vehicle object is dangerous based on the on-off state of the direction indicator light of the target vehicle object and the size of any one separation distance; Operation method of a dangerous vehicle warning device comprising.

Images