KR101344066B1 - Apparatus for illumination controlling of black box using car and method thereof - Google Patents

Abstract

Disclosed is a lighting control device and a method of a vehicle black box, which automatically obtains an optimal accident image even in low light by automatically lighting a light as required during parking monitoring in a vehicle black box.
The lighting control apparatus of the disclosed vehicle black box includes an image input unit configured to receive an image acquired through a camera; A G-sensor for detecting a shock of the vehicle, detecting a movement of an object by using gravity, and outputting a motion detection signal as a driving signal; A control unit for determining whether the parking mode is the driving mode or the driving mode, controlling the storage of the image output from the image input unit in the parking mode, and generating a control signal for lighting lighting when the brightness of the image is dark; An image processor which senses brightness of the image from the image, detects a moving object, and transmits a detection result to the controller; The lighting unit for lighting the light according to the control signal for lighting the light is provided.

Description

Apparatus for illumination controlling of black box using car and method

The present invention relates to a light control device and a method of a black box for a vehicle, and more particularly, for a vehicle in which a black box for a vehicle can automatically turn on lighting as needed during parking monitoring to obtain an optimal accident image even in low light. The present invention relates to a lighting control device of a black box and a method thereof.

Recently, a black box applied to a vehicle provides a function of storing surveillance images while parking to prepare for theft or damage of the vehicle. It is obvious that such surveillance images are most importantly used to reenact the situation at the time of the accident.

In order to acquire a surveillance image, a CMOS sensor is used in a general vehicle black box. Such a CMOS sensor has a characteristic deterioration at a low light level due to its characteristics, and thus a monitoring image using a CMOS sensor in low light conditions (especially at night) is used. Even when the image was obtained, it was often impossible to read the image.

In order to solve the problem of reading a surveillance image obtained in such a low light situation, a conventionally proposed technique is disclosed in Republic of Korea Patent Application Publication No. 10-2011-0055779 (2011.5.26. Publication) (name of the invention: A method and apparatus for recording image data dynamically corresponding to an illumination environment) (hereinafter, abbreviated as "prior art").

The disclosed prior art includes an image sensor capable of detecting external image information and controlling the same width, an illuminance sensor detecting external illuminance information, and an illuminance sensitive control unit for setting an illuminance mode using image information and illuminance information, In the low situation, the NIR generator generates NIR. When the generated near infrared rays are reflected from the surrounding objects, the reflected infrared rays are detected by the image sensor and combined with the image information to increase the visibility of the image.

1. Republic of Korea Patent Publication No. 10-2011-0055779 (2011.5.26. Published) (name of the invention: a method and apparatus for recording image data by dynamically responding to the external lighting environment)

However, the prior art as described above is equipped with a near-infrared generator and a near-infrared generator for generating near-infrared at low light and an illuminance sensor for detecting the illuminance, thus complicating the configuration of the black box and improving the cost of the black box. As a result, it acted as a problem in the generalization of the black box.

In addition, in the prior art, when it is determined that the illuminance detected by the illuminance sensor is low, power consumption is severe because the near-infrared generator is unconditionally driven regardless of image capturing, which leads to insufficient power in case of accidental image capturing. In some cases, it could not acquire a surveillance image because the black box could not be driven.

Accordingly, the present invention has been made to solve the above-mentioned problems occurring in the prior art,

The problem to be solved by the present invention is to provide a lighting control device and a method of the vehicle black box to obtain an optimal accident image even in low light by automatically lighting the light as needed during parking monitoring in the vehicle black box have.

Another problem to be solved by the present invention is to provide a lighting control apparatus and method for a vehicle black box to prevent the waste of the battery by efficiently controlling the lighting during parking monitoring in the vehicle black box.

Another problem to be solved by the present invention is to enable the motion detection from the image even in the absence of lighting in software, and by using the illumination of the visible light region without using an expensive infrared camera (IR camera) existing black The present invention provides a lighting control device and a method of a vehicle black box, which minimize power consumption by using a camera mounted on a box as well as obtain the best surveillance image.

Lighting control device for a vehicle black box according to the present invention for solving the above problems,

An image input unit receiving an image acquired through a camera;

A G-sensor for detecting a shock of the vehicle, detecting a movement of an object by using gravity, and outputting a motion detection signal as a driving signal;

The driving signal output from the G sensor unit determines whether the parking mode or the driving mode, and in the parking mode, controls the storage of the image output from the image input unit, and generates a control signal for lighting lighting when the brightness of the image is dark. A control unit;

An image processor which senses brightness of an image from an image output from the image input unit, detects a moving object, and transmits a detection result to the controller;

It characterized in that it comprises a lighting unit for lighting the light in accordance with a control signal for lighting the light generated by the control unit.

In the above, the control unit

The lighting timing of the lighting and the storage timing of the image output from the image input unit may be determined based on the detection result transmitted from the image processing unit.

The image processing unit in the above

It detects the brightness of the input image to determine when to turn on the light and detects a moving object from the input image to generate a signal for determining when the image should be stored as a detection result, and transmits it to the control unit It is done.

The illumination unit may include a rotating member for adjusting an illumination angle to change an angle with a lens receiving an object.

Lighting control method of the vehicle black box according to the present invention for solving the above problems,

(a) determining a parking mode based on a driving signal output from the G sensor and detecting whether a shock occurs in the vehicle in the parking mode;

(b) detecting a movement from the captured image when the impact does not occur in the vehicle;

(c) detecting brightness from the captured image when motion is detected from the captured image or an impact occurs in the vehicle;

(d) turning on illumination when the brightness of the image is dark as a result of the brightness detection;

(e) storing the image photographed after the lighting is turned on or the image photographed before the lighting is turned on together with event information.

In the step (b), (b-1) generating an image frame from an image input from the camera, (b-2) generating a foreground image by applying an adaptive threshold value to the generated image frame, ( b-3) classifying objects by labeling the generated foreground image, and then detecting movement of the classified objects.

In the step (b-1), the Y component is extracted from the image input from the camera to generate an 8-bit gray image having a W × H resolution, and an image frame is generated by reducing the resolution through pyramidic downsampling the 8-bit gray image. It is characterized by.

In the step (b-3), the outermost rectangle (bounding box) of the classified label is defined as a region of interest (ROI), and the size of the region of interest is compared with a preset threshold to determine whether to move. When the size of the ROI is larger than a threshold, the object is determined to move.

Wherein (c) step (c-1) updating the background image using a variable weighted index moving average filter to the foreground image, (c-2) comprising the step of measuring the brightness of the updated background image It features.

In step (c-1), a background is obtained by combining a moving average filter for storing previous frames in a buffer and averaging each data with the same weight, and a weighted exponential moving average filter for averaging more recent data with a larger weight value. And updating the image.

According to the present invention, there is an advantage in that the vehicle black box automatically turns on lighting as needed during parking monitoring to obtain an optimal accident image even in low light.

In addition, according to the present invention there is an advantage that can effectively control the lighting during parking monitoring in the vehicle black box to prevent waste of the battery.

In addition, according to the present invention can be detected from the image even in the absence of lighting in software, by using the illumination of the visible light region, the camera mounted on the existing black box without using an expensive infrared camera (IR camera) By using it as it is, it has the advantage of minimizing power consumption and obtaining the best surveillance image.

1 is a block diagram of a lighting control device for a vehicle black box according to the present invention.
Figure 2 is a flow chart showing a lighting control method of a vehicle black box according to the present invention.
3 is a flowchart illustrating an embodiment of an image processing process for detecting the motion of FIG. 2.
4 is a flowchart illustrating an embodiment of a process for detecting brightness from the image of FIG. 2.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the following description, well-known functions or constructions are not described in detail since they would obscure the invention in unnecessary detail.

1 is a block diagram of a lighting control device for a vehicle black box according to the present invention, the image input unit 10, the G sensor unit 2, the control unit 30, the image storage unit 40, the image processing unit 50 and The lighting unit 60 is configured.

The image input unit 10 is a part that receives the image obtained through the camera, the G sensor unit 20 detects the impact of the vehicle, detects the movement of the object using gravity, and drives the motion detection signal It outputs a signal and transmits it to the control unit 30. The G-sensor unit 20 uses a three-axis G-sensor, the force applied to the three-axis G-sensor is expressed by the gravity and the three-dimensional vector sum, the sensor is a unit of the scalar of the x-axis, y-axis, and z-axis respectively Output to indicate the force.

The controller 30 determines whether the vehicle is in the parking mode or the driving mode by the driving signal output from the G sensor unit 20, and controls the storage of the image output from the image input unit 10 in the parking mode, and the brightness of the image. If is dark, the control signal for lighting the illumination is generated. The controller 30 determines the lighting time and the storage time of the image based on the detection result transmitted from the image processing unit 50.

The image processor 50 detects the brightness of the image from the image output from the image input unit 10, detects a moving object, and delivers the detection result to the controller 30. The image processing unit 50 detects the brightness of the input image and generates a signal for determining when to turn on the illumination and a signal for determining when the image should be stored by detecting a moving object from the input image, as a detection result. Transfer to the control unit 30.

The lighting unit 60 serves to light the lighting according to a control signal for lighting lighting generated from the control unit 30. In this case, the lighting unit 60 preferably includes a rotating member for adjusting the lighting angle so as to change the angle with the lens receiving the subject.

The image storage unit 40 stores the image input through the image input unit 10 under the control of the controller 30.

The lighting control device for a vehicle black box according to the present invention configured as described above transmits the image acquired through the camera from the image input unit 10 to the controller 30.

When the image is input, the controller 30 determines whether the current vehicle mode is a driving mode or a parking mode, and controls the storage time of the input image according to the mode determination result.

The mode determination of the vehicle is based on the driving signal output from the G sensor unit 20.

를 제외하고 횡축 성분인

와

의 합의 노름(norm)을 사용하며, 이를 수식으로 표현하면 아래의 수학식 1로 표현할 수 있다.For example, the G-sensor unit 20 uses a three-axis G-sensor, the force applied to the three-axis G-sensor is represented by the sum of gravity and three-dimensional vector, the sensor is a scalar of the x-axis, y-axis, and z-axis, respectively Output power in units. In order to judge the accident and enter the parking mode, gravity

Except for the transverse component

Wow

, Which can be represented by the following equation (1). &Quot; (1) "

)과 비교하여,

가 되면 사고(이벤트) 발생으로 판단을 한다. 아울러 상기 출력 값을 주차 모드를 판단하기 위한 임계값(

)과 비교하여, 상기

인 상태가 지속하는 시간을 카운트하여, 상기

인 상태가 일정 시간 이상 지속하면 주차 모드로 판단을 한다. 또한,

가 되면 주행 모드로 판단을 한다.Therefore, the control unit 30 is set in advance to determine the output signal and the accident (event) output from the G sensor unit 20 (

),

It is judged that an accident (event) has occurred. The output value is used as a threshold value for determining the parking mode

),

Counts the time that the in-state continues,

The parking mode is determined. Also,

It is determined that the vehicle is in the traveling mode.

As a result of the determination, in the driving mode, the video storage unit 40 is controlled in the same manner as the image storage control of the general black box so that the recording is always performed, and when the shock is detected (accident) by the G sensor unit 20. Control to record the event as a trigger.

On the other hand, if it is determined that the parking mode, the image processing unit 50 determines the lighting time and the image storage time based on the detection result transmitted after performing the image processing.

The image processor 50 detects brightness of an image from an image output from the image input unit 10, detects a moving object, and transmits a detection result to the controller 30.

Such an image processing unit 50 should be able to detect motion even at night with little lighting and memory amount required in consideration of an embedded system and little lighting. The unit of time t mentioned later is one frame, and the image processor 50 is an image processing algorithm at time t + 1.

For example, the image processor 50 extracts the Y component from the image input from the camera to generate an 8-bit gray image having a W × H resolution. Since Motion Detect does not have a large performance difference depending on the resolution, the video frame F _{t + 1} is generated by reducing the resolution to W / 2 × H / 2 through Pyramidic Downsampling. The memory size occupied by the image frame F _{t +1} is reduced to W / 2 × H / 2 × 8 bits.

Next, foreground labeling with adaptive threshold is performed using the adaptive threshold value, which is expressed by Equation 2 below.

이다. M은 마스크 영상으로 영상 O와 서브 마스크 w와 Dilation 연산을 통해 구한다. 주간에는 영상의 윤곽이 뚜렷이 나타나기 때문에 Background subtraction을 위하여 에지 영상(Fe_t+1, Be_t)를 사용하고, 조명이 없는 야간에는 노이즈가 많이 발생하고 물체가 뚜렷이 보이지 않기 때문에 노이즈가 포함된 DC 성분을 제거하고 영상의 AC 성분만 남은 영상(Fdr_{t +1}, Bdr_t)을 사용하여 물체의 윤곽을 두드러지게 한다. 이때 영상이 안정적인 주간에는 고정 임계값(Threshold)을 사용할 수 있지만, 야간에는 주변의 밝기에 따라 가변하도록 하는 적응 임계값을 사용하여야 조명이 있는 야간과 거의 없는 야간 상황을 구분하여 동작 감지를 수행할 수 있다. 따라서 야간 임계값인 ObjTh_night에 계수 k를 곱하여 사용하되, k는 영상의 밝기에 따라 가변하는 값이지만 선형적으로 변하지는 않는다. 이는 사용된 카메라와 렌즈, 센서에 따라 heuristic하며 밝기 값에 따라 몇 가지 단계로 구분하여 사용한다.Where ∋ght means Night, O is the image containing the motion detected object, Fe _{t +1} is the edge image of the input image at time t + 1, and Be _t is the background image at time t. The edge image, Fdr _{t +1} is an image from which the DC component of the input image is removed at time t + 1, and Bdr _t is an image from which the DC component of the background image is removed at time t. w is a 3 × 3 submask, w =

to be. M is a mask image and is obtained by dilation operation with image O and sub mask w. The edge of the image (Fe _{t + 1} , Be _t ) is used for background subtraction because the outline of the image appears clearly during the day, and the DC component that contains noise because there is a lot of noise and the object is not visible at night without lighting. Remove the image and use the image (Fdr _{t +1} , Bdr _t ) where only the AC component of the image is left to make the outline of the object stand out. At this time, a fixed threshold can be used during the day when the image is stable. However, at night, an adaptive threshold that can vary according to the brightness of the surroundings must be used to distinguish motion at night with little or no lighting. Can be. Therefore, ObjTh _night , which is a night threshold value, is used by multiplying the coefficient k, where k is a value that varies depending on the brightness of the image, but does not change linearly. It is heuristic depending on the camera, lens and sensor used, and it is divided into several steps according to the brightness value.

When the mask image M is classified through labeling, the outermost rectangle of the classified label is defined as a region of interest (ROI), and the size of the ROI is expressed as w × h. When w> Th _w and h> Th _{h, the} corresponding label is determined as a foreground (moving object), an event is generated, and transmitted to the controller 30.

In addition, the image processor 50 updates the background image using an Exponentially Variable Weighted Moving Average filter and measures the brightness from the updated background image.

The variable weighted exponential moving average filter is a combination of the existing moving average filter and the weighted exponential moving average filter.

The moving average filter for the previous n frames is expressed by Equation 3 below.

The moving average filter is affected by the size of the buffer because it stores previous frames in the buffer and averages each data with the same weight, and is not suitable for fast dynamic systems. The buffer requires n times the memory size of the frame size. . Larger buffers introduce time delays, increase the required memory space, and smaller ones are susceptible to noise.

Exponentially Weighted Moving Average filter is expressed by Equation 4 below.

Where α is a constant that is greater than 0 and less than 1 as the learning rate, and always has a larger weight as the latest data comes. The characteristic of the value of α is that when the value of α is large, the weight of the past data becomes low, which makes it vulnerable to noise and reduces the time delay. When α is small, the weight of the past data is increased, which improves the noise removing ability and increases the time delay. This method does not have additional memory besides the background image.

Therefore, the variable weighted exponential moving average filter combining the characteristics of the moving average filter and the weighted exponential moving average filter as described above may be expressed by Equation 5 below.

Vehicles parked in front of the black box after entering the parking mode are classified as the first moving object and classified as the foreground, but are converted to the background as time passes. Since the update delay time of the foreground and the background is the same at a single α (learning rate), the vehicle may be temporarily stopped and the vehicle may be classified again as a moving object. In order to solve this problem, when the (x, y) coordinate is the foreground, the learning rate value is changed by the coefficient k, unlike the background, thereby increasing the time delay of the foreground. The coefficient k has an initial value of 0 in the first frame in which the foreground is detected, and the value of k is gradually increased as the time to stay and reaches 1 after a certain time is updated with the same weight as the background.

Then, the brightness of the image is measured in the updated image frame B _{t +1} . The brightness at time t + 1 is the average value of all the pixel brightnesses of B _{t +1} , which is expressed by Equation 6 below.

Here, the entry conditions of night mode and day mode are as follows.

BrightTh _day ＞ BrightTh _night

Therefore, if Brightness _{t +1} <BrightTh _night , it enters night mode, and if Brightness _{t + 1} > BrightTh _day Enter the day mode.

Based on the object motion detection signal and the image brightness signal processed and provided by the image processor 50, the lighting lighting timing and the image storage timing are controlled. For example, the controller 30 controls the storage of the image output from the image input unit 10 in the parking mode, and generates a control signal for lighting lighting when the brightness of the image is dark.

The lighting unit 60 serves to light the lighting according to a control signal for lighting lighting generated from the control unit 30. The lighting unit 60 is composed of at least one LED flash. The LED flashlight consumes less power, requires less space in the drive circuit, and can be used as a continuous light source, which is enough for a black box lighting device.

The point to consider when the LED flash is embedded in the black box body is the angle of the light source. Because most of the black box is installed directly behind the windshield of the vehicle, if the light from the light source is reflected on the glass and enters the camera lens, it is impossible to capture a proper image. In addition, since the angle of the windshield is different for each vehicle, there is a lot of problems to use the fixed angle of the light source.

Therefore, when the illumination of the visible light region is installed in the black box, the structure of the lens and the illumination angle should be adjusted to avoid such a problem, in order to prevent the problem in the present invention in order to change the angle with the lens receiving the subject A rotating member for adjusting the illumination angle is provided to adjust only the illumination part differently from the angle of the lens. The rotating member may be rotated by a user's manual operation to adjust an angle. The rotating member may automatically drive the rotating member by using a driving motor and generating a driving signal for controlling the driving motor by the controller 30. It is also possible to automatically adjust the angle of 60.

The image storage unit 40 records at all times under the control of the controller 30, or records a motion detection event (motion detection) and a vehicle shock detection event as a trigger.

2 is a flowchart illustrating a lighting control method of a vehicle black box according to the present invention, in which S represents a step and illustrates an algorithm performed by software in the controller 30 of FIG. 1.

As shown therein, (a) determining whether the parking mode is generated based on the driving signal output from the G sensor, and detecting whether a shock occurs in the vehicle in the parking mode (S11 to S12); (b) detecting a movement from the captured image when the impact does not occur in the vehicle (S13); (c) detecting brightness from the captured image when motion is detected from the captured image or an impact occurs in the vehicle (S14 to S15); (d) lighting the illumination when the brightness of the image is dark as a result of the brightness detection (S16 to S17); (e) storing the image photographed after the lighting is turned on or the image photographed before the lighting is turned on together with event information (S18).

In the lighting control method of the vehicle black box according to the present invention, the parking mode or the driving mode is first determined based on the driving signal output from the G sensor in step S11. As a result of the determination, in the driving mode, the recording of the image is controlled in the same way as the image is stored in the existing black box. In contrast, in the parking mode, the control proceeds to step S12 to determine whether an impact occurs in the vehicle. The impact of the vehicle may be determined by comparing a value calculated by the well-known <Equation 1> with a threshold set in advance to determine the impact.

As a result of this determination, if no impact occurs in the vehicle, the process moves to step S13 to perform image processing for motion detection. In the image processing for motion detection, as shown in FIG. 3, an image frame is generated from an image input from a camera (S21 to S22). For example, an Y component is extracted from an image input from a camera to generate an 8-bit gray image having a W × H resolution, and an image frame is generated by reducing the resolution through pyramidic downsampling the 8-bit gray image.

을 생성하게 된다(S23). 이후 생성한 전경 영상을 레이블링(labeling)을 하여 물체를 분류한 후, 분류한 물체의 움직임을 감지하게 된다(S24). 예컨대, 분류한 레이블의 최외각 사각형(bounding box)을 관심영역(ROI: Region of interest)으로 정의하고, 관심영역 크기를 움직임 여부를 판단하기 위해 미리 설정된 임계치와 비교하여, 상기 관심영역의 크기가 임계치보다 클 경우 물체가 움직이는 것으로 판단하게 된다.Next, the foreground image (

It will generate (S23). After the object is classified by labeling the generated foreground image, the motion of the classified object is detected (S24). For example, the outermost rectangle of the sorted label is defined as a region of interest (ROI), and the size of the region of interest is compared with a preset threshold to determine whether the region of interest moves. If it is larger than the threshold, it is determined that the object is moving.

On the other hand, if the motion is detected as a result of detecting the movement of the object (S14), or if an impact occurs in the vehicle, the process moves to step S15 to perform image processing for brightness detection. As shown in FIG. 4, the image processing process for detecting the brightness is performed by updating the background image B (t + 1) using the variable weighted index moving average filter to the foreground image (S31) and updating the background image. Measure the brightness of (S32).

Here, the variable weighted index moving average filter is a combination of a moving average filter that stores previous frames in a buffer and averages each data with the same weight, and a weighted exponential moving average filter that averages with a larger weight as the latest data always comes up. Since it is implemented as in Equation 5, detailed description is omitted in order to avoid redundant description.

Then, the brightness of the image is measured in the updated image frame B _{t +1} . Brightness at the time of t + 1 is the average of all the pixel brightness of the B _{t +1,} and measure the brightness by using the aforementioned <Equation 6>.

The measured brightness values are compared to determine whether the night mode and the day mode are entered. Here, the entry conditions of the night mode and the day mode are as follows.

BrightTh _day ＞ BrightTh _night

Therefore, if Brightness _{t +1} <BrightTh _night , it enters night mode, and if Brightness _{t + 1} > BrightTh _day Enter the day mode.

If it is determined that the brightness of the image is dark (night mode), the control unit proceeds to step S17 to illuminate the light, and then, to step S18, the event occurrence information and the image are stored together. On the other hand, if it is determined that the brightness of the image is bright (weekly mode), the control unit immediately moves to step S18 without lighting, and stores the event occurrence information and the image together.

As such, the present invention determines whether the parking mode or driving mode, and when the parking mode, by determining the image storage time through the motion detection of the image by software, and automatically controls the lighting lighting by detecting the brightness of the image, at night It has the advantage of automatically lighting the lights and taking images so that the images taken at night can be accurately read later.

It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined in the appended claims and their equivalents. Of course, such modifications are within the scope of the claims.

10... Image input section
20... G sensor unit
30 ... The control unit
40 ... The image storage unit
50 ... The image processor
60 ... Illumination unit

Claims (10)

An image input unit receiving an image acquired through a camera;
A G-sensor for detecting a shock of the vehicle, detecting a movement of an object by using gravity, and outputting a motion detection signal as a driving signal;
The driving signal output from the G sensor unit determines whether the parking mode or the driving mode, and in the parking mode, controls the storage of the image output from the image input unit, and generates a control signal for lighting lighting when the brightness of the image is dark. A control unit;
An image processor which senses brightness of an image from an image output from the image input unit, detects a moving object, and transmits a detection result to the controller;
Including a lighting unit for lighting the light according to the control signal for lighting the light generated by the control unit,
The lighting unit lighting control device of a vehicle black box, characterized in that it comprises a rotating member for adjusting the angle of illumination in order to change the angle with the lens receiving the subject.
The apparatus of claim 1,
The lighting control device of the vehicle black box, characterized in that for determining the lighting time and the storage time of the image output from the image input unit based on the detection result transmitted from the image processing unit.
The method of claim 1, wherein the image processing unit
It detects the brightness of the input image to determine when to turn on the light and detects a moving object from the input image to generate a signal for determining when the image should be stored as a detection result, and transmits it to the control unit Lighting control device of the vehicle black box.
delete (a) determining a parking mode based on a driving signal output from the G sensor and detecting whether a shock occurs in the vehicle in the parking mode;
(b) detecting a movement from the captured image when the impact does not occur in the vehicle;
(c) detecting brightness from the captured image when motion is detected from the captured image or an impact occurs in the vehicle;
(d) turning on illumination when the brightness of the image is dark as a result of the brightness detection;
(e) storing the image photographed after the lighting is turned on or the image photographed before the lighting is turned on with event information,
The step (c) includes updating the background image by using the variable weighted index moving average filter on the foreground image, and (c-2) measuring the brightness of the updated background image. Lighting control method of the vehicle black box characterized in that.
The method of claim 5, wherein the step (b) comprises: (b-1) generating an image frame from an image input from the camera; and (b-2) generating an foreground image by applying an adaptive threshold value to the generated image frame. And (b-3) classifying objects by labeling the generated foreground image, and detecting the movement of the classified objects.
The method of claim 6, wherein the step (b-1) extracts the Y component from the image input from the camera to generate an 8-bit gray image, and generates an image frame by reducing the resolution through pyramidic downsampling the 8-bit gray image. Lighting control method of the vehicle black box characterized in that.
The method of claim 6, wherein the step (b-3) defines a region of interest (ROI) as the outermost bounding box of the classified label, and is set in advance to determine whether the region of interest moves. Compared with a threshold value, when the size of the region of interest is larger than the threshold value control method of the vehicle black box, characterized in that the moving object.
delete The method of claim 5, wherein the step (c-1) comprises: a moving average filter for storing previous frames in a buffer and averaging each data with the same weight, and a weighted exponential moving average filter for always averaging with a larger weight value for more recent data. Lighting control method of a vehicle black box, characterized in that for updating the background image by combining.

Images