RU2618863C1 - Warning device of possible collision with vehicle moving on wrong lane - Google Patents

Abstract

FIELD: physics.

SUBSTANCE: warning device of a possible collision with the vehicle moving on the wrong lane, mounted on the vehicle and comprising a video recorder connected to the image analyzing unit, which is connected to the data storage device. The image analyzing unit is configured to recognize the contour of the ahead moving vehicle and to determine the area value of the selected contour, which is an output unit signal, the data storage device performs storing the input values and outputting the value, which has been on the device input Δt seconds (0,001≤Δt≤0.5) ago relative to the current time. The output of the image analyzing unit and the output of the data storage unit are connected to the relation unit unputs, the output of which is connected to the first comparator input, the threshold unit is connected to the second comparator input, the comparator output is connected to the alarm unit.

EFFECT: advance warning the driver of a possible collision.

3 dwg

Description

The invention relates to the field of automation and can be used to create equipment that serves to prevent collisions of cars and other vehicles.

A well-known automobile registrar of traffic accidents [1], which performs the determination of the occurrence of traffic accidents (accident). The disadvantage of using such a registrar is that it does not prevent accidents, in particular, it does not signal the danger of a collision of vehicles.

Closest to the proposed system (prototype) is a system for identifying vehicles that illegally invaded a dedicated lane. This system comprises a video recorder mounted on a patrol vehicle connected to a video buffer of a data storage device, an image analysis unit connected to a video recorder, a data storage device and a decision and fixing module, connected to a video recorder and a control and initialization unit connected to a data storage device, which also contains a buffer of recognized numbers. The video recorder is configured to form a video image and individual photo frames of the front near field of view with a resolution sufficient to recognize road markings and signs, as well as vehicles ahead and their data; the image analysis unit is capable of analyzing the video image and individual photo frames, with the allocation and recognition of road markings and road signs, vehicles ahead and their data, as well as with the possibility of transmitting the recognized data to the decision-making and fixing module; the control and initialization unit is configured to obtain user-selected external conditions for determining an unlawful intrusion into a designated lane, namely, conditions for automatic analysis or manual analysis, as well as the ability to transfer user-selected external conditions to a decision and fixing module, which is configured to determine a fact the movement of a patrolling vehicle in a selected lane automatically by analyzing predefined coordinates or road markings and road signs indicating a designated lane, as well as with the possibility of detecting the fact of an unlawful intrusion of a vehicle ahead in front of a designated lane, or automatically by analyzing a set of signs, including preset coordinates or road markings and road signs indicating permitted / forbidden sections of a dedicated vehicle lanes, state registration number of the vehicle in front, indicating its type (private or not), light signal left or right turn indicating intention to the vehicle in front off the allocated band at its portion that is feasible to rebuild or manually by the driver of the vehicle patrolling signal; the video buffer and the buffer of recognized numbers are made with the possibility of cyclical recording in video data, individual data, including time stamps, and metadata coming from the DVR, as well as from the blocks and modules of the system [2].

The disadvantage of this system is that it recognizes the possibility of a collision based on the coordinates of the road marking and signs. However, there are no clear criteria for predicting the possibility of a collision some time before the collision. In this regard, early warning of a possible collision of vehicles may be difficult.

The proposed warning device about a possible collision with a vehicle moving in the oncoming lane, eliminates these disadvantages of the prototype.

The operation of the device is based on the following approach.

Consider vehicles TC1 and TC2 moving towards each other with speeds V ₁ and V _2, respectively. The linear dimensions of TC1, i.e. width and height, we take equal to A and B, respectively. The distance between TC1 and TC2, which is a variable, is denoted as D (Fig. 1).

Let a video recorder be installed on TC2, on which the linear dimensions of the image of TC1, i.e. width and height, have values a and b respectively. The distance from the objective lens to the screen in the TC2 dashcam is taken equal to d.

When objects approach each other, the linear dimensions of the TC1 image on the screen of the DVR increase.

From the likeness of triangles formed by the base and height of TC1 and similar elements of the image of its image, it follows that

и

.

and

.

Then

и

.

and

.

Denote c1 = A⋅d and c2 = B⋅d. Then

Since the values c1 and c2 are constant, the linear dimensions a and b of the image of the image of TC1 on the screen are inversely proportional to the distance D between TC1 and TC2.

The distance D between TC1 and TC2 over time t is determined by the speed V _{sb of} their approach according to the formula:

where D ₀ is the distance between TC1 and TC2 at time t = 0.

The convergence rate of vehicles can be calculated by the formula:

V _sb = V ₁ + V ₂ .

It is obvious that at the moment of collision TC1 and TC2, i.e. at t = t _article , the distance between TC1 and TC2 will be 0, i.e. D (t _article ) = 0. Then it follows from formula (2) that

D ₀ -V _sb ⋅t _st = 0, and

D ₀ = V _sb ⋅t _st

We substitute the resulting expression in formula (2) and after the conversion we get:

Formula (1), the dependence of the linear dimensions of the image of TC1 on the screen on the distance between TC1 and TC2, taking into account dependence (3), can be represented as:

Suppose that the approach velocity in a short time changes insignificantly, i.e. V _sb = const. Then the formula (4) of the dependence of the linear dimensions on the screen when TC1 and TC2 approach each other will take the form:

.Where

.

Assume that the measurement of the linear dimensions of the TC1 image on the screen occurs at equal time intervals Δt. Then the ratio F (t _i ) of the linear dimensions of the image TC1 on the screen at the current time t _i to the previous t _i-1 in accordance with expression (5) will be equal to

With the approach of objects a (t _i ) is greater than a (t _i-1 ), respectively, F (t _i ) changes from 1 to an infinitely large number. In FIG. 2 shows a series of dependency curves F (t _i ) for different values of t _article and Δt = 0.1.

Suppose that at time t = t _pr there is an early warning of a collision of vehicles TC1 and TC2. From formula (5) it follows that at this moment the linear dimensions of TC1 on the screen of the DVR will be equal to

.

.

At the same time, it follows from formula (6) that the ratio of the current value of the linear dimensions of the image of TC1 on the screen and the previous one will look like:

As can be seen from formula (7), the value of F _CR does not depend on the speed of approach of the vehicles, nor on the dimensions of the vehicle moving towards, but is determined only by the period of time for the driver to be warned in advance of a collision t _st -t _CR and the period Δt of the measurement of the linear relation the image size of the TC1 image on the screen of the DVR at the current and previous time.

Therefore, by calculating the ratio F (t _i ) of the current linear dimensions of the image of TC1 on the screen of the DVR to the previous ones and comparing this ratio with the ratio for the moment of warning of a collision F _pr , it is possible to warn the driver in advance of a dangerous situation and thereby reduce the likelihood of an accident.

For example, if the time interval for the driver’s early warning of a collision is chosen to be t _st -t _pr = 3 s, and the linear dimensions are measured on the screen of the DVR after Δt = 0.5 s, then at the moment of the warning about the collision, the ratio of the current linear dimensions of the image prior TC1 will be at least F _ave = 1 + 0.5 / 3 = 1 + 0.167 = 1.167 . In this case, a possible collision should be signaled if F (t _i ) is equal to or exceeds the value of 1.167.

If the program for processing signals from the DVR provides for the allocation of the contour of the image of a car moving in the oncoming lane, then calculate the area S (t _i ) of this circuit and measure the ratio of the current value of the area of the circuit with its previous value S (t _i-1 ):

,

,

it is quite obvious that this ratio of areas will be in a quadratic dependence with respect to the value of t _st -t _ave

Thus, if you constantly measure the ratio of the linear dimensions or the area of the contours of the image of the car on the screen of the DVR moving in the oncoming lane, and comparing it with the set value, you can warn the driver in advance of a possible collision with him and thereby prevent an accident.

The invention is illustrated by the following graphic material:

- FIG. 1 is a diagram of the movement of vehicles;

- FIG. 2 - a series of dependency curves F (t _i ) for different values of t _article and Δt = 0.1.

- FIG. 3 - functional diagram of the device.

To achieve a technical result, which consists in increasing the effectiveness of a warning about a collision of vehicles moving in an oncoming lane, a device containing a video recorder connected to an image analysis unit that is connected to a data storage device is additionally introduced with a relationship unit, a comparison unit, a threshold level unit , a signaling unit, wherein the image analysis unit is configured to recognize the contour of the vehicle in front and determine the value the area of the selected circuit, which is the output signal of the unit, the data storage device stores the input values and supplies the value that was at the input of this device Δt seconds (0.001≤Δt≤0.5) back relative to the current time, the output of the image analysis unit and the output of the data storage device is connected to the inputs of the ratio unit, the output of which is connected to the first input of the comparison unit, the threshold level unit is connected to the second input of the comparison unit, the output of the comparison unit is connected to the signal unit tion.

The device consists (Fig. 2) of a DVR 1, an image analysis unit 2, a data storage device 3, a relationship unit 4, a comparison unit 5, a threshold level unit 6, an alarm unit 7.

The warning device about a possible collision with a vehicle moving in the oncoming lane contains a video recorder 1 connected to an image analysis unit 2, which is connected to a data storage device 3, the output of the image analysis unit 2 and the output of the data storage device 3 are connected to the inputs of the ratio unit 4 the output of which is connected to the first input of the comparison unit 5, the threshold level unit 6 is connected to the second input of the comparison unit 5, the output of the comparison unit 5 is connected to the signaling unit 7.

The proposed warning device about a possible collision with a vehicle moving in the oncoming lane installed on the vehicle operates as follows.

The image of the traffic situation, obtained from the DVR 1, is sent to the image analysis unit 2, where the presence of the vehicle ahead is recognized by the image. If such a vehicle is detected, then its contour is selected and the area of this contour is calculated. The value of the area is formed at the output of block 3 and is fed to the input of the data storage device 3, where it is stored in memory. Similarly, other incoming contour area values for other time instants are stored in the memory. At the output of device 3, a signal is generated that is equal to the input signal of this device 3 Δt seconds (0.001≤Δt≤0.5) back relative to the current point in time. The output signals of block 2 and device 3 are fed to the inputs of block 4 of the relationship, which performs the operation of dividing the first signal into a second. The result obtained is the output signal of block 4 and is supplied to the first input of block 5 of comparisons, which performs its comparison with the threshold level received at the second input from block 6 of the threshold level. As the vehicle approaches, the result of the area ratio of the contours begins to increase. If the distance between the vehicles becomes less than critical, then the result of the ratio exceeds the threshold level set by block 6, and a high level logic signal appears at the output of the comparison block 5. This signal enters the alarm unit 7, as a result of which the driver is notified of a possible collision with the vehicle in front.

An implementation option for the blocks of the proposed device. Block 2 image analysis can be implemented on the basis of a microprocessor device that compares the image received from the DVR 1 with standards, for example, by correlation methods [3]. The data storage device 3 can be implemented on the basis of a FIFO ring buffer, which at each time point Δt remembers the input values in the cells in accordance with the position of the pointer, and a value taken from the previous cell is generated at the output [4]. Block 4 relations can be performed either on a microprocessor device or on discrete elements of digital technology [4]. Block 5 comparison can be implemented on the basis of the comparator, block 6 of the threshold level on the basis of a constant voltage generator, block 7 alarm - based on the devices of light (for example, LED) and sound (for example, sound speaker) signaling.

The technical and economic effect of the proposed device is to signal a possible collision of vehicles, which will prevent a head-on collision with a vehicle driving into the oncoming lane. The efficiency of collision detection and prediction was achieved by isolating the vehicle contour in the image obtained from the DVR and determining the ratio of the areas of the two selected contours obtained at different points in time.

Literature

1. Patent for utility model RU No. 127498 U1 "Automobile registrar of traffic accidents" 2013.

2. Patent for utility model RU No. 127219 U1 “System for identifying vehicles that illegally invaded a designated lane” 2013.

3. Fisenko V.T., Fisenko T.Yu. Computer processing and image recognition: textbook. allowance. - St. Petersburg: St. Petersburg State University ITMO, 2008.

4. Ugryumov EP Digital circuitry. SPb .: BHV - St. Petersburg, 2000 .-- 89 p.

Claims (1)

A warning device for a possible collision with a vehicle moving in the oncoming lane, mounted on the vehicle and containing a video recorder connected to an image analysis unit, which is connected to a data storage device, characterized in that a ratio unit, a comparison unit, a unit are additionally introduced into the system threshold level, the alarm unit, and the image analysis unit is configured to recognize the contour of the vehicle in front and determine the value I the area of the selected circuit, which is the output signal of the unit, the data storage device stores the input values and supplies the output value that was at the input of this device Δt seconds (0.001≤Δt≤0.5) back relative to the current time, the output of the image analysis unit and the output of the data storage device is connected to the inputs of the ratio unit, the output of which is connected to the first input of the comparison unit, the threshold level unit is connected to the second input of the comparison unit, the output of the comparison unit is connected to the signal unit tion of.

Images