A warning and active defense device for sudden braking of a vehicle ahead

Abstract

The present invention discloses an early warning and active defense device for sudden braking of a vehicle ahead. The device uses a method of integrating the recognition of the taillights of a vehicle ahead with the recognition of acceleration to determine the sudden braking tendency and behavior of the vehicle ahead, and issues an audible and visual alarm to the driver. Under different lighting and climatic conditions, the taillights of the vehicle ahead are often not clear enough or cannot be seen clearly, which affects the recognition of sudden braking by the following vehicle and causes a collision accident. In view of this situation, the present invention uses recognition enhancement technology for the image of the taillights of the vehicle ahead captured by the camera, and simultaneously integrates the acceleration sensor of the vehicle and the radar speed measurement to monitor the driving state of the vehicle ahead in real time, which can more accurately determine the braking intention of the vehicle ahead and improve the driving safety of the vehicle. The present invention integrates information from different information sources such as image recognition and acceleration state recognition, and can more deeply and accurately determine the sudden braking intention of the vehicle ahead. It is superior to the current single recognition method and can better ensure driving safety.

Description

Technical Field

[0001] The present invention relates to the field of vehicle safety auxiliary driving devices, and in particular to an early warning and active defense device for sudden braking of a vehicle ahead. Background Art

[0002] With the explosive growth of the number of motor vehicles, social problems such as traffic accidents, traffic congestion, and deteriorating traffic environment have become increasingly prominent. Frequent traffic accidents have caused huge losses to the lives and property of the people. According to traffic accident data released by relevant departments, rear-end collisions account for 30%-40% of traffic accidents, among which rear-end collisions are the most common type of traffic accidents. Traditional intelligent collision avoidance systems for automobiles ignore the comprehensive impact of drivers, vehicle conditions, and road traffic environments on driving safety, and thus cannot adapt to complex and changeable road traffic environments. The active defense device for automobile driving safety proposed in the present invention can accurately identify the driving intention of the vehicle in front by integrating machine vision, multi-sensor information fusion, and auxiliary decision-making. It has the characteristics of real-time, accuracy, and a wide range, and is of great significance to improving driver safety, alleviating urban traffic congestion, reducing emission pollution, and promoting urban economic development.

[0003] The present invention firstly uses the relevant methods of image processing to realize image denoising, filtering and grayscale conversion, and then effectively integrates the image processing information with the acceleration sensor information to analyze and predict the driving intention of the vehicle ahead, with the focus on the semantic recognition of the taillights, so as to realize the early warning and active defense of the sudden braking of the vehicle ahead. Summary of the invention

[0004] The purpose of the present invention is to provide an active defense method and device for predicting the sudden braking of the preceding vehicle in advance and providing an audible and visual alarm by judging the behavior of the preceding vehicle and identifying the preceding vehicle's sudden braking intention, so as to solve the problems raised in the above-mentioned background technology.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0006] A warning and active defense device for sudden braking of a leading vehicle, comprising a central processing unit module, wherein a video signal processing hub is fixedly provided inside the central processing unit module, the video signal processing hub is electrically connected to a camera data preprocessing module, and the camera data preprocessing module is electrically connected to a vehicle-mounted camera to collect an image of a leading vehicle's taillights; a leading vehicle acceleration processing hub is fixedly provided inside the central processing unit module, the leading vehicle acceleration processing hub is electrically connected to an acceleration sensor of the vehicle to collect the acceleration of the vehicle, and the leading vehicle acceleration processing hub is electrically connected to a radar speed meter to collect a relative speed of the leading vehicle relative to the vehicle; a leading vehicle intention judgment module is fixedly provided inside the central processing unit module, the leading vehicle intention judgment module is electrically connected to a leading vehicle sudden braking alarm module, the leading vehicle sudden braking alarm module is electrically connected to a center console display module, and the leading vehicle sudden braking alarm module is electrically connected to an audio alarm module.

[0007] As a preferred technical solution of the present invention, the image data processing algorithm of the video signal processing center includes: extracting taillight pairs using directional gradient histogram, position histogram and color correlation features for the front vehicle rear image preprocessed by the camera data preprocessing module, and realizing convolutional neural network taillight status detection based on image separation areas using the histogram features of the taillight area image.

[0008] As a preferred technical solution of the present invention, the image data preprocessing algorithm of the camera data preprocessing module includes: converting the image captured by the vehicle-mounted camera from the RGB color space to the hue saturation value (HSV) color space, thereby achieving more accurate taillight segmentation and solving the difficulty of detecting red taillights in the RGB space.

[0009] As a preferred technical solution of the present invention, the preceding vehicle acceleration processing center can obtain real-time acceleration information of the preceding vehicle by processing the acceleration of the preceding vehicle collected by the acceleration sensor of the preceding vehicle and the relative speed of the preceding vehicle collected by the radar speed meter.

[0010] As a preferred technical solution of the present invention, the preceding vehicle intention judgment algorithm of the preceding vehicle intention judgment module includes: 1. analyzing the probability that the preceding vehicle's taillights are red; 2. analyzing the probability that the preceding vehicle suddenly decelerates; 3. using evidential reasoning and fuzzy mathematics methods to fuse the probability that the preceding vehicle's taillights are red and the probability that the preceding vehicle suddenly decelerates to obtain the probability that the preceding vehicle suddenly brakes.

[0011] As a preferred technical solution of the present invention, the acceleration sensor of the vehicle is electrically connected to the preceding vehicle emergency brake alarm module.

[0012] As a preferred technical solution of the present invention, the judgment algorithm for the preceding vehicle emergency brake alarm module to issue different levels of alarms includes: 1. no alarm when the vehicle distance is far and the vehicle has decelerated; 2. issuing the highest level alarm when the vehicle distance is close and the vehicle has not decelerated; 3. using the vehicle distance and relative speed as risk indicators, and the acceleration of the vehicle's deceleration as a risk mitigation indicator, performing gradient classification on the current risk, and issuing different alarms.

[0013] The beneficial effects of the present invention are as follows: the warning and active defense device for the sudden braking of the front vehicle gets rid of the excessive reliance on the front vehicle taillight recognition in the traditional front vehicle sudden braking recognition method, and draws the driving habit behavior spectrum of the front vehicle according to the acceleration and deceleration state of the front vehicle by measuring the acceleration and deceleration of the front vehicle, so as to analyze the driving intention of the front vehicle in real time for the currently collected front vehicle acceleration, and take the active defense of sound and light alarm to the driver, thereby reducing the occurrence of rear-end collision accidents and improving the safety level of automobile driving. The main advantages of the active defense method by identifying the sudden braking intention of the front vehicle are: first, the driving intention judgment can complement and correct the taillight recognition of the front vehicle, so that the driver of the vehicle can timely avoid the dangerous driving behavior of the front vehicle when the taillight recognition of the front vehicle is not obvious or there is no time for taillight recognition; second, the method does not require high camera clarity, and the intention recognition technology based on algorithm optimization can realize the multi-source data fusion of the camera, the acceleration sensor of the vehicle and the vehicle-mounted radar and make up for the deficiency of the data of each sensor itself, and improve the data unreliability caused by the instability of a single sensor, thereby improving the driving safety of the vehicle. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0015] Figure 1 This is an overall system block diagram of a warning and active defense device for sudden braking of a vehicle ahead of the present invention;

[0016] Figure 2 It is a flowchart of judging the sudden braking intention of a vehicle ahead of the present invention in a warning and active defense device for sudden braking of a vehicle ahead;

[0017] Figure 3 The present invention discloses a flowchart of an image data preprocessing algorithm for an early warning and active defense device for sudden braking of a vehicle ahead.

[0018] In the figure: 1. Central processing unit module; 2. Video signal processing central module; 3. Camera data preprocessing module; 4. Vehicle-mounted camera; 5. Front vehicle acceleration processing central module; 6. Vehicle acceleration sensor; 7. Radar speedometer; 8. Front vehicle intention judgment module; 9. Front vehicle emergency brake alarm module; 10. Center console display module; 11. Audio alarm module. DETAILED DESCRIPTION

[0019] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods.

[0020] like Figure 1-2 As shown, the present invention provides an early warning and active defense device for sudden braking of a vehicle ahead, comprising a central processing unit module 1, characterized in that: a video signal processing hub module 2 is fixedly arranged inside the central processing unit module 1, the video signal processing hub module 2 is electrically connected to a camera data preprocessing module 3, the camera data preprocessing module 3 is electrically connected to a vehicle-mounted camera 4 to collect an image of the taillights of the vehicle ahead; a front vehicle acceleration processing hub module 5 is fixedly arranged inside the central processing unit module 1, the front vehicle acceleration processing hub module 5 is electrically connected to an acceleration sensor 6 of the vehicle to collect the acceleration of the vehicle ahead, the front vehicle acceleration processing hub module 5 is electrically connected to a radar speed meter 7 to collect the relative speed of the front vehicle relative to the vehicle ahead; a front vehicle intention judgment module 8 is fixedly arranged inside the central processing unit module 1, the front vehicle intention judgment module 8 is electrically connected to a front vehicle sudden braking alarm module 9, the front vehicle sudden braking alarm module 9 is electrically connected to a center console display module 10, and the front vehicle sudden braking alarm module 9 is electrically connected to an audio alarm module 11.

[0021] The image data processing algorithm of the video signal processing central module 2 includes: extracting taillight pairs using the directional gradient histogram, position histogram and color correlation features for the front vehicle rear image preprocessed by the camera data preprocessing module 3, and realizing the convolutional neural network taillight state detection based on the image separation area using the histogram features of the taillight area image. The specific algorithm is as follows:

[0022] The convolutional neural network based on image segmentation area is mainly composed of convolution layer, pooling layer and output layer after activation function processing. Among them, the convolution layer extracts features from the front vehicle rear image preprocessed by the camera data preprocessing module 3, and then uses the extracted features to classify, identify and predict the front vehicle taillight image. Assuming the input image size is M×N, the filter kernel size is K×K, the step size is S, and the number of padded pixels is P, then the formula for calculating the size of the feature map output by the layer is:

[0023] O = (I-K+2P) / S+1 (Formula 1)

[0024] The pooling layer is usually connected after the convolution layer to aggregate features by reducing dimensionality and reduce computational complexity. First, the feature map must pass through the fully connected layer and the softmax layer to calculate the classification proposal and output the detected probability vector; then, the position offset of each solution is obtained through bounding box probability regression, thereby regressing a more accurate target detection boundary. The formula for calculating the size of the feature map output by the pooling layer is:

[0025] O = (IK) / S+1 (Formula 2)

[0026] The output layer inputs the data into two structures. One part uses the softmax classifier to classify the anchor points. After classification, the foreground (detection object) and background can be obtained; the other part is used to calculate the bounding box regression offset of the anchor point. After classification, the foreground (detection object) and background can be obtained, and more accurate suggestions can be obtained. The final output layer combines the foreground anchor and the bounding box regression offset to obtain the candidate bounding box of the front vehicle target. Some inappropriate suggestions are excluded. Then the suggestion layer is used to achieve target positioning and determine the brightness state of the front vehicle taillights.

[0027] The image data preprocessing algorithm of the camera data preprocessing module 3 includes: converting the image captured by the vehicle camera 4 from the RGB color space to the hue saturation value (HSV) color space, thereby achieving more accurate taillight segmentation and solving the difficulty of detecting red taillights in the RGB space. The process of the image data preprocessing algorithm of the camera data preprocessing module 3 is as follows: Figure 3 shown.

[0028] The front vehicle acceleration processing central module 5 processes the vehicle speed v collected by the vehicle acceleration sensor 6. 0t and acceleration a 0t The speed v of the vehicle ahead collected by the radar speed meter 7 t , the real-time acceleration information of the front vehicle can be obtained, and the calculation method is:

[0029]

[0030] Among them, v t is the speed of the preceding vehicle measured at the current moment, v t-1 is the speed of the preceding vehicle measured at the previous moment, and the speed and acceleration of the vehicle at the current moment are v 0t , a 0t , the acquisition period of the radar speed meter is Δt.

[0031] Among them, the preceding vehicle intention judgment algorithm of the preceding vehicle intention judgment module 8 includes: 1. Obtaining the probability P1 that the preceding vehicle taillight is red according to the result of the video signal processing central module 2; 2. Obtaining the probability P2 that the preceding vehicle suddenly decelerates according to the preceding vehicle acceleration processing central module 5; 3. Using evidence reasoning and fuzzy mathematics methods to fuse the probability that the preceding vehicle taillight is red and the probability that the preceding vehicle suddenly decelerates, obtaining the probability of the preceding vehicle suddenly braking, the calculation formula is as follows:

[0032] We use a two-element set Θ = {P, NP} as the recognition framework, where P represents the reasoning that supports the sudden braking of the front car, and NP represents the opposite. The power set of Θ is where {P,NP} is a universal set, is an empty set. Let m be the basic probability assignment function on Θ, and assign quantitative confidence to m({P}) and m({NP}), according to which the behavior of interest can be classified as supporting or unsupported inferences of the current preceding vehicle's sudden braking reasoning. In particular, m({P,NP}) specifies the degree of uncertainty that a given preceding vehicle can be identified as supporting or unsupported. Its constraints are m({P})+m({NP})+m({P,NP})=1, and the benchmark function is:

[0033]

[0034] The curves are defined as follows:

[0035] m({P}) f1(x)=f0(1.5x,0.6,0.7)

[0036] m({NP}) f2(x)=f0(2-1.5x,0.6,0.7)

[0037] m({NP,P}) f3(x)=1-f1(x)-f2(x)

[0038] Among them, the judgment algorithm for the preceding vehicle emergency brake alarm module 9 to issue different levels of alarms includes: 1. no alarm when the vehicle distance is far and the vehicle has decelerated; 2. issuing the highest level alarm when the vehicle distance is close and the vehicle has not decelerated; 3. using the vehicle distance and relative speed as risk indicators, and the acceleration of the vehicle's deceleration as a risk mitigation indicator, to grade the current risk and make different alarms.

[0039] The specific implementation steps at work are as follows Figure 2, the vehicle-mounted camera 4 is used to collect the image of the front vehicle's taillights, and the front vehicle's taillight signal is transmitted to the camera data preprocessing module 3 to preprocess the front vehicle's taillight image data. Then, the preprocessed front vehicle's rear image in the video signal processing central module 2 uses the directional gradient histogram, position histogram and color correlation features to extract the taillight pair, and the histogram features of the taillight area image are used to implement the convolutional neural network taillight state detection based on the image separation area. The video signal processing central module 2 transmits the processed probability P1 that the front vehicle's taillight is a red light to the central processing unit module 1 and then to the front vehicle intention judgment module 8. The front vehicle acceleration processing central module 5 can obtain the real-time acceleration information of the front vehicle by processing the vehicle's speed and acceleration collected by the vehicle's acceleration sensor 6 and the front vehicle's speed collected by the radar speed meter 7. The vehicle acceleration processing central module 5 transmits the probability P2 of the front vehicle's sudden deceleration to the central processing unit module 1 and then to the front vehicle intention judgment module 8. The preceding vehicle emergency brake alarm module 9 gives different types of alarm reminders such as no alarm, preliminary warning and emergency alarm according to the preceding vehicle intention judgment module 8's judgment on the preceding vehicle emergency brake situation, and issues corresponding alarms through the center console display module 10 and the sound alarm module 11.

[0040] Finally, it should be noted that in the description of the present invention, it should be noted that the terms "vertical", "up", "down", "horizontal", etc. indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, they cannot be understood as limitations on the present invention.

[0041] In the description of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0042] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention is described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A warning and active defense device for sudden braking of a vehicle ahead, comprising a central processing unit module (1), characterized in that: A video signal processing central module (2) is fixedly arranged inside the central processing unit module (1), the video signal processing central module (2) is electrically connected to a camera data preprocessing module (3), the camera data preprocessing module (3) is electrically connected to a vehicle-mounted camera (4) to collect an image of the taillights of the preceding vehicle; a preceding vehicle acceleration processing central module (5) is fixedly arranged inside the central processing unit module (1), the preceding vehicle acceleration processing central module (5) is electrically connected to an acceleration sensor of the vehicle (6) to collect the acceleration of the vehicle, the preceding vehicle acceleration processing central module (5) is electrically connected to a radar speed meter (7) to collect the relative speed of the preceding vehicle relative to the vehicle; a preceding vehicle intention judgment module (8) is fixedly arranged inside the central processing unit module (1), the preceding vehicle intention judgment module (8) is electrically connected to a preceding vehicle emergency brake alarm module (9), the preceding vehicle emergency brake alarm module (9) is electrically connected to a central console display module (10), and the preceding vehicle emergency brake alarm module (9) is electrically connected to an audio alarm module (11); The preceding vehicle intention judgment module (8) comprises the following steps:

1. analyzing the probability that the preceding vehicle's taillight is red; 2. analyzing the probability that the preceding vehicle decelerates suddenly; and 3. using evidence reasoning and fuzzy mathematics methods to fuse the probability that the preceding vehicle's taillight is red and the probability that the preceding vehicle decelerates suddenly, thereby obtaining the probability that the preceding vehicle brakes suddenly.

2. The device for early warning and active defense of sudden braking of the vehicle ahead according to claim 1, characterized in that: The image data processing algorithm of the video signal processing central module (2) comprises: extracting taillight pairs from the front vehicle rear image preprocessed by the camera data preprocessing module (3) using directional gradient histogram, position histogram and color correlation features, and implementing taillight state detection based on a convolutional neural network of image separation regions using the histogram features of the taillight region image.

3. The device for early warning and active defense of sudden braking of a vehicle ahead according to claim 1, characterized in that: The image data preprocessing algorithm of the camera data preprocessing module (3) comprises: converting the image collected by the vehicle-mounted camera (4) from the RGB color space into the hue saturation value (HSV) color space, thereby achieving more accurate taillight segmentation and solving the difficulty of detecting red taillights in the RGB space.

4. The device for early warning and active defense of sudden braking of a vehicle ahead according to claim 1, characterized in that: The preceding vehicle acceleration processing central module (5) can obtain real-time acceleration information of the preceding vehicle by processing the acceleration of the vehicle collected by the preceding vehicle acceleration sensor (6) and the relative speed of the preceding vehicle collected by the radar speed meter (7).

5. The device for early warning and active defense of sudden braking of a vehicle ahead according to claim 1, characterized in that: The host vehicle acceleration sensor (6) is electrically connected to the preceding vehicle emergency brake alarm module (9).

6. The device for early warning and active defense of sudden braking of a vehicle ahead according to claim 1, characterized in that: The judgment algorithm for the preceding vehicle emergency brake alarm module (9) to issue different levels of alarms includes: first, no alarm when the distance between vehicles is relatively far and the vehicle has decelerated; second, issuing the highest level of alarm when the distance between vehicles is relatively close and the vehicle has not decelerated; and third, taking the distance between vehicles and the relative speed as risk indicators, taking the acceleration of the vehicle's deceleration as a risk mitigation indicator, performing gradient classification on the current risk, and issuing different alarms.

Images