CN201240344Y - Embedded integrated vision auxiliary driving safety system - Google Patents

Abstract

An embedded integrated vision assisted driving safety system is characterized in that the system consists of an acquisition module composed of a video image collector and a range-finding radar; a transmission module based on a bus; an information processing module and a storage module. It consists of a main control module and a man-machine interface module consisting of sound alarm, vibration alarm, video display and operation buttons. The utility model utilizes computer vision and information fusion technology. On the one hand, when there are vehicles approaching at high speed from the rear or other objects that may cause accidents in the adjacent lane, the system can inform the driver of the potential danger in advance; once the driver changes lanes, The system will trigger a warning so that the driver can react in time to avoid a collision; on the other hand, if the vehicle does not deviate from the lane under controlled conditions, the system can issue a warning to ensure driving safety.

Description

Embedded integrated vision-assisted driving safety system

technical field

The utility model relates to a visual aid driving safety system, more specifically a safety driving system mainly used to prevent vehicles from colliding or deviating from a lane.

Background technique

Rising vehicle production coupled with growing safety concerns are driving the active and passive safety systems market. According to a study by the German Insurance Association, 25% of traffic accidents are caused by driver inattention, delayed reaction, or even dozing off in the car. This makes the research on the auxiliary traffic safety system become particularly important.

At present, as the conventional configuration of vehicles, auxiliary driving safety systems mainly have the following two forms, but each has its own problems:

1. The radar is used for passive detection in order to detect accident objects in the warning area, but the specific location of the accident object is not intuitively marked for the driver.

2. The radar is applied to the lane change assist system LCA, but only when the existence of an accident object is detected when the vehicle changes lanes, the alarm will be issued, which cannot avoid the involuntary lane change behavior caused by the driver's inattention.

3. There are visual blind spots. The area within 3 meters on the left and right sides of a motor vehicle and within 3 meters behind it can easily form a visual blind spot on the exterior rearview mirror.

Utility model content

The utility model is to avoid the shortcomings of the above-mentioned prior art, and uses computer vision and information fusion technology to provide an embedded integrated vision-assisted driving safety system. When the vehicle or other objects that may cause an accident, the system can inform the driver of the potential danger in advance; once the driver changes lanes, the system will trigger a warning so that the driver can react in time to avoid collision; on the other hand, if the vehicle is not If the vehicle departs from the lane without control, the system can issue a warning to ensure driving safety.

The utility model solves the technical problem and adopts the following technical solutions:

The feature of the utility model embedded integrated vision assisted driving safety system is that its system composition includes:

An acquisition module composed of a video image collector and a ranging radar, which collects the environmental conditions around the vehicle body in real time;

A bus-based transmission module, which communicates between modules of the system and between the system and other on-board electronic devices in the vehicle;

A main control module composed of an information processing module and a storage module, accepts the collected information about the environmental conditions around the vehicle body from the collection module, realizes the recognition and tracking of the driving lane through image stitching, information fusion and synchronous image recognition, And capture and track potential accident objects in the prevention area during driving, cooperate with detecting whether the driver has actively changed the driving direction, and output execution signals to the human-machine interface module;

A human-machine interface module composed of sound alarm, vibration alarm, video display and operation buttons receives the output signal of the main control module and executes corresponding display or early warning.

The structural features of the utility model system also lie in:

The prevention area is an area within 3 meters on the left and right sides of the motor vehicle and within 3 meters behind it.

The video image collector is a CMOS camera capable of infrared night vision and a viewing angle of 105° to 120°, at least four of which are respectively installed in the front of the car, the left side of the car, the right side of the car and the rear of the car.

The range-finding radar adopts micron-wave radars, at least three of which are respectively installed on the left side of the vehicle body, the right side of the vehicle body and the rear of the vehicle body.

The detection distance of the ranging radar installed on the left side and the right side of the vehicle body is not less than 50 meters, and the detection distance of the ranging radar installed at the rear of the vehicle body is not less than 100 meters.

The video screen display in the man-machine interface module is a color LCD display installed on the front panel in the vehicle.

The vibration alarm in the man-machine interface module is a vibrator installed in the steering wheel of the car.

The data transmission module based on the bus form adopts the general CAN bus protocol interface of automobiles.

The early warning executed by the man-machine interface module in the utility model system includes:

a. When the driver does not actively change the driving direction and the vehicle body deviates from the lane, the system will sound an alarm with an audible alarm;

b. When the driver does not actively change lanes during normal driving, the video monitor displays the fused video images around the vehicle body, and flickers to delineate potential accident objects in the video images;

c. When the driver actively changes lanes, there is a potential accident object in the video image and the radar detects the accident object, the vibration alarm will vibrate and give an early warning.

Compared with the prior art, the beneficial effects of the utility model are reflected in:

1. The utility model uses the acquisition signal of the video image collector to output the "synthesized" video in the detection area after video stitching and information fusion; at the same time, it uses image recognition technology to simultaneously identify adjacent vehicles and pedestrians, overcoming the passivity of radar detection And non-visibility; By flickering and delineating potential accident objects in the video, it reminds the driver to pay attention, realizes active and safe driving, and provides a strong guarantee for safe driving.

2. The utility model monitors the position of the vehicle body relative to the lane marking line in real time through the video image collector, and at the same time performs lane marking line recognition and safety rule matching, and can effectively identify whether the driver is involuntarily changing lanes. When the driver does not actively change the driving direction and the vehicle body deviates from the lane, the system will sound an alarm with the sound siren, which can effectively avoid accidents.

3. The operating buttons in the man-machine interface module of the present invention can be used to set the operating parameters of the system and the active warning level when the system is running. Associate the active warning level with the degree of danger, and determine three danger levels according to the degree of danger, which are high, medium and low. According to driving habits, the driver can set the operating state of the system to three operating states of high safety, medium safety and low safety. The three operating states determine the threshold for determining the danger level from the degree of danger. The higher the safety of the operating state , the smaller the threshold of the danger level, the higher the safe driving factor. According to the threshold value determined by the operating state, the accident objects circled by flickering in the video image are dynamically given three danger levels: high, medium and low.

Description of drawings

Fig. 1 is a structural diagram of the utility model system.

Fig. 2 is a structural block diagram of the utility model equipment.

Figure 3 is a schematic diagram of the radar installation location and detection range.

Fig. 4 is a schematic diagram of the principle of the blind spot detection BSD and the lane change assist system LCA of the present invention.

Fig. 5 is a functional block diagram of the utility model BSD and LCA.

Fig. 6 is a schematic diagram of the principle of the lane departure warning system LDW of the present invention.

Fig. 7 is a functional block diagram of the utility model LDW.

The utility model will be further described below in conjunction with the accompanying drawings through specific embodiments.

Detailed ways

Referring to Fig. 1, present embodiment is based on the embedded control platform of high performance, adopts COMS camera and ranging radar, comprises radar and camera on the left side of the car, radar and camera on the right side of the car, rear radar and camera, and front camera real-time Monitor the visual guard zone and the position of the vehicle relative to the lane. The signal acquisition, storage, transmission, MCU, and information processing modules are packaged into a control box. Based on the transmission module in the form of a bus, it communicates between the modules of the system and between the system and other electronic devices in the vehicle. The main control module composed of the information processing module and the storage module receives the collected information about the surrounding environment of the vehicle body from the collection module, realizes the identification and tracking of the driving lane, and captures and tracks the driving process through image stitching, information fusion and synchronous image recognition Potential accident objects in the medium guard area. The human-machine interface module consists of sound alarm, vibration alarm, video display and operation buttons.

According to conventional techniques in this field, image stitching is based on Opencv's adaptive template matching algorithm, which extracts image corners through the Harris operator, and uses the number and position information of corners to adaptively adjust the size and position of the matching template. This template matching method can not only change the size and position of the template adaptively, but also solve the mismatching problem caused by the illumination change of the color image. Information fusion is to ensure that the stitched image can maintain a smooth visual effect. First, the two statistical parameters of variance and mean are used to adjust the image to be stitched, and the statistical characteristics of one image are assigned to the other image to be stitched. The brightness of the stitched image is basically consistent. Then by introducing the optimal fusion factor. The synchronous image recognition algorithm is a feature-based recognition algorithm. It uses primitive extraction, feature clustering and matching to track lanes, moving vehicles, and pedestrians. The features used are local features of line segments and corner points. The operation speed of the algorithm is fast, and the synchronous image recognition is realized. And it is suitable for complex traffic environments, especially when multiple targets are connected or mutually occluded, it can use a few feature points for identification, and make full use of the high signal-to-noise ratio of local features.

The four cameras shown in Figure 2 are respectively installed in the front of the car, the left side of the car, the right side of the car and the rear of the car. Adopting COMS camera, it has a high image dynamic capture range, high resolution, night vision function and meets the needs of the driving environment. The three radars shown in Figure 2 are micron-wave radars installed on the left side of the vehicle body, the right side of the vehicle body and the rear of the vehicle body respectively, and their detection range should cover the range defined in Figure 3; the digital signal processor 8 is an integrated CAN bus interface and video The high-performance digital signal processor of the peripheral interface such as interface, its processing capacity should be able to meet the real-time processing requirement of multitasking; Memory 9 is the storage unit that SDRAM and FLASH RAM are jointly formed, and SDRAM adopts bidirectional interface type, and capacity meets the video collection volume The needs of FLASH RAM storage software, system parameters and other firmware; LCD display 11 adopts color liquid crystal display, is installed in the instrument panel position in the car, is used for realizing the early warning after decision-making and the function of displaying blind spot video. The vibrator 12 is installed in the steering wheel of the car, and when the driver actively changes lanes and there is an object that may potentially cause an accident, a vibration warning is generated; the electronic device 10 in the car includes a steering wheel rotation sensor and an indicator light controller in the car, which are used to accept driving The driver actively changes the driving direction signal.

Referring to Fig. 3, the micron-wave radar is adopted, two side radars 31 installed on the left side of the vehicle body and the right side of the vehicle body and a tail radar 32 installed on the rear of the vehicle body respectively provide the distances of at least 50 meters and the rear of the vehicle body at least 50 meters on both sides of the vehicle body. The detection distance is at least 100 meters, as shown in the area A in Figure 3; according to the radar wave transmission and reception signals, the relative speed and distance between the object and the vehicle body in a certain volume range in the area are accurately measured, and the relative speed and distance are taken as The basis for calculating the hazard level of an object.

Referring to Fig. 4, according to the prior art, the visual blind spot on the exterior rearview mirror is in the area of 3 meters on the left and right sides of the vehicle and 3 meters behind. In the figure, β is the viewing angle range of the rearview mirror. At present, Audi, Mercedes-Benz, and TRW all use medium-range radar or short-range radar as sensors for detecting blind spots, but they can only give a "yes or no" result of whether there are pedestrians or adjacent vehicles in the detection area.

In this embodiment, the lack of radar detection is compensated by using a large viewing angle video monitoring blind area method. In order to not only provide the driver with a single message of whether there is danger in the blind spot, but also allow the driver to more intuitively and clearly understand the situation of pedestrians and adjacent vehicles in the blind spot. The angle of view α is 105°～120°, and the small color infrared CMOS cameras are respectively arranged on the left side of the vehicle body, the right side of the vehicle body and the rear of the vehicle body, that is, two side cameras 41 and one rear camera 42. According to the size of the vehicle body, three cameras are adjusted synchronously. The angle of a CMOS camera makes the overlapping area of video images larger than 20% of a single video and less than 50% of a single video. Three CMOS cameras detect all blind spots left by the rearview mirror in real time.

Referring to Figure 5, BSD and LCA are composed of three CMOS cameras, ranging radar, signal acquisition and storage module, information fitting module, image fitting module, siren, and video display, among which:

The CMOS camera is a small color infrared camera, and the camera angle can be adjusted.

Ranging radar, using micron wave radar. In order to reduce the cost and improve the cost performance, it is necessary to improve the performance of the radar, reduce the cost of the radar, and reduce the amount of radar used in the system.

The signal collection and storage module is used to collect blind area information and store various drivers, application software and collected video information. BSD and LCA share the signal acquisition module.

The information fitting module is based on image measurements. Using image processing and image fitting technology, the blind spot information monitored by the three CMOS cameras is stitched into a more complete image, truly realizing blind spot-free driving.

The image fitting module is based on image recognition and segmentation. Apply image recognition technology to more intuitively identify the location of potential accident objects in the blind spot. That is, for the multi-sensor "synthetic" image output by the information fitting module, synchronously identify and estimate the nearby (shortest distance) accident objects, and remind the driver to pay attention to the identified potential accident objects by flickering.

When the driver actively changes lanes and there is a potential accident object in the video image and the radar detects the accident object, the alarm will vibrate and give an early warning. The vibration alarm is installed on the vibrator in the steering wheel of the car. BSD and LCA share annunciators.

The video display is used to display a "composite" video of the detection area captured by the 3 CMOS cameras. The video display is a color LCD display mounted on the front panel of the vehicle.

Referring to Figure 6, the lane departure warning LDW monitors the position of the vehicle body relative to the lane markings through a CMOS front-end camera 61 installed on the front bumper of the vehicle. The viewing angle range γ of the front-end camera 61 is 105°-120°. When the driver changes lanes involuntarily without paying attention, the alarm will sound an alarm to remind the driver to pay attention.

Referring to Figure 7, LDW consists of a camera, signal acquisition module, storage, lane marking line recognition safety rule matching module and siren, wherein:

The signal collection module is used to collect the position information of the vehicle body relative to the lane marking line;

The storage module is used to store various drivers, application software and collected video information;

Lane marking recognition safety rule matching module includes parallel lane marking recognition algorithm, image fitting algorithm, adjacent vehicle/pedestrian recognition extraction algorithm, and image-based measurement algorithm;

Siren, when the driver does not actively change the driving direction and the vehicle body deviates from the lane, the system will sound an alarm with the sound siren to avoid the occurrence of traffic accidents.

Claims (7)

1, a kind of built-in integrated visual sensation auxiliary driving safety system is characterized in that system's formation comprises:

An acquisition module of being made up of jointly video image acquisition device and range only radar is gathered vehicle body environmental aspect all around in real time;

Transport module based on bus form communicates between each module of system and communicates between other vehicle electronic device in system and car;

A main control module of forming jointly by message processing module and memory module, acceptance is from the Information Monitoring about the environmental aspect around the vehicle body of described acquisition module, realize the identification and the tracking of driving lane by image stitching, information fusion and synchronous images identification, and catch and follow the tracks of and take precautions against latent defect object in the zone in the driving process, cooperate and detect the action whether driver has active conversion direction of traffic, output is carried out signal to human-machine interface module;

A human-machine interface module of being made up of audible alarm, vibration alarm device, video display unit (VDU) and action button is accepted the output signal of described main control module, carries out corresponding the demonstration or early warning.

2, built-in integrated visual sensation auxiliary driving safety system according to claim 1, it is characterized in that described video image acquisition device is but that infrared night vision, visual angle are 105 °～120 ° CMOS camera, be at least four, be installed on car respectively before, car left side, car right side and the tailstock.

3, built-in integrated visual sensation auxiliary driving safety system according to claim 1 is characterized in that described range only radar adopts the micron wave radar, at least three, is installed on vehicle body left side, vehicle body right side and tail of body respectively.

4, built-in integrated visual sensation auxiliary driving safety system according to claim 3, it is characterized in that the described detection range that is installed on the range only radar on vehicle body left side and vehicle body right side is no less than 50 meters, the detection range that is installed in the range only radar of tail of body is no less than 100 meters.

5, built-in integrated visual sensation auxiliary driving safety system according to claim 1 is characterized in that the screen telltale in the described human-machine interface module is to be installed on the color LCD display on the front panel in the car.

6, built-in integrated visual sensation auxiliary driving safety system according to claim 1 is characterized in that the vibration alarm in the described human-machine interface module is the vibrator that is installed in the vehicle steering.

7, built-in integrated visual sensation auxiliary driving safety system according to claim 1 is characterized in that described data transmission module based on bus form is to adopt the general CAN bus protocol of automobile interface.

Images