CN110633653B - Doubling auxiliary system based on streaming media inside rearview mirror - Google Patents
Doubling auxiliary system based on streaming media inside rearview mirror Download PDFInfo
- Publication number
- CN110633653B CN110633653B CN201910800923.8A CN201910800923A CN110633653B CN 110633653 B CN110633653 B CN 110633653B CN 201910800923 A CN201910800923 A CN 201910800923A CN 110633653 B CN110633653 B CN 110633653B
- Authority
- CN
- China
- Prior art keywords
- vehicle
- rear view
- doubling
- lane
- view image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60R—VEHICLES, VEHICLE FITTINGS, OR VEHICLE PARTS, NOT OTHERWISE PROVIDED FOR
- B60R1/00—Optical viewing arrangements; Real-time viewing arrangements for drivers or passengers using optical image capturing systems, e.g. cameras or video systems specially adapted for use in or on vehicles
- B60R1/12—Mirror assemblies combined with other articles, e.g. clocks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
- G06V20/584—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of vehicle lights or traffic lights
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/588—Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60R—VEHICLES, VEHICLE FITTINGS, OR VEHICLE PARTS, NOT OTHERWISE PROVIDED FOR
- B60R1/00—Optical viewing arrangements; Real-time viewing arrangements for drivers or passengers using optical image capturing systems, e.g. cameras or video systems specially adapted for use in or on vehicles
- B60R1/12—Mirror assemblies combined with other articles, e.g. clocks
- B60R2001/1253—Mirror assemblies combined with other articles, e.g. clocks with cameras, video cameras or video screens
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60R—VEHICLES, VEHICLE FITTINGS, OR VEHICLE PARTS, NOT OTHERWISE PROVIDED FOR
- B60R2300/00—Details of viewing arrangements using cameras and displays, specially adapted for use in a vehicle
- B60R2300/80—Details of viewing arrangements using cameras and displays, specially adapted for use in a vehicle characterised by the intended use of the viewing arrangement
- B60R2300/8046—Details of viewing arrangements using cameras and displays, specially adapted for use in a vehicle characterised by the intended use of the viewing arrangement for replacing a rear-view mirror system
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Mechanical Engineering (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention discloses a merging auxiliary system based on a streaming media inside rear view mirror, which utilizes the streaming media inside rear view mirror widely used on a vehicle to track the position information of a rear vehicle and judge whether the running conditions of the merging vehicle and the rear vehicle belong to one of four situations of merging danger or not based on the partition information of a preset rear view image so as to realize the auxiliary function of providing merging safety information for a driver when preparing merging; in addition, the invention does not need to add additional equipment on the vehicle, and has the advantage of low cost.
Description
Technical Field
The invention relates to the field of vehicles, in particular to a doubling auxiliary system based on a streaming media inside rearview mirror.
Background
The streaming media rearview mirror has gradually replaced the traditional rearview mirror by virtue of wide visual angle and many advantages. The visual angle of the existing streaming media rearview mirror reaches more than 150 degrees, and although the streaming media rearview mirror can see a vehicle behind, the streaming media rearview mirror cannot actively warn dangers caused by the fact that the vehicle comes from the side and the rear when the vehicle is in a parallel line.
The doubling assistance function has been widely used in vehicles. In the prior art, a microwave radar or cameras arranged in a left rearview mirror and a right rearview mirror are mostly adopted to realize the doubling auxiliary function. However, if the streaming inner rear view mirror is installed, the use of both techniques increases the cost and requires additional installation.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: the utility model provides a doubling auxiliary system based on streaming media inside rear-view mirror to solve current vehicle doubling auxiliary function and need increase the problem that the additional equipment causes the cost-push on the vehicle and need additionally carry out equipment fixing.
The technical scheme adopted by the invention is as follows:
a merging auxiliary system based on a streaming media inside rearview mirror comprises the streaming media inside rearview mirror installed on a merging vehicle, wherein the streaming media inside rearview mirror is provided with a rear camera and a processor, the processor can acquire a rear view image behind the merging vehicle in real time through the rear camera, and the rear view image covers a current lane where the merging vehicle is located and a left lane and a right lane which are respectively located on the left side and the right side of the current lane; the method is characterized in that:
the processor is provided with an image processing module and a doubling auxiliary module;
the image processing module is configured to: acquiring one rear-view image every preset detection period, and identifying the position information of a rear vehicle displayed in the rear-view image, wherein the position information comprises: the vehicle rear view system comprises a rectangular positioning frame and coordinates of a central point of the rectangular positioning frame in the rear view image, wherein four sides of the rectangular positioning frame are tangent to the rear vehicle, and the rear vehicle comprises two conditions: the rear vehicle is a whole vehicle which is completely displayed in the rear view image and is marked as a complete vehicle, and the rear vehicle is a part of the whole vehicle which is displayed in the rear view image and is marked as a vehicle part; thus, the position tracking of the rear vehicle is realized.
The merging assist module presets a left lane boundary and a right lane boundary as the divisional marks of the rear view image such that: in the rear view image, the left lane is located in the areas on the two sides of the left lane boundary and respectively refers to a left lane blind area which cannot be observed by the driver of the parallel vehicle through a left rearview mirror and a visible area which can be observed by the driver of the parallel vehicle, and the right lane is located in the areas on the two sides of the right lane boundary and respectively refers to a right lane blind area which cannot be observed by the driver of the parallel vehicle through a right rearview mirror and a visible area which can be observed by the driver of the parallel vehicle;
the doubling auxiliary module is used for generating doubling safety information of the doubling vehicle, and comprises:
in the first situation, if the central point of the rectangular positioning frame is located in a left lane blind area of the rear-view image, it is judged that the merging vehicle merges to the left lane at the current moment as a danger;
in a second situation, if the central point of the rectangular positioning frame is located in the right lane blind area of the rear-view image, judging that the merging vehicle merges towards the right lane at the current moment to be dangerous;
when the rear vehicle is a complete vehicle, if the central point of the rectangular positioning frame falls in a visible area of a left lane of the rear-view image, calculating the pre-judging time of the left lane when the rear vehicle reaches the blind area of the left lane; if the rear vehicle does not appear in the rear view image at the timing ending time of the left lane, judging that the merging vehicle merges into the left lane at the timing ending time of the left lane as a danger;
when the rear vehicle is a complete vehicle, if the central point of the rectangular positioning frame falls in a visible region of a right lane of the rear-view image, calculating the right lane prejudgment time when the rear vehicle reaches the blind region of the right lane; if the rear vehicle does not appear in the rear view image at the right lane timing end time, judging that the merging vehicle merges into the right lane at the right lane timing end time as dangerous;
and judging that the doubling vehicle is safe when being doubled at the current moment except the four situations.
Therefore, the invention utilizes the streaming media inside rear-view mirror widely used on the vehicle to track the position information of the rear vehicle and judge whether the running conditions of the parallel vehicle and the rear vehicle belong to one of four situations of parallel danger or not based on the preset partition information of the rear view image so as to realize the auxiliary function of providing parallel safety information for the driver when preparing for parallel; in addition, the invention has the advantage of low cost without adding extra equipment on the vehicle.
As a preferred embodiment of the present invention: the processor is also provided with an early warning module;
the early warning module is used for: the access the indicator light information of doubling vehicle to based on the doubling safety information that doubling auxiliary module produced, produce doubling early warning signal, include:
when the information of the steering lamps of the doubling vehicle is left-turning, if the doubling safety information is the situation I or the situation III, generating a doubling early warning signal;
and when the information of the steering lamp of the doubling vehicle is right turning, if the doubling safety information is the situation two or the situation four, generating the doubling early warning signal.
Therefore, the invention utilizes the early warning module to actively warn a driver that the parallel wire is dangerous by generating the parallel wire early warning signal when the parallel wire vehicle prepares the parallel wire and accords with one of four parallel wire dangers, thereby greatly improving the driving safety.
Preferably, the following components: the doubling auxiliary system also comprises an early warning execution device; the early warning execution device can send out an alarm when the early warning module generates a parallel early warning signal.
As a preferred embodiment of the present invention: the image processing module is preset with a vehicle position detection model obtained based on the training of a MobileNet-V2 neural network model, and the rear view image is converted into data in a TFRecord format and then input into the vehicle position detection model to obtain the position information of the rear vehicle.
Therefore, the vehicle position detection model trained on the MobileNet-V2 neural network model is adopted, the position information of the rear vehicle in the rear view image is identified, the operation speed is high, and the method and the device are suitable for the existing streaming media inside rear view mirror.
Preferably: the mode of training to obtain the vehicle position detection model is as follows:
step one, collecting a plurality of rear view images to obtain an image data sample;
manually marking the position information and the vehicle information of the rear vehicle in each rear-view image to obtain a position information sample and a vehicle information sample, wherein the vehicle information indicates that the rear vehicle is a complete vehicle or a local vehicle;
step three, converting the image data sample, the corresponding position information sample and the corresponding vehicle information sample into TFRecord format data, and dividing the TFRecord format data into a training set and a testing set according to a preset proportion;
inputting the training set into a MobileNet-V2 neural network model for training to obtain a model to be verified;
and step five, verifying the model to be verified by using the test set, wherein if the model to be verified passes the verification, the model to be verified is the vehicle position detection model, and otherwise, the step one to the step five are executed again.
As a preferred embodiment of the present invention: in the third situation, the calculation method of the left lane anticipation time is as follows: the method comprises the steps of firstly calculating the speed of a rear vehicle at the current moment according to position information of the rear vehicle in N continuous rear view images, and then calculating the pre-judging time of the left lane by combining the distance between the rear vehicle at the current moment and the blind zone of the left lane, wherein N is a positive integer larger than 2.
As a preferred embodiment of the present invention: in the fourth case, the calculation method of the right lane pre-judgment time is as follows: the method comprises the steps of firstly calculating the speed of a rear vehicle at the current moment according to position information of the rear vehicle in N continuous rear view images, and then calculating the right lane prejudgment time by combining the distance between the rear vehicle at the current moment and the right lane blind area, wherein N is a positive integer larger than 2.
Compared with the prior art, the invention has the following beneficial effects:
firstly, the invention utilizes the streaming media inside rear view mirrors widely used on the vehicle to track the position information of the rear vehicle and judge whether the running conditions of the parallel vehicle and the rear vehicle belong to one of four situations of parallel danger or not based on the preset partition information of the rear view image so as to realize the auxiliary function of providing parallel safety information for the driver when preparing for parallel; in addition, the invention does not need to add additional equipment on the vehicle, and has the advantage of low cost.
Secondly, the invention utilizes the early warning module to actively warn a driver that the parallel wire is dangerous by sending out the parallel wire early warning signal when the parallel wire vehicle prepares for parallel wire and accords with one of four parallel wire dangers, thereby greatly improving the driving safety.
Thirdly, the vehicle position detection model obtained by training based on the MobileNet-V2 neural network model is adopted, the position information identification of the rear vehicle in the rear view image is realized, the method has the advantage of high operation speed, and can be suitable for the existing streaming media inside rear view mirror.
Drawings
The invention is described in further detail below with reference to the following figures and specific examples:
FIG. 1 is a functional block diagram of a doubling assistance system of the present invention;
FIG. 2 is a schematic view of a rear view image 1 according to the present invention;
FIG. 3 is a block diagram of a process for training a vehicle position detection model according to the present invention.
Detailed Description
The present invention will be described in detail with reference to the following embodiments and the accompanying drawings to help those skilled in the art better understand the inventive concept of the present invention, but the scope of the claims of the present invention is not limited to the following embodiments, and it will be apparent to those skilled in the art that all other embodiments obtained without inventive efforts without departing from the inventive concept of the present invention belong to the scope of the present invention.
Example one
As shown in fig. 1 and fig. 2, the present invention discloses a merging auxiliary system based on a streaming media interior rear view mirror, which includes a streaming media interior rear view mirror installed on a merging vehicle, wherein the streaming media interior rear view mirror is provided with a rear camera and a processor, the processor can acquire a rear view image 1 behind the merging vehicle in real time through the rear camera, and the rear view image 1 covers a current lane 2 where the merging vehicle is located, and a left lane 3 and a right lane 4 respectively located on the left side and the right side of the current lane 2;
the processor is provided with an image processing module and a doubling auxiliary module;
the image processing module is configured to: acquiring one rear view image 1 every preset detection period, and identifying the position information of a rear vehicle displayed in the rear view image 1, wherein the position information comprises: a rectangular positioning frame a having four sides tangential to the rear vehicle, and coordinates of a center point A1 of the rectangular positioning frame a in the rear view image 1, wherein the rear vehicle includes two cases: firstly, the rear vehicle is a whole vehicle which is completely displayed in the rear view image 1 and is marked as a complete vehicle, and secondly, the rear vehicle is a part of the whole vehicle which is displayed in the rear view image 1 (namely, part of the whole vehicle is positioned in a blind area of a rear camera and cannot be displayed in the rear view image 1) and is marked as a vehicle part; thus, the position tracking of the vehicle behind is realized.
The rear view image 1 may be obtained by the rear camera through a photographing function, or may be a video frame read from a video of the rear camera.
The merging support module is preset with a left lane boundary 3a and a right lane boundary 4a as the division marks of the rear view image 1 such that: in the rear view image 1, the left lane 3 is located in the areas on both sides of the left lane boundary 3a, namely, a left lane blind area 3-1 and a left lane visible area 3-2 which cannot be observed by the driver of the merging vehicle through a left rearview mirror, and the right lane 4 is located in the areas on both sides of the right lane boundary 4a, namely, a right lane blind area 4-1 and a right lane visible area 4-2 which cannot be observed by the driver of the merging vehicle through a right rearview mirror; after the streaming media inside rear view mirror is installed, the positions of the left lane boundary 3a and the right lane boundary 4a in the rear view image 1 can be obtained by manual marking.
The doubling auxiliary module is used for generating doubling safety information of the doubling vehicle, and comprises:
in the first situation, if the central point A1 of the rectangular positioning frame A falls in a left lane blind area 3-1 of the rearview image 1, judging that the merging vehicle merges to the left lane 3 at the current moment as a danger; this case one applies to both cases where the rear vehicle is a complete vehicle or a vehicle part.
In the second situation, if the central point A1 of the rectangular positioning frame A falls in the right lane blind area 4-1 of the rear view image 1, judging that the merging vehicle merges to the right lane 4 at the current moment as a danger; the second case is applicable to the two cases that the rear vehicle is a complete vehicle or a partial vehicle.
Thirdly, when the rear vehicle is a complete vehicle, if the central point A1 of the rectangular positioning frame A falls in a left lane visible area 3-2 of the rear view image 1, calculating the left lane prejudgment time of the rear vehicle reaching the left lane blind area 3-1; then, when the time counted from the current time reaches the left lane prediction time, the time counted from the current time is a left lane timing end time, and if the rear vehicle does not appear in the rear view image 1 at the left lane timing end time (the rear vehicle may suddenly accelerate within the left lane prediction time and reach a dead zone of the rear camera, for example, when the rear vehicle and a merging vehicle are side by side), it is determined that the merging vehicle is dangerous to merge into the left lane 3 at the left lane timing end time;
fourthly, when the rear vehicle is a complete vehicle, if the central point A1 of the rectangular positioning frame A falls into the right lane visible area 4-2 of the rear view image 1, calculating the right lane prejudgment time when the rear vehicle reaches the right lane blind area 4-1; and starting timing from the current time, and when the timing reaches the right lane pre-judgment time, the timing is the right lane timing end time, if the rear vehicle does not appear in the rear view image 1 at the right lane timing end time, it is judged that the merging vehicle merges to the right lane 4 at the right lane timing end time as a danger;
and judging that the doubling vehicle is safe when being doubled at the current moment except the four situations.
Therefore, the invention utilizes the streaming media inside rear-view mirror widely used on the vehicle, tracks the position information of the rear vehicle, and judges whether the running conditions of the parallel vehicle and the rear vehicle belong to one of four situations of parallel danger or not based on the preset partition information of the rear view image 1, so as to realize the auxiliary function of providing parallel safety information for the driver when preparing for parallel; in addition, the invention does not need to add additional equipment on the vehicle, and has the advantage of low cost.
Example two
On the basis of the first embodiment, the second embodiment also adopts the following preferred embodiments:
the processor is also provided with an early warning module;
the early warning module is used for: the access the indicator light information of doubling vehicle to based on the doubling safety information that doubling auxiliary module produced, produce doubling early warning signal, include:
when the information of the steering lamps of the doubling vehicle is left-turning, if the doubling safety information is the situation I or the situation III, generating a doubling early warning signal;
and when the information of the steering lamp of the doubling vehicle is right turning, if the doubling safety information is the situation two or the situation four, generating the doubling early warning signal.
Therefore, the invention utilizes the early warning module to actively warn a driver that the parallel wire is dangerous by generating the parallel wire early warning signal when the parallel wire vehicle prepares the parallel wire and accords with one of four parallel wire dangers, thereby greatly improving the driving safety.
The above is the basic implementation of the second embodiment, and further optimization, improvement and limitation can be performed on the basis of the basic implementation:
preferably: the doubling auxiliary system also comprises an early warning execution device; the early warning execution device can send out an alarm when the early warning module generates a parallel early warning signal. For example: the early warning execution device can be a display screen of the streaming media inside rearview mirror, and gives an alarm by displaying information on the display screen; alternatively, the warning executing device may be a buzzer, a warning lamp, a vibrator, or the like, and the warning executing device may issue a warning by sound, light, vibration, or the like.
EXAMPLE III
On the basis of the first embodiment or the second embodiment, the third embodiment further adopts the following preferred embodiments:
the image processing module is preset with a vehicle position detection model obtained based on training of a MobileNet-V2 neural network model, and the rear-view image 1 is converted into data in a TFRecord format and then input into the vehicle position detection model to obtain position information of the rear vehicle.
Therefore, the vehicle position detection model obtained based on the training of the MobileNet-V2 neural network model is adopted, the position information of the rear vehicle in the rear view image 1 is identified, the operation speed is high, and the method and the device are suitable for the existing streaming media inside rearview mirror.
The above is the basic implementation of the third embodiment, and further optimization, improvement and limitation can be made on the basis of the basic implementation:
preferably, the following components: as shown in fig. 3, the manner of training the vehicle position detection model is as follows:
step one, collecting a plurality of rear view images 1, wherein the quantity of the rear view images accords with the training requirement of a MobileNet-V2 neural net model, and obtaining image data samples;
step two, manually marking the position information and the vehicle information of the rear vehicle in each rear view image 1 to obtain a position information sample and a vehicle information sample, wherein the vehicle information indicates that the rear vehicle is a complete vehicle or a local vehicle;
converting the image data samples, the corresponding position information samples and the corresponding vehicle information samples into TFRecord format data, and dividing the TFRecord format data into a training set and a test set according to a preset proportion;
inputting the training set into a MobileNet-V2 neural network model for training to obtain a model to be verified;
and step five, verifying the model to be verified by using the test set, wherein if the model to be verified passes the verification, the model to be verified is the vehicle position detection model, and otherwise, the step one to the step five are executed again.
Example four
On the basis of any one of the first to third embodiments, the fourth embodiment further adopts the following preferred embodiments:
in the third case, the calculation method of the left lane pre-judgment time is as follows: the method comprises the steps of firstly calculating the speed of a rear vehicle at the current moment according to the position information of the rear vehicle in N continuous rear view images 1, and then calculating the left lane prejudgment time by combining the distance between the rear vehicle at the current moment and a left lane blind area 3-1, wherein N is a positive integer larger than 2.
In the fourth situation, the right lane anticipation time is calculated in the following manner: the method comprises the steps of firstly calculating the speed of a rear vehicle at the current moment according to the position information of the rear vehicle in N continuous rear view images 1, and then calculating the right lane prejudgment time by combining the distance between the rear vehicle at the current moment and a right lane blind area 4-1, wherein N is a positive integer larger than 2.
The present invention is not limited to the above embodiments, and various other equivalent modifications, substitutions and alterations can be made without departing from the basic technical concept of the invention as described above, according to the common technical knowledge and conventional means in the field.
Claims (5)
1. A merging auxiliary system based on a streaming media inside rear view mirror comprises the streaming media inside rear view mirror installed on a merging vehicle, wherein the streaming media inside rear view mirror is provided with a rear camera and a processor, the processor can collect a rear view image (1) behind the merging vehicle in real time through the rear camera, the rear view image (1) covers a current lane (2) where the merging vehicle is located, and a left lane (3) and a right lane (4) which are respectively located on the left side and the right side of the current lane (2); the method is characterized in that:
the processor is provided with an image processing module and a doubling auxiliary module;
the image processing module is configured to: acquiring one rear view image (1) every preset detection period, and identifying the position information of a rear vehicle displayed in the rear view image (1), wherein the position information comprises: -a rectangular positioning frame (a) having four sides tangent to the rear vehicle, and-the coordinates of a central point (A1) of the rectangular positioning frame (a) in the rear view image (1), wherein the rear vehicle comprises two situations: firstly, the rear vehicle is a whole vehicle which is completely displayed in the rear view image (1) and is marked as a complete vehicle, and secondly, the rear vehicle is a part of the whole vehicle which is displayed in the rear view image (1) and is marked as a vehicle part;
the image processing module is preset with a vehicle position detection model obtained based on training of a MobileNet-V2 neural network model, and the rear view image (1) is converted into TFRecord format data and then is input into the vehicle position detection model to obtain position information of the rear vehicle;
the mode of training to obtain the vehicle position detection model is as follows:
step one, collecting a plurality of rear view images (1) to obtain an image data sample;
step two, manually marking the position information and the vehicle information of the rear vehicle in each rear view image (1) to obtain a position information sample and a vehicle information sample, wherein the vehicle information indicates that the rear vehicle is a complete vehicle or a local vehicle;
converting the image data samples, the corresponding position information samples and the corresponding vehicle information samples into TFRecord format data, and dividing the TFRecord format data into a training set and a test set according to a preset proportion;
inputting the training set into a MobileNet-V2 neural network model for training to obtain a model to be verified;
step five, the test set is used for verifying the model to be verified, if the model to be verified passes the verification, the model to be verified is the vehicle position detection model, and if the model to be verified does not pass the verification, the step one to the step five are executed again;
the merging support module is pre-set with a left lane boundary (3 a) and a right lane boundary (4 a) as the division marks of the rear view image (1) such that: in the rear view image (1), the left lane (3) is positioned in the areas on both sides of the left lane boundary (3 a) and respectively is a left lane blind area (3-1) and a visible left lane visible area (3-2) which cannot be observed by the driver of the merging vehicle through a left rearview mirror, and the right lane (4) is positioned in the areas on both sides of the right lane boundary (4 a) and respectively is a right lane blind area (4-1) and a visible right lane visible area (4-2) which cannot be observed by the driver of the merging vehicle through a right rearview mirror;
the doubling auxiliary module is used for generating doubling safety information of the doubling vehicle and comprises:
in the first situation, if the central point (A1) of the rectangular positioning frame (A) is located in a left lane blind area (3-1) of the rear view image (1), judging that the merging vehicle merges towards the left lane (3) at the current moment to be dangerous;
in a second situation, if the central point (A1) of the rectangular positioning frame (A) falls in a right lane blind area (4-1) of the rear view image (1), judging that the merging vehicle is dangerous when merging to the right lane (4) at the current moment;
when the rear vehicle is a complete vehicle, if the central point (A1) of the rectangular positioning frame (A) falls into a left lane visible area (3-2) of the rear view image (1), calculating the left lane prejudgment time of the rear vehicle reaching the left lane blind area (3-1); starting timing from the current time, and when the timing reaches the left lane pre-judgment time, the timing is a left lane timing end time, if the rear vehicle does not appear in the rear view image (1) at the left lane timing end time, the fact that the merging vehicle merges to the left lane (3) at the left lane timing end time is judged to be dangerous;
when the rear vehicle is a complete vehicle, if the central point (A1) of the rectangular positioning frame (A) is located in a right lane visible area (4-2) of the rear view image (1), calculating the right lane prejudgment time of the rear vehicle reaching the right lane blind area (4-1); and starting timing from the current time, and when the timing reaches the right lane pre-judgment time, the timing is the right lane timing end time, if the rear vehicle does not appear in the rear view image (1) at the right lane timing end time, the vehicle is judged to be dangerous when merging to the right lane (4) at the right lane timing end time;
and judging that the doubling vehicle is safe when being doubled at the current moment except the four situations.
2. The streaming media rearview mirror-based merge assisting system of claim 1, wherein: the processor is also provided with an early warning module;
the early warning module is used for: the access the indicator light information of doubling vehicle to based on the doubling safety information that doubling auxiliary module produced, produce doubling early warning signal, include:
when the information of the steering lamps of the doubling vehicle is left-turning, if the doubling safety information is the situation I or the situation III, generating a doubling early warning signal;
and when the turn signal information of the doubling vehicle is right turn, if the doubling safety information is the second situation or the fourth situation, generating the doubling early warning signal.
3. The streaming media rearview mirror-based merge assist system of claim 2, wherein: the doubling auxiliary system also comprises an early warning execution device; the early warning execution device can send out an alarm when the early warning module generates a parallel early warning signal.
4. The streaming media-based rearview mirror doubling assistance system according to any one of claims 1 to 3, wherein: in the third case, the calculation method of the left lane pre-judgment time is as follows: the method comprises the steps of firstly calculating the speed of a rear vehicle at the current moment according to the position information of the rear vehicle in N continuous rear view images (1), and then calculating the left lane prejudgment time by combining the distance between the rear vehicle and a left lane blind area (3-1) at the current moment, wherein N is a positive integer larger than 2.
5. The streaming media-based rearview mirror doubling assistance system according to any one of claims 1 to 3, wherein: in the fourth case, the calculation method of the right lane pre-judgment time is as follows: the method comprises the steps of firstly calculating the speed of a rear vehicle at the current moment according to the position information of the rear vehicle in N continuous rear view images (1), and then calculating the right lane prejudgment time by combining the distance between the rear vehicle and a right lane blind area (4-1) at the current moment, wherein N is a positive integer larger than 2.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811170324 | 2018-10-09 | ||
| CN2018111703244 | 2018-10-09 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110633653A CN110633653A (en) | 2019-12-31 |
| CN110633653B true CN110633653B (en) | 2023-04-18 |
Family
ID=68969357
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910800923.8A Active CN110633653B (en) | 2018-10-09 | 2019-08-28 | Doubling auxiliary system based on streaming media inside rearview mirror |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110633653B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1726513A1 (en) * | 2005-05-02 | 2006-11-29 | Iveco S.p.A. | Driving assistance system for lane keeping support, for lane change assistance, and for driver status monitoring for a vehicle |
| CN108162866A (en) * | 2017-12-21 | 2018-06-15 | 宁波吉利汽车研究开发有限公司 | A kind of lane recognition system and method based on Streaming Media external rearview mirror system |
-
2019
- 2019-08-28 CN CN201910800923.8A patent/CN110633653B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1726513A1 (en) * | 2005-05-02 | 2006-11-29 | Iveco S.p.A. | Driving assistance system for lane keeping support, for lane change assistance, and for driver status monitoring for a vehicle |
| CN108162866A (en) * | 2017-12-21 | 2018-06-15 | 宁波吉利汽车研究开发有限公司 | A kind of lane recognition system and method based on Streaming Media external rearview mirror system |
Non-Patent Citations (1)
| Title |
|---|
| 基于Arduino单片机的汽车并线辅助系统设计;李明娟;《实验室研究与探索》;20171215(第12期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110633653A (en) | 2019-12-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10186147B2 (en) | Wrong-way determination apparatus | |
| CN112349144B (en) | Monocular vision-based vehicle collision early warning method and system | |
| WO2010077316A1 (en) | Multiple object speed tracking system | |
| CN113808418B (en) | Road condition information display system, method, vehicle, computer device and storage medium | |
| JP4528283B2 (en) | Vehicle periphery monitoring device | |
| CN111492416A (en) | Violation monitoring system and violation monitoring method | |
| CN112001208B (en) | Target detection method and device for vehicle blind area and electronic equipment | |
| CN112758088A (en) | Dangerous source reminding method and advanced driving assistance system | |
| CN108766001A (en) | A kind of automobile data recorder for prompting traffic lights | |
| JP4813304B2 (en) | Vehicle periphery monitoring device | |
| JP6459809B2 (en) | Image processing apparatus, traffic management system including the same, and image processing method | |
| CN110633653B (en) | Doubling auxiliary system based on streaming media inside rearview mirror | |
| CN111815958B (en) | Traffic electronic police snapshot control device and method | |
| CN112907979A (en) | Motor vehicle illegal driving track tracking system and method based on multiple cameras | |
| JP4848644B2 (en) | Obstacle recognition system | |
| CN107009964A (en) | A kind of parked servicing unit of high-precision vehicle | |
| JPH0628596A (en) | Image processing type overspeed warning device | |
| CN114170798B (en) | Message reminding system and method | |
| WO2008037473A1 (en) | Park assist system visually marking up dangerous objects | |
| CN211349574U (en) | 360-degree all-round-looking early warning system for vehicle | |
| KR102764189B1 (en) | A device and system for sending emergency alerts on roads using CCTV, sound and VMS | |
| CN113570901B (en) | Vehicle driving assisting method and device | |
| JP7145227B2 (en) | sign recognition device | |
| CN212604814U (en) | Blind spot detection and early warning device for large-scale operating machinery and equipment | |
| KR20170129466A (en) | System for Displaying a Conditional Limiting Speed of Vehicle |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |