阅读说明：本技术 一种基于后拉摄像头图像adas分析的倒车安全系统 (Back-up safety system based on analysis of back-pull camera image ADAS ) 是由 黎逵 孙建伟 孙德生 于 2019-09-25 设计创作，主要内容包括：本发明涉及一种基于后拉摄像头图像ADAS分析的倒车安全系统,包括后拉摄像头、CPU数据处理模块、预警诊断模块、APP处理模块、显示模块以及人机交互显示模块以及人机交互语音模块；所述后拉摄像头用于将获取倒车图像数据并将该数据输入至CPU数据处理模块；所述CPU数据处理模块用于将图像数据按每帧的形式发送至预警诊断模块；所述预警诊断模块用于图像数据识别成相应的物体模型信息并将该模型信息返回至CPU数据处理模块；所述CPU数据处理模块用于将相应的物体模型信息发送至APP处理模块；所述人机交互显示模块用于将APP处理模块处理好的物体模型信息进行显示,所述人机交互语音模块用于将APP处理模块处理好的物体模型信息转成危险级别语音信号进行播放。(The invention relates to a reversing safety system based on ADAS analysis of a rear-pull camera image, which comprises a rear-pull camera, a CPU data processing module, an early warning diagnosis module, an APP processing module, a display module, a human-computer interaction display module and a human-computer interaction voice module, wherein the CPU data processing module is used for processing a back-pull camera image; the rear pull camera is used for acquiring backing image data and inputting the data to the CPU data processing module; the CPU data processing module is used for sending the image data to the early warning diagnosis module in a form of each frame; the early warning diagnosis module is used for identifying the image data into corresponding object model information and returning the model information to the CPU data processing module; the CPU data processing module is used for sending corresponding object model information to the APP processing module; the man-machine interaction display module is used for displaying the object model information processed by the APP processing module, and the man-machine interaction voice module is used for converting the object model information processed by the APP processing module into dangerous level voice signals to be played.)

1. The utility model provides a safety coefficient that backs a car based on analysis of back pull camera image ADAS which characterized in that: the system comprises a back-pull camera, a CPU data processing module, an early warning diagnosis module, an APP processing module, a human-computer interaction display module and a human-computer interaction voice module;

the rear pull camera is used for acquiring backing image data and inputting the data to the CPU data processing module;

the CPU data processing module is used for sending the image data to the early warning diagnosis module in a form of each frame;

the early warning diagnosis module is used for identifying the image data into corresponding object model information and returning the model information to the CPU data processing module;

the CPU data processing module is used for sending corresponding object model information to the APP processing module;

the man-machine interaction display module is used for displaying the object model information processed by the APP processing module, and the man-machine interaction voice module is used for converting the object model information processed by the APP processing module into dangerous level voice signals to be played.

2. The reverse safety system of claim 1, wherein the early warning diagnosis module is configured to determine a region of interest of the image by filtering analysis of each frame of image data, and then extract the body model information using an adaboost iterative algorithm.

3. The reverse safety system of claim 2, wherein: the early warning diagnosis module firstly extracts the picture characteristics of each area in each frame of image data, then compares the characteristics with the pre-stored object model information, identifies and screens out the corresponding object model information and determines the coordinate position of the object model on the image.

4. The reverse safety system of claim 3, wherein: the man-machine interaction display module can display the reversing image data acquired by the rear-pull camera into an image plane, an alarm line is displayed on a vertical middle horizontal line of the image plane, the alarm line can move up and down on the image plane, the image plane is divided into an early warning monitoring area and an early warning area by the alarm line, and when an object model appears on the image plane and falls into the early warning monitoring area, the man-machine interaction voice module cannot give an alarm; when an object appears on the image plane and falls into the early warning alarm area, the man-machine interaction voice module gives an alarm, and the function of reminding a driver is enhanced through TTS voice broadcasting 'please notice to back a car'.

Technical Field

The invention relates to a reversing safety system based on ADAS analysis of a rear-pull camera image.

Background

With the continuous and high-speed increase of economy in China, the automobile industry is also rapidly developed, automobiles have entered thousands of households, the number of motor vehicles in China currently exceeds 3.7 hundred million by 6 months in 2019, and the number of cars is two hundred million, which is huge. The automobiles bring great convenience to people and bring many safety problems, for example, accidents caused by automobile backing show a trend of increasing year by year. However, in order to avoid the above car backing accidents, radar car backing anti-collision instruments are adopted in the prior art, and the early radar car backing anti-collision instruments can detect obstacles in a certain distance range behind a car, so as to give an alarm, and then, the alarm is developed according to the distance in a segmented mode. Although the radar anti-collision instrument for backing a car can solve the car backing accident to a certain extent, some defects still exist, for example: the driver can not visually see the field of vision behind the vehicle, and a large number of accidents can be caused due to the dead angle problem and the coverage problem of the radar. The existing millimeter wave radar configured for some high-end vehicles can quickly identify obstacles behind the vehicles, but even if the obstacles exist, no image display is available, it is unclear what specific obstacles are, how far away from the vehicles is, the experience feeling is not good, the cost is very high, and the technology cannot be popularized and used by medium-low-end vehicles.

In the current market, the intelligent rearview mirrors popular with car owners are basically equipped with a reversing image function, and the reversing image display function supported after the system is started is only the display of images. When a vehicle owner backs a vehicle, the vehicle owner needs to look at the display screen all the time to observe the rear dynamic state, sometimes the vehicle owner needs to observe not only the rear of the vehicle but also the driver needs to observe the left side, the right side or the front state, so that the driver cannot always look at the screen to back the vehicle, and when the driver leaves the rear view, if pedestrians, children, electric vehicles or other vehicles enter the rear of the vehicle, the driver cannot observe the situation, traffic accidents easily occur. The experience is poor, and the backing safety is also greatly discounted.

Disclosure of Invention

The invention aims to solve the defects of the prior art, and provides a backing safety system based on ADAS analysis of a back-pull camera image.

The rear pull camera is used for acquiring backing image data and inputting the data to the CPU data processing module;

the CPU data processing module is used for sending the image data to the early warning diagnosis module in a form of each frame;

the early warning diagnosis module is used for identifying the image data into corresponding object model information and returning the model information to the CPU data processing module;

the CPU data processing module is used for sending corresponding object model information to the APP processing module;

the man-machine interaction display module is used for displaying the object model information processed by the APP processing module, and the man-machine interaction voice module is used for converting the object model information processed by the APP processing module into dangerous level voice signals to be played.

Further, the early warning diagnosis module is used for determining a region of interest (ROI) of the image through filtering analysis of image data of each frame, and then extracting body model information by adopting an adaboost iterative algorithm.

Further, the early warning diagnosis module firstly extracts picture features of the region of interest, and then compares the picture features with pre-stored object model information to identify and screen out corresponding object model information and coordinate positions of the object models on the image.

Furthermore, the man-machine interaction display module can display the reversing image data acquired by the rear-pull camera into an image plane, an alarm line is displayed on a vertical middle horizontal line of the image plane, the alarm line can move up and down on the image plane, the alarm line divides the image plane into an early warning monitoring area and an early warning area, and when an object model appears on the image plane and falls into the early warning monitoring area, the man-machine interaction voice module cannot give an alarm; when an object appears on the image plane and falls into the early warning alarm area, the man-machine interaction voice module gives an alarm, and the function of reminding a driver is enhanced through TTS voice broadcast

The invention has the beneficial effects that: the invention adds an ADAS intelligent recognition function on the basis of the automobile intelligent rearview mirror to realize that pedestrians, children, electric bicycles, other automobiles and the like suddenly rushing behind the automobile can be quickly recognized in the process of backing, and the automobile owner is reminded in time through prompt tones, voice characters and pictures and texts, thereby avoiding traffic accidents and improving the safety, reliability and effectiveness of backing. The driving safety is improved to the maximum extent.

Drawings

FIG. 1 is a block diagram of the system of the present invention;

FIG. 2 is a flow chart of the early warning diagnostic module of the system of the present invention;

FIG. 3 is a schematic diagram of a back-pull camera imaging system of the present invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

Referring to fig. 1, a car backing safety system based on analysis of a back-pull camera image ADAS comprises a back-pull camera, a CPU data processing module, an early warning diagnosis module, an APP processing module, a human-computer interaction display module and a human-computer interaction voice module.

The rear pull camera is used for acquiring backing image data and inputting the data to the CPU data processing module;

the CPU data processing module is used for sending the image data to the early warning diagnosis module in a form of each frame;

the early warning diagnosis module is used for identifying the image data into corresponding object model information and returning the model information to the CPU data processing module;

the CPU data processing module is used for sending corresponding object model information to the APP processing module;

the man-machine interaction display module is used for displaying the object model information processed by the APP processing module, and the man-machine interaction voice module is used for converting the object model information processed by the APP processing module into dangerous level voice signals to be played.

Further, referring to fig. 2, the forewarning diagnosis module is configured to determine a region of interest (ROI region) of the image through a filtering analysis of each frame of image data, and then extract the body model information using the adaboost iterative algorithm. Specifically, firstly, the picture characteristics of the region of interest are extracted, and then the picture characteristics are compared with the pre-stored object model information to identify and screen out the corresponding object model information and the coordinate position of the object model on the image.

In order to obtain good identification effect, target small areas with multiple sizes, such as 20X20, 36X36 and 48X48, need to be traversed and screened, and the identified effect is the best in several traversals.

In the invention, the reversing picture data is analyzed by complete ADAS to obtain corresponding object model information and the information is sent to the APP processing module by the CPU data processing module, the APP processing module outputs the information to the man-machine interaction display module for display after corresponding logic operation, and the danger degree prompt tone is played by the man-machine interaction voice module, so that a driver can perform corresponding operation according to the danger degree prompt tone and the content displayed by the display module, TTS voice broadcast can be added for determining the type of the corresponding object model, such as cautious pedestrians and please notice of reversing! The driver is reminded in time to pay attention to the rear safety, and corresponding object modules are drawn on the right side of the LCD display screen at the left rear part, the right rear part and the rear part of the vehicle or in three conditions, so that the driver can pay attention to the condition.

Because the intelligent rear-view mirror that supports the image function of backing a car at present does not basically install when the loading factory and accomplish, if will install such intelligent rear-view mirror on the automobile body of different grade type, for example car, SUV, small truck, van etc. this moment, in order to let the image of shooing have better visual angle scope, the position of back pull camera installation just need to correspond with corresponding automobile body, for example: the SUV is higher by a little, and the height of the camera above the ground is higher by a little; the car is a little short, and the height of the camera from the ground is a little short. In addition, the height of the camera from the ground also affects the imaging range and the distance feeling difference of the back-pull camera. Even if the same type of vehicle is used, the rear pull cameras are arranged at different positions of the vehicle, such as the middle position, the left position or the right position, and the imaging range and the distance feeling are different. Therefore, the factors are comprehensively considered in the invention, a reasonable early warning alarm range is determined in the imaging range, specifically, as shown in fig. 3, the back-pull camera sets the vertical middle horizontal line of the formed image plane as an alarm line, the alarm line divides the image plane into an early warning monitoring area and an early warning area, the early warning area is an area nearest to the tail of the vehicle, and the alarm line can move up and down on the image plane. The user can move up and down according to the condition of the vehicle body to adjust the position of the alarm line on the image plane, and can enlarge or reduce the early warning video range according to the driving level of the user.

The function of the alarm line in the invention is further explained, the back-pull camera displays the acquired image plane on the man-machine interaction display module, and the alarm line can be displayed in the image plane, the alarm line divides the image plane into an early warning monitoring area and an early warning area, when an object appears on the image plane and falls into the early warning monitoring area, the man-machine interaction voice module can not give an alarm; when an object appears on the image plane and falls into the early warning alarm area, the human-computer interaction voice module gives an alarm, and meanwhile, the function of reminding a driver through TTS voice broadcasting 'please notice to back a car' is enhanced.

Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.