CN106092114A - The automobile real scene navigation apparatus of a kind of image recognition and method - Google Patents

Abstract

The invention discloses an image recognition vehicle real-scene navigation device and method, comprising a video analysis module, an image recognition module, a GPS positioning module, a route planning module, a map management module, and a real-scene navigation management module; the video analysis module and the image The identification module is connected; the real scene navigation management module is connected with the image recognition module, the GPS positioning module, the map management module, and the path planning module respectively; the map display is provided by the car navigation equipment, and the path is marked, and the image is carried out in conjunction with the image collected by the driving recorder Recognition, especially for the recognition of traffic signs with variable lanes, prompts and reminds users how to choose a turning lane or a straight lane when approaching an intersection, and reminds users to drive or park according to the actual traffic signal conditions to realize real-world navigation. Added new functions of car navigation and enhanced user experience.

Description

A real-scene car navigation device and method for image recognition

technical field

The invention belongs to the technical field of automobile control, and in particular relates to an image recognition real-scene navigation device and method for automobiles.

Background technique

With the popularization and development of digital camera technology, GPS positioning technology and wireless Internet, related car navigation equipment, driving recorder and other car auxiliary equipment have the video acquisition capability of 1080P or higher resolution, centimeter-level positioning GPS, stable 100Mbps wireless transmission rate. Traditional driving recorders generally record the video images and sounds of the whole process of car driving, without other extended functions; traditional car navigation devices generally use built-in or network maps combined with GPS positioning technology, and use map display and route marking methods for car navigation , without combining the images collected by the driving recorder for real-scene navigation; some car real-scene navigation devices have the method of collecting real-scene video combined with map display and route marking, or calculating the traffic flow in multiple road lanes, but they cannot be combined with image recognition Technology that prompts and reminds users how to choose a turning lane or a straight lane when approaching an intersection. It cannot judge the traffic signs of variable lanes, and cannot remind users to drive or park according to the actual traffic signal conditions.

Contents of the invention

The technical problem to be solved by the present invention is to provide a car real-scene navigation device and method for image recognition, using existing video analysis technology, image recognition technology, GPS positioning technology, path planning technology, etc., through the design of information technology and hardware equipment , to realize the real scene navigation function of the car.

In order to solve the problems of the technologies described above, the specific technical solutions adopted in the present invention are as follows:

A car real-scene navigation device for image recognition, comprising: a video analysis module, an image recognition module, a GPS positioning module, a path planning module, a map management module, and a real-scene navigation management module;

The video analysis module is connected with the image recognition module;

Described real scene navigation management module is connected with image recognition module, GPS positioning module, map management module, route planning module respectively;

The map management module is used to provide relevant road information of the driving area of the vehicle;

The real-scene navigation management module is used to integrate real-scene image recognition information, longitude and latitude information, path information and road information, and display it in real time on the display interface;

A car real-scene navigation method for image recognition is as follows:

Real-world navigation management module executes the following steps cyclically

Step S1, the GPS positioning module extracts the latitude and longitude information of the vehicle;

Step S2, the map management module is used to provide relevant road information of the driving area of the vehicle;

In step S3, the route planning module provides lane selection methods such as turning around, turning left, going straight or turning right at the previous intersection according to the latitude and longitude information of the car, relevant road information and the route information required to reach the destination;

Step S4, the video analysis module collects the real-time video information of the running of the car, analyzes and extracts video key frames, and stores them as key frame image files;

Step S5, the image recognition module analyzes the key frame image file to analyze the current moment, if it is a daytime mode, go to step S6, otherwise it is a night mode, go to step S7;

Step S6, setting daytime mode parameters, and turning to step S8;

Step S7, set the night mode parameters, and go to step S8;

Step S8, the image recognition module analyzes the key frame image file, traffic sign information, road traffic marking information and traffic signal light information, and stores the information;

Step S9, judging whether the road traffic marking information is an intersection lane marking, if yes, proceed to step S10, if not, proceed to step 12;

Step S10, judging whether there is traffic sign information within the time threshold T1, if yes, then go to step S11, if not, go to step 12;

Step S11, judging the U-turn, left turn, straight line or right turn in the traffic sign information, mapping it to the road traffic marking information U-turn, left turn, straight line or right turn, according to the lane selection method provided by the path planning module, determine Lane selection information;

Step S12, judging whether there is traffic signal light information, if yes, proceed to step S13, if not, proceed to step 16;

Step S13, judging whether there is lane selection information within the time threshold T2, if yes, then go to step S14, if not, go to step 15;

Step S14, according to the lane selection information, judge the lane release marking situation in the traffic signal light information: driving or parking, go to step S16;

Step S15, directly judging the status of the lane clearance marking in the traffic signal light information: driving or parking;

Step S16, the real-scene navigation management module integrates the real-scene image recognition information, latitude and longitude information, path information and road information, and displays the real scene on the display interface to provide the driver with lane selection, driving or parking prompts;

The present invention has beneficial effects: the present invention provides a new method for car real-scene navigation of image recognition, which provides map display and path marking through car navigation equipment, and performs image recognition in combination with images collected by driving recorders, especially for variable lanes The recognition of traffic signs, prompts and reminds users how to choose a turning lane or straight lane when approaching an intersection, and reminds users to drive or park according to the actual traffic signal conditions, realizing real-world navigation. Added new functions of car navigation and increased user experience.

The specific beneficial effects of the patent of the present invention can be reflected in the following aspects:

The video analysis module adopts key frame technology, which can remove most of the repeated video frames in the live video information, reduce the matching data calculation amount, and improve the processing efficiency;

The image recognition module uses median filtering and morphological processing methods to eliminate noise in key frame image files, reduces the error rate of matching, uses the method of constructing polygonal boundaries to identify regional boundaries, and improves the recognition rate of specific targets. The translation, rotation and scaling of the target are invariant, so that the specific target has a higher recognition rate;

In an image recognition method for car real scene navigation, by analyzing key frame image files, it is analyzed that the current moment is a day or night mode, and different parameters suitable for the day or night mode can be used to target different scenes during the day or night. Has a higher recognition rate;

In an image recognition method for real-world vehicle navigation, by collecting and judging: whether the road traffic marking information on the ground is an intersection lane marking, the U-turn, left-turn, straight-going or right-turn status in the traffic sign information set up on the roadside, Multi-aspect comprehensive information related to real-world navigation, such as traffic signal information at intersections, and multi-aspect information sources, enable drivers to choose lanes, and calculate decision-making information for driving or parking with higher accuracy;

Description of drawings

FIG. 1 is a schematic diagram of an overall structure of an image-recognition vehicle real-scene navigation device.

Fig. 2 is a flow chart of an image recognition method for car real scene navigation.

Fig. 3 is a schematic diagram of relevant map road information and route planning in the driving area of the vehicle.

Figure 4 is a key frame of traffic sign information in a daytime scene.

Fig. 5 is a schematic diagram of extracting traffic sign information in a daytime scene.

Figure 6 is a key frame of road traffic marking information in a daytime scene.

Fig. 7 is a schematic diagram of extracting road traffic marking information from a daytime scene.

Figure 8 is a key frame of the red light information of the traffic lights in the daytime scene.

FIG. 9 is a schematic diagram of extracting red light information of traffic lights in a daytime scene.

Fig. 10 is a key frame of traffic signal green light information in a daytime scene.

Fig. 11 is a schematic diagram of extracting green light information of traffic lights in a daytime scene.

Figure 12 is a key frame of traffic sign information in a variable lane scene.

Fig. 13 is a schematic diagram of extracting traffic sign information in a scene with variable lanes.

Fig. 14 is a key frame of road traffic marking information in a variable lane scene.

Fig. 15 is a schematic diagram of extracting road traffic marking information in a scene with variable lanes.

Fig. 16 is a key frame of traffic sign information in the U-turn scene.

Fig. 17 is a schematic diagram of extracting traffic sign information in a U-turn scene.

Fig. 18 is a key frame of road traffic marking information in a U-turn scene.

Fig. 19 is a schematic diagram of extracting road traffic marking information in a U-turn scene.

Fig. 20 is a key frame of traffic sign information in a night scene.

FIG. 21 is a schematic diagram of extracting traffic sign information in a night scene.

Fig. 22 is a key frame of road traffic marking information in a night scene.

Fig. 23 is a schematic diagram of extracting road traffic marking information in a night scene.

Reference numerals: 1-video analysis module, 2-image recognition module, 3-GPS positioning module, 4-route planning module, 5-map management module, 6-real scene navigation management module.

detailed description

The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention.

It can be seen from the overall structural diagram of an image recognition vehicle real scene navigation device shown in Figure 1 that it includes 1-video analysis module, 2-image recognition module, 3-GPS positioning module, 4-path planning module, 5-map Management module, 6-real scene navigation management module.

Described video analysis module 1 is connected with image recognition module 2;

Described real scene navigation management module 6 is connected with image recognition module 2, GPS positioning module 3, map management module 4, path planning module 5 respectively;

Described video analysis module 1 is used for analyzing the real scene video information that equipment collects, extracts video key frame, is stored as key frame image file;

The image recognition module 2 is used to recognize the traffic sign information in the key frame image file, the road traffic marking information, and the traffic signal light information;

Described GPS positioning module 3 is used for providing the longitude and latitude information of automobile driving;

The path planning module 4 is used to provide path information needed to reach the destination;

The map management module 5 is used to provide relevant road information of the driving area of the automobile;

The real-scene navigation management module 6 is used to integrate real-scene image recognition information, latitude and longitude information, route information and road information, and display it in real time on the display interface;

The video analysis module is used to analyze the real-scene video information collected by the equipment; extract the video key frame and store it as a key frame image file; the key frame is a combination of one or more frames of main content in the real-scene video information, and The extraction method is to judge the similarity between all consecutive image frames in the real-scene video information, using visual characteristics such as color, motion, etc. as the measurement standard;

The image recognition module is used to recognize the traffic sign information in the key frame image file, the road traffic marking information, and the traffic signal light information; the specific method is: firstly, the median filter and the morphological processing method in the image recognition are used to eliminate The noise in the key frame image file, and then use the method of constructing polygonal boundary in image recognition to identify the area boundary and identify the specific target, and finally use the calculation of the specific target invariant moment characteristics and traffic signs, road traffic markings, traffic lights The distance of the invariant moment feature, to judge the actual information of the specific target;

The GPS positioning module adopts the L1 carrier with a frequency of 1575.42MHz and the L2 carrier with a frequency of 1227.60MHz to provide the latitude and longitude information of the vehicle;

The path planning module uses the Dijkstra algorithm to provide path information needed to reach the destination.

Below are the embodiments given by the inventor:

The process is shown in Figure 2.

Example 1: Car real scene navigation daytime scene

Carry out the following steps sequentially from the driving route

Frame image processing steps shown in Figure 4:

GPS positioning module 3 extracts the latitude and longitude information of automobile driving;

Map management module 5 provides the relevant road information of the automobile driving area as shown in Figure 3 according to the longitude and latitude information of automobile driving;

The path planning module 4 provides the prompt for the automobile to go straight at the previous intersection according to the latitude and longitude information of the automobile, the relevant road information and the path information required to reach the destination;

Video analysis module 1 collects the real-time video information of automobile operation, analyzes and extracts video key frames, and stores them as key frame image files as shown in Figure 4;

Image recognition module 2 analyzes the key frame image file, and analyzes that the current moment is the daytime mode;

Set day mode parameters, adjust appropriate brightness, contrast, noise removal parameters;

Image recognition module 2 analyzes the key frame image file as shown in Figure 4, traffic sign board information, road traffic marking information and traffic signal light information as shown in Figure 5, and store information;

Judging whether the road traffic marking information is an intersection lane marking, no;

Determine whether there is traffic signal light information, no;

Real-scene navigation management module 6 integrates real-scene image recognition information, latitude and longitude information, path information and road information, and displays the real scene on the display interface to provide driving tips for drivers;

Frame image processing steps shown in Figure 6

Image recognition module 2 analyzes the key frame image file as shown in Figure 6, traffic sign board information, road traffic marking information and traffic signal light information as shown in Figure 7, and store information;

Judging whether the road traffic marking information is an intersection lane marking, yes;

Judging whether there is traffic sign information within the time threshold T1, T1=100 seconds, yes;

Judging the traffic sign information shown in Figure 5 identified from Figure 4, from left to right: go straight 1, go straight 2, turn right; map it to the traffic sign shown in Figure 7 identified from Figure 6 Road traffic marking information, from left to right: go straight 1, go straight 2, turn right; according to the prompt of going straight at the intersection provided by the path planning module 4, determine whether to choose the lane to go straight 1 or go straight 2;

Judging whether the traffic signal light information is recognized from Fig. 6, as shown in Fig. 7, no;

Real-scene navigation management module 6 integrates real-scene image recognition information, latitude and longitude information, route information and road information, and displays the real scene on the display interface to provide drivers with lane selection prompts;

Frame image processing steps shown in Figure 8

Image recognition module 2 analyzes the key frame image file as shown in Figure 8, traffic sign board information, road traffic marking information and traffic signal light information as shown in Figure 9, and store information;

Judging whether the road traffic marking information is an intersection lane marking, yes;

Judging whether there is traffic sign information within the time threshold T1, T1=100 seconds, yes;

Judging the traffic sign information shown in Figure 5 identified from Figure 4, from left to right: go straight, go straight, turn right; map it to the road traffic identified from Figure 6 as shown in Figure 7 Marking information, from left to right: go straight, go straight, turn right; according to the prompt of going straight at the intersection provided by the path planning module 4, it is determined to choose the lane to go straight 2;

Judging whether the traffic signal light information is recognized from Fig. 8, as shown in Fig. 9, yes;

Judging whether there is lane selection information within the time threshold T2, T2=90 seconds, yes;

According to the lane selection information, judge the lane release sign in the traffic signal information, because it is a red light, and there is a waiting time of 24 seconds, then park;

Real-scene navigation management module 6 integrates real-scene image recognition information, latitude and longitude information, path information and road information, and displays the real scene on the display interface to provide parking prompts for drivers;

Frame image processing steps shown in Figure 10

Image recognition module 2 analyzes the key frame image file as shown in Figure 10, traffic sign board information, road traffic marking information and traffic signal light information as shown in Figure 11, and store information;

Judging whether the traffic signal light information is identified from Figure 10, as shown in Figure 11, yes;

Judging whether there is traffic sign information within the time threshold T1, T1=100 seconds, yes;

Judging the traffic sign information shown in Figure 5 identified from Figure 4, from left to right: go straight, go straight, turn right; map it to the road traffic identified from Figure 6 as shown in Figure 7 Marking information, from left to right: go straight, go straight, turn right; according to the prompt of going straight at the intersection provided by the path planning module 4, it is determined to choose the lane to go straight 2;

Judging whether the traffic signal light information is identified from Figure 10, as shown in Figure 11, yes;

Judging whether there is lane selection information within the time threshold T2, T2=90 seconds, yes;

According to the lane selection information, judge the lane release sign in the traffic signal light information, and drive because it is a green light;

Real-scene navigation management module 6 integrates real-scene image recognition information, latitude and longitude information, path information and road information, and displays the real scene on the display interface to provide driving tips for drivers;

Embodiment 2: Vehicle real scene navigation variable lane scene;

Carry out the following steps sequentially by the driving route of the car;

Frame image processing steps shown in Figure 12:

GPS positioning module 3 extracts the latitude and longitude information of automobile driving;

The map management module 5 provides relevant road information of the driving area of the vehicle according to the latitude and longitude information of the driving of the vehicle;

The path planning module 4 provides the prompt for the automobile to go straight at the previous intersection according to the latitude and longitude information of the automobile, the relevant road information and the path information required to reach the destination;

Video analysis module 1 collects the real-time video information of automobile operation, analyzes and extracts video key frames, and is stored as a key frame image file as shown in Figure 12;

Image recognition module 2 analyzes the key frame image file, and analyzes that the current moment is the daytime mode;

Set day mode parameters, adjust appropriate brightness, contrast, noise removal parameters;

Image recognition module 2 analyzes the key frame image file as shown in Figure 12, and the traffic sign information, road traffic marking information and traffic signal light information are as shown in Figure 13, and store information;

Judging whether the road traffic marking information is an intersection lane marking, no;

Determine whether there is traffic signal light information, no;

Real-scene navigation management module 6 integrates real-scene image recognition information, latitude and longitude information, path information and road information, and displays variable lanes in real scene on the display interface to provide driving tips for drivers;

Frame image processing steps shown in Figure 14

Image recognition module 2 analyzes the key frame image file as shown in Figure 14, and the traffic sign information, road traffic marking information and traffic signal light information are as shown in Figure 7, and store information;

Judging whether the road traffic marking information is an intersection lane marking, yes;

Judging whether there is traffic sign information within the time threshold T1, T1=100 seconds, yes;

Judging the traffic sign information shown in Figure 13 identified from Figure 12, from left to right: turn left, go straight 1, go straight 2, turn right; map it to the traffic sign information identified in Figure 14 as shown in Figure 15 For the road traffic marking information shown, the order from left to right before mapping is: turn left, variable lane, go straight, variable lane; after mapping, the order from left to right is: turn left, go straight 1, go straight 2, right Turn; According to the prompt of the crossing going straight provided by the path planning module 4, it is determined to select the lane going straight 2;

Judging whether the traffic signal light information is identified from Figure 14, as shown in Figure 15, no;

Real-scene navigation management module 6 integrates real-scene image recognition information, latitude and longitude information, route information and road information, and displays the real scene on the display interface to provide drivers with lane selection prompts;

Embodiment 3: U-turn scene of car real scene navigation

Carry out the following steps sequentially from the driving route

Frame image processing steps shown in Figure 16:

GPS positioning module 3 extracts the latitude and longitude information of automobile driving;

The map management module 5 provides relevant road information of the driving area of the vehicle according to the latitude and longitude information of the driving of the vehicle;

The path planning module 4 provides the prompt for the automobile to go straight at the previous intersection according to the latitude and longitude information of the automobile, the relevant road information and the path information required to reach the destination;

Video analysis module 1 collects the real-time video information of automobile operation, analyzes and extracts video key frames, and stores them as key frame image files as shown in Figure 16;

Image recognition module 2 analyzes the key frame image file, and analyzes that the current moment is the daytime mode;

Set day mode parameters, adjust appropriate brightness, contrast, noise removal parameters;

Image recognition module 2 analyzes the key frame image file as shown in Figure 16, and the traffic sign information, road traffic marking information and traffic signal light information are as shown in Figure 17, and store the information;

Judging whether the road traffic marking information is an intersection lane marking, no;

Determine whether there is traffic signal light information, no;

Real-scene navigation management module 6 integrates real-scene image recognition information, latitude and longitude information, path information and road information, and displays the real scene on the display interface to provide driving tips for drivers;

Frame image processing steps shown in Figure 18

Image recognition module 2 analyzes the key frame image file as shown in Figure 18, and the traffic sign information, road traffic marking information and traffic signal light information are as shown in Figure 19, and store the information;

Judging whether the road traffic marking information is an intersection lane marking, yes;

Judging whether there is traffic sign information within the time threshold T1, T1=100 seconds, yes;

Judging the traffic sign information as shown in Figure 17 identified from Figure 16, from left to right is: turn around or turn left, turn left, turn right, turn right; map it to the information identified in Figure 18 as The road traffic marking information shown in Figure 19, from left to right, is: U-turn or left turn, left turn, right turn, right turn; change;

Judging whether the traffic signal light information is identified from Figure 18, as shown in Figure 19, no;

Real-scene navigation management module 6 integrates real-scene image recognition information, latitude and longitude information, route information and road information, and displays the real scene on the display interface to provide drivers with lane selection prompts;

Embodiment 4: Car real scene navigation night scene

Carry out the following steps sequentially from the driving route

Frame image processing steps shown in Figure 20:

GPS positioning module 3 extracts the latitude and longitude information of automobile driving;

The map management module 5 provides relevant road information of the driving area of the vehicle according to the latitude and longitude information of the driving of the vehicle;

The path planning module 4 provides the prompt for the automobile to go straight at the previous intersection according to the latitude and longitude information of the automobile, the relevant road information and the path information required to reach the destination;

Video analysis module 1 collects the real-time video information of automobile operation, analyzes and extracts video key frames, and is stored as a key frame image file as shown in Figure 20;

Image recognition module 2 analyzes the key frame image file, and analyzes that the current moment is night mode;

Set the night mode parameters, adjust the appropriate brightness, contrast, noise removal parameters;

Image recognition module 2 analyzes the key frame image file as shown in Figure 20, traffic sign information, road traffic marking information and traffic signal light information as shown in Figure 21, and store information;

Judging whether the road traffic marking information is an intersection lane marking, no;

Determine whether there is traffic signal light information, no;

Real-scene navigation management module 6 integrates real-scene image recognition information, latitude and longitude information, path information and road information, and displays the real scene on the display interface to provide driving tips for drivers;

Frame image processing steps shown in Figure 22

Image recognition module 2 analyzes the key frame image file as shown in Figure 22, traffic sign board information, road traffic marking information and traffic signal light information as shown in Figure 7, and store information;

Judging whether the road traffic marking information is an intersection lane marking, yes;

Judging whether there is traffic sign information within the time threshold T1, T1=100 seconds, yes;

Judging the traffic sign information shown in Figure 21 identified from Figure 20, from left to right: go straight 1, go straight 2, go straight and turn right; map it to the traffic sign information identified from Figure 22 as shown in Figure 23 The road traffic marking information, from left to right, is: go straight 1, go straight 2, go straight and turn right; according to the prompt of going straight at the intersection provided by the path planning module 4, it is determined to choose the lane to go straight and turn right;

Judging whether the traffic signal light information is identified from Figure 22, as shown in Figure 23, no;

The real-scene navigation management module 6 integrates real-scene image recognition information, latitude and longitude information, path information and road information, and displays the real scene on the display interface to provide the driver with a lane selection prompt.

In the description of this specification, references to the terms "one embodiment," "some embodiments," "exemplary embodiments," "example," "specific examples," or "some examples" are intended to mean that the implementation A specific feature, structure, material, or characteristic described by an embodiment or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

Although the embodiments of the present invention have been shown and described, those skilled in the art can understand that various changes, modifications, substitutions and modifications can be made to these embodiments without departing from the principle and spirit of the present invention. The scope of the invention is defined by the claims and their equivalents.

Claims (4)

1. the automobile real scene navigation apparatus of an image recognition, it is characterised in that include analysis module, image recognition mould Block, GPS locating module, path planning module, management map module, real scene navigation management module；

Described analysis module is connected with picture recognition module；

Described real scene navigation management module is advised with picture recognition module, GPS locating module, management map module, path respectively Draw module to connect；

The outdoor scene video information that described analysis module gathers for analytical equipment, extracts key frame of video, is stored as closing Key two field picture file；

Described picture recognition module is for identifying the traffic marking board information in key frame images file, road traffic marking Information, traffic light information；

Described GPS locating module is for providing the latitude and longitude information of running car；

Described path planning module arrives at the routing information of needs for providing；

Described management map module is for providing the related roads information in running car region；

Described real scene navigation management module is for integrated real scene image identification information, latitude and longitude information, routing information and road Information, and showing that interface shows in real time.

The automobile real scene navigation apparatus of a kind of image recognition the most according to claim 1, it is characterised in that described GPS L1 carrier wave that locating module uses frequency to be 1575.42MHz and L2 carrier wave that frequency is 1227.60MHz.

3. the automobile real scene navigation method of an image recognition, it is characterised in that comprise the following steps:

Step S1, GPS locating module extracts the latitude and longitude information of running car；

Step S2, management map module is for providing the related roads information in running car region；

Step S3, path planning module is according to latitude and longitude information, the related roads information of running car and arrives at needs Routing information, it is provided that automobile is to turn around, turn left, keep straight on or the track selection mode such as right-hand bend at previous crossing；

Step S4, analysis module gathers the real-time video information of automobilism, analyzes and extract key frame of video, storage For key frame images file；

Step S5, picture recognition module analysis of key two field picture file, analyze the current moment, if day mode, turn Enter step S6, be otherwise night mode, proceed to step S7；

Step S6, arranges day mode parameter, proceeds to step S8；

Step S7, arranges night mode parameter, proceeds to step S8；

Step S8, picture recognition module analysis of key two field picture file, traffic marking board information, road traffic marking information and friendship Ventilating signal lamp information, and store information；

Step S9, it is judged that whether road traffic marking information is that road junction roadway indicates, if it is, proceed to step S10, if it does not, Proceed to step 12；

Step S10, it is judged that within time threshold T1, if having traffic marking board information, if it is, proceed to step S11, as The most no, proceed to step 12；

Step S11, it is judged that turning around, turn left, keep straight on or turning right in traffic marking board information, maps that to road and hands over Logical graticule information turns around, turns left, keeps straight on or turns right, and the choosing lane mode provided according to path planning module determines Choosing lane information；

Step S12, it may be judged whether have traffic light information, if it has, proceed to step S13, if it does not, proceed to step 16；

Step S13, it is judged that within time threshold T2, if having choosing lane information, if it is, proceed to step S14, if No, proceed to step 15；

Step S14, judges the track clearance sign situation in traffic light information: travel or stay according to choosing lane information Car, proceeds to step S16；

Step S15, directly judges the track clearance sign situation in traffic light information: travel or parking；

Step S16, real scene navigation management module integrated real scene image identification information, latitude and longitude information, routing information and road letter Breath, and showing that interface outdoor scene shows, it is provided that to driver's choosing lane, travel or parking prompting.

The automobile real scene navigation method of a kind of image recognition the most according to claim 3, it is characterised in that described step S8 Method particularly includes: eliminate in key frame images file initially with the medium filtering in image recognition and Morphological scale-space method Noise, then use the method identification zone boundary of the structure Polygonal Boundary in image recognition and identify objectives, Finally using and calculate objectives invariant moment features and traffic marking board, road traffic marking, the not bending moment of traffic light is special The distance levied, it is judged that the actual information of objectives.