CN114199265A - Auxiliary navigation system and navigation method based on target detection algorithm - Google Patents
Auxiliary navigation system and navigation method based on target detection algorithm Download PDFInfo
- Publication number
- CN114199265A CN114199265A CN202111383578.6A CN202111383578A CN114199265A CN 114199265 A CN114199265 A CN 114199265A CN 202111383578 A CN202111383578 A CN 202111383578A CN 114199265 A CN114199265 A CN 114199265A
- Authority
- CN
- China
- Prior art keywords
- module
- target detection
- navigation
- navigation system
- detection system
- 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.)
- Pending
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 80
- 238000000034 method Methods 0.000 title claims abstract description 21
- 238000012544 monitoring process Methods 0.000 claims abstract description 29
- 238000012549 training Methods 0.000 claims abstract description 23
- 238000012545 processing Methods 0.000 claims description 13
- 238000004891 communication Methods 0.000 claims description 10
- 238000012360 testing method Methods 0.000 claims description 6
- 238000012805 post-processing Methods 0.000 claims description 5
- 230000005540 biological transmission Effects 0.000 claims description 2
- 238000005457 optimization Methods 0.000 claims description 2
- 238000004088 simulation Methods 0.000 claims description 2
- 238000013461 design Methods 0.000 claims 1
- 206010039203 Road traffic accident Diseases 0.000 description 8
- 238000005516 engineering process Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 230000004888 barrier function Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000002427 irreversible effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
- G01C21/34—Route searching; Route guidance
- G01C21/3453—Special cost functions, i.e. other than distance or default speed limit of road segments
- G01C21/3492—Special cost functions, i.e. other than distance or default speed limit of road segments employing speed data or traffic data, e.g. real-time or historical
Abstract
The invention discloses an auxiliary navigation system and a navigation method based on a target detection algorithm, which comprises an intelligent automobile data recorder system and a navigation system, wherein a video monitoring module and a target detection system module are embedded in the intelligent automobile data recorder system, and a target object is identified for road condition videos shot in real time in the intelligent automobile data recorder system through the target detection system module; the navigation system comprises a basic function module, a monitoring module, a voice broadcasting module and a display module; firstly, starting an intelligent automobile data recorder, operating a target detection system, and loading a pre-training weight; starting a navigation system and monitoring a message sent by a target detection system in real time; the target detection system monitors road condition information in real time, and the navigation system prompts a driver through an interface and voice after receiving the early warning message sent by the target detection system.
Description
Technical Field
The invention relates to a navigation system and a navigation method thereof, in particular to an auxiliary navigation system and a navigation method based on a lightweight YOLO target detection algorithm, and belongs to the technical field of navigation.
Background
With the continuous development of city motorization in recent years, the holding amount of motor vehicles always shows a rising trend, and the intensity and range of urban traffic travel activities are obviously improved and become an irreversible trend. Under the condition that the intensity of the demand for travel is larger and larger, urban traffic supply can not follow the demand for traffic gradually, the unbalanced contradiction between the urban traffic supply and the demand for traffic is more and more obvious, a plurality of traffic problems are caused, and the occurrence of traffic accidents is stimulated to a great extent. The traffic accidents which occur every year bring great harm to people all over the world, how to carry out safe driving and safe traveling is the focus problem discussed by people, the safe driving assisting technology is optimized along with the transportation of the safe driving assisting technology and starts from the perspective of computer vision, the traffic accidents which can be avoided during the driving of the automobile are analyzed through research, the occurrence rate of the traffic accidents is reduced, certain guarantee is provided for the safe driving of the automobile, and the safe driving assisting technology is innovated and upgraded.
Disclosure of Invention
In view of the above problems in the prior art, the present invention provides an auxiliary navigation system and a navigation method based on a target detection algorithm, which can realize accurate navigation and avoid traffic accidents.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows: an auxiliary navigation system based on a target detection algorithm comprises an intelligent automobile data recorder system and a navigation system which are in communication connection, wherein a video monitoring module and a target detection system module are embedded in the intelligent automobile data recorder system, the video monitoring module comprises a video monitoring function, the target detection module comprises a target object monitoring function, and target object identification is carried out on road condition videos shot in real time in the intelligent automobile data recorder system through the target detection system module; navigation system includes basic function module, monitors the module, voice broadcast module and display module, and basic function module contains the location of navigation itself and basic function such as navigation, and the module of monitoring contains real-time monitoring and receives the function of vehicle event data recorder transmission early warning information, and the voice broadcast module contains the voice broadcast function, and display module contains and shows current traffic signal lamp, lane turn to arrow point and pedestrian's obstacle warning at the navigation interface.
Furthermore, the target detection system module consists of a configuration file module, a communication module, a target detection system main program and a core processing module;
the target detection system main program uses an embedded program developed based on a lightweight YOLO detection algorithm model, and loads weights pre-trained on a server to realize real-time monitoring of road condition information;
the configuration file module is mainly used for storing configuration files such as a predefined prediction box, weight, super parameters and the like which are needed to be used by a target detection system program;
the communication module is mainly responsible for communicating with a mobile phone through Bluetooth or WIFI and transmitting early warning information;
the core processing module is mainly a lightweight YOLO network algorithm model and comprises functions of data loading, model detection, NMS post-processing and the like.
Furthermore, the configuration file module can be upgraded and replaced with new optimization weights, and the communication module supports Bluetooth and WIFI communication.
A navigation method of an assisted navigation system based on an object detection algorithm according to claim 1, comprising the steps of:
step one, starting an intelligent automobile data recorder, operating a target detection system, and loading a pre-training weight;
starting a navigation system, and monitoring a message sent by a target detection system in real time;
and step three, the target detection system monitors road condition information in real time and sends early warning information to special conditions in the front road, wherein the special conditions at least comprise that an obstacle exists in the front road, the red light state of a traffic signal lamp and the state of a lane is inconsistent with the state of the traffic signal lamp.
And step four, after receiving the early warning message sent by the target detection system, the navigation system prompts the driver through an interface and voice.
Further, the target detection algorithm is a lightweight YOLO target detection algorithm, and is an algorithm designed and implemented based on a YOLO-V4-Tiny or a YOLO Nano lightweight target detection network.
Further, in the first step, the pre-training weight is an optimal weight selected after the server simulation training test, the classification of all data sets is adjusted during training, and all data sets are converted into the VOC format for training.
Further, when pre-trained on the server, the data sets used include the BDD100K data set, the TT100K data set, and the live shot data set; the BDD100K data set comprises a large number of preliminarily processed driving video materials, at least comprising detailed labels, road surface object recognition, lane marks and feasible space information; the pictures in the TT100K data set are 8192 x 4096 pixel high-resolution panoramas which are shot by a vehicle-mounted camera and formed by post-processing; the real shooting data set is picture data formed by processing videos shot by a real vehicle-mounted automobile data recorder;
a plurality of BDD100K picture data, TT100K picture data and real shooting data pictures are taken as training and testing for use, a lightweight target monitoring system is simulated on a server for pre-training, and the weight with better training is exported and uploaded to a target detection system to serve as a configuration file.
Further, in the third step, the target detection system monitors the road condition information in real time, and performs target identification on the traffic signal lamps, the lanes and the vehicles;
in the normal driving process, a target detection system in the automobile data recorder is used for identifying whether objects which comprise pedestrians and electric vehicles and block motor vehicle passing exist in a road in front of a lane, and if the objects exist, a warning message is sent to a navigation system after the objects are confirmed by a core processing module;
and identifying a traffic signal lamp and a lane, judging whether the vehicle can pass through by the core processing module according to the actual situations of the current lane and the traffic signal lamp, and sending a message and information of the current lane to the navigation system.
Compared with the prior art, the invention has the beneficial effects that:
1. the system detects whether obstacles such as pedestrians exist in front of the road in the driving process in real time, informs drivers of slowing down or avoiding steering in advance, and prevents traffic accidents.
2. The system detects the state of the traffic light in the driving process in real time and informs drivers in advance whether the drivers can fast run at the same time or not so as to prevent traffic accidents caused by running the red light. 3. The system detects the lane in the driving process in real time and simultaneously assists the state of the traffic signal lamp to remind a driver of the lane, so that traffic accidents caused by the fact that the driver does not drive according to the lane are avoided.
Drawings
FIG. 1 is a block diagram of an object detection system of the present invention.
FIG. 2 is a block diagram of a navigation system according to the present invention.
FIG. 3 is a flow chart of an assisted navigation system according to the present invention.
Detailed Description
In order to better understand the technical content of the present invention, specific embodiments are described below with reference to the accompanying drawings.
The invention is further described below with reference to the accompanying drawings. The following examples are only for illustrating the technical solutions of the present invention more clearly, and the protection scope of the present invention is not limited thereby.
An auxiliary navigation system based on a lightweight YOLO target detection algorithm comprises the following steps:
step one, starting an intelligent automobile data recorder, operating a target detection system, and loading a pre-training weight;
starting a navigation system, and monitoring a message sent by a target detection system in real time;
step three, the target detection system monitors road condition information in real time and sends early warning information for special conditions;
and step four, after receiving the early warning message sent by the target detection system, the navigation system prompts the driver through an interface and voice.
The method comprises the following steps that firstly, an intelligent automobile data recorder is started, a target detection system is operated, and pre-training weights are loaded;
embedding a target detection system in the intelligent automobile data recorder, and identifying a target object by the target detection system on the road condition video shot in real time in the intelligent automobile data recorder, wherein the target detection system module is shown as figure 1.
The main program of the target detection system uses an embedded program developed based on a lightweight YOLO detection algorithm model, and loads weights which are pre-trained on a server, so that real-time monitoring of road condition information is realized.
The lightweight YOLO detection algorithm model can be adapted and improved using a YOLO-V4-Tiny or a YOLO Nano lightweight target detection network.
The data sets used when pre-trained on the server include the BDD100K data set, the TT100K data set, and the live beat data set. The BDD100K data set contains a large amount of primarily processed driving video material, including detailed labels (Annotations), Road Object Detection (Road Object Detection), Lane markers (Lane markers), feasible space (driving Areas), and so on, which can be said to include most of the data in the driving vision aspect. The picture source and the Tencent street view picture in the TT100K data set are 8192 x 4096 pixel high-resolution panoramic pictures which are shot by a vehicle-mounted camera and are formed through post-processing. The real shooting data set is picture data formed by processing videos shot by the real vehicle-mounted automobile data recorder.
3 ten thousand BDD100K picture data, 1 ten thousand TT100K picture data and 1 ten thousand live shooting data pictures are taken as training and testing. 5000 real-shot pictures are extracted as a test set, the rest are training sets, and the categories except traffic lights, lanes, vehicles and the like are reset to be one category. The method comprises the steps of pre-training by simulating a lightweight target monitoring system on a server, exporting the weight with better training and uploading the weight to the target monitoring system as a configuration file.
Starting a navigation system, and monitoring a message sent by a target detection system in real time;
and starting the navigation system, and running a self-defined Listener module to monitor the message in real time, wherein the self-defined Listener module can be developed and designed based on the API opened by the existing mobile phone navigation system.
During actual driving navigation, the SDK of the navigation system transmits navigation information and data in real time through the listeners such as AMapNaviListener, parallelrowdelistener, and AimlessModeListener. The function interfaces of amapnavi.getinstance () and mamapnavaj.addamapnnavilisterner () can be called in the navigation system to register a new data monitoring module, and the module is specially used for receiving the real-time road condition information transmitted by the target detection system.
And acquiring a NaviInfo object through an onanaviInfoUpdate callback interface in the AMapaNaviListener, wherein the NaviInfo object is newly added with self-defined information such as traffic lights, lanes, non-motor vehicles and the like.
And newly adding custom voice broadcast, transparently transmitting the broadcast text content through the AMapNaviListener class callback, and converting the content into sound information by selecting a voice synthesis SDK of a third party to finish voice broadcast.
The target detection system monitors road condition information in real time and sends early warning information for special conditions;
the target monitoring system monitors road condition information in real time and identifies targets of traffic signal lamps, lanes and vehicles.
In the normal driving process, whether non-motor vehicle objects, such as pedestrians, electric vehicles and other obstructing objects, exist in a road in front of a lane, and for finding the special conditions, the core processing module sends a warning message to the navigation system after confirming the special conditions.
And identifying a traffic signal lamp and a lane, judging whether the vehicle can pass through by the core processing module according to the actual situation of the current lane and the traffic signal lamp, and sending ok and forbidden messages and current lane information (left-turn lane, straight lane and right-turn lane) to a navigation system.
And fourthly, after receiving the early warning message sent by the target detection system, the navigation system prompts the driver through an interface and voice.
After the navigation system receives various early warning messages sent by a target monitoring system through a self-defined monitoring module, the current traffic signal lamp state and the current lane are displayed on a navigation interface through a self-defined NaviInfo module, and meanwhile voice information is broadcasted to remind a driver of a traffic light in front, a lane, whether pedestrians exist, whether barriers exist and the like. .
Although the present invention has been described with reference to the preferred embodiments, it is not intended to be limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the invention. Therefore, the protection scope of the present invention should be determined by the appended claims.
Claims (8)
1. An aided navigation system based on a target detection algorithm, characterized in that: the intelligent automobile data recorder system comprises an intelligent automobile data recorder system and a navigation system which are in communication connection, wherein a video monitoring module and a target detection system module are embedded in the intelligent automobile data recorder system, the video monitoring module comprises a video monitoring function, the target detection module comprises a target object detection function, and the target object detection system module is used for identifying a road condition video shot in real time in the intelligent automobile data recorder system; navigation system includes basic function module, monitors the module, voice broadcast module and display module, and basic function module contains the location of navigation itself and basic function such as navigation, and the module of monitoring contains real-time monitoring and receives the function of vehicle event data recorder transmission early warning information, and the voice broadcast module contains the voice broadcast function, and display module contains and shows current traffic signal lamp, lane turn to arrow point and pedestrian's obstacle warning at the navigation interface.
2. The object detection algorithm-based assisted navigation system of claim 1, wherein: the target detection system module consists of a configuration file module, a communication module, a target detection system main program and a core processing module;
the target detection system main program uses an embedded program developed based on a lightweight YOLO detection algorithm model, and loads weights pre-trained on a server to realize real-time monitoring of road condition information;
the configuration file module is mainly used for storing configuration files such as a predefined prediction box, weight, super parameters and the like which are needed to be used by a target detection system program;
the communication module is mainly responsible for communicating with a mobile phone through Bluetooth or WIFI and transmitting early warning information;
the core processing module is mainly a lightweight YOLO network algorithm model and comprises functions of data loading, model detection, NMS post-processing and the like.
3. The object detection algorithm-based assisted navigation system of claim 2, wherein: the configuration file module can be upgraded and replaced with new optimization weights, and the communication module supports Bluetooth and WIFI communication.
4. The navigation method of the aided navigation system based on the object detection algorithm of claim 1, characterized in that: the method comprises the following steps:
step one, starting an intelligent automobile data recorder, operating a target detection system, and loading a pre-training weight;
starting a navigation system, and monitoring a message sent by a target detection system in real time;
step three, the target detection system monitors road condition information in real time and sends early warning information to special conditions in the front road, wherein the special conditions at least comprise that an obstacle exists in the front road, the red light state of a traffic signal lamp and the state of a lane is inconsistent with the state of the traffic signal lamp;
and step four, after receiving the early warning message sent by the target detection system, the navigation system prompts the driver through an interface and voice.
5. The navigation method of the aided navigation system based on the target detection algorithm of claim 4, wherein the target detection algorithm is a lightweight YOLO target detection algorithm, and is an algorithm implemented based on a YOLO-V4-Tiny or a YOLO Nano lightweight target detection network design.
6. The navigation method of the aided navigation system based on the object detection algorithm as claimed in claim 4, wherein: in the first step, the pre-training weight is the optimal weight selected after the simulation training test of the server, the classification of all data sets is adjusted during training, and all data sets are converted into the VOC format for training.
7. The navigation method of the aided navigation system based on the object detection algorithm of claim 5, characterized in that: the data sets used when pre-trained on the server include the BDD100K data set, the TT100K data set, and the live shot data set; the BDD100K data set comprises a large number of preliminarily processed driving video materials, at least comprising detailed labels, road surface object recognition, lane marks and feasible space information; the pictures in the TT100K data set are 8192 x 4096 pixel high-resolution panoramas which are shot by a vehicle-mounted camera and formed by post-processing; the real shooting data set is picture data formed by processing videos shot by a real vehicle-mounted automobile data recorder;
a plurality of BDD100K picture data, TT100K picture data and real shooting data pictures are taken as training and testing for use, a lightweight target monitoring system is simulated on a server for pre-training, and the weight with better training is exported and uploaded to a target detection system to serve as a configuration file.
8. The navigation method of the aided navigation system based on the object detection algorithm as claimed in claim 4, wherein: in the third step, the target detection system monitors the road condition information in real time and performs target identification on traffic signal lamps, lanes and vehicles;
in the normal driving process, a target detection system in the automobile data recorder is used for identifying whether objects which comprise pedestrians and electric vehicles and block motor vehicle passing exist in a road in front of a lane, and if the objects exist, a warning message is sent to a navigation system after the objects are confirmed by a core processing module;
and identifying a traffic signal lamp and a lane, judging whether the vehicle can pass through by the core processing module according to the actual situations of the current lane and the traffic signal lamp, and sending a message and information of the current lane to the navigation system.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111383578.6A CN114199265A (en) | 2021-11-22 | 2021-11-22 | Auxiliary navigation system and navigation method based on target detection algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111383578.6A CN114199265A (en) | 2021-11-22 | 2021-11-22 | Auxiliary navigation system and navigation method based on target detection algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114199265A true CN114199265A (en) | 2022-03-18 |
Family
ID=80648274
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111383578.6A Pending CN114199265A (en) | 2021-11-22 | 2021-11-22 | Auxiliary navigation system and navigation method based on target detection algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114199265A (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110276988A (en) * | 2019-06-26 | 2019-09-24 | 重庆邮电大学 | A kind of DAS (Driver Assistant System) based on collision warning algorithm |
| CN110979350A (en) * | 2019-12-18 | 2020-04-10 | 奇瑞汽车股份有限公司 | AR navigation based on automobile data recorder |
-
2021
- 2021-11-22 CN CN202111383578.6A patent/CN114199265A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110276988A (en) * | 2019-06-26 | 2019-09-24 | 重庆邮电大学 | A kind of DAS (Driver Assistant System) based on collision warning algorithm |
| CN110979350A (en) * | 2019-12-18 | 2020-04-10 | 奇瑞汽车股份有限公司 | AR navigation based on automobile data recorder |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20230311749A1 (en) | Communication between autonomous vehicle and external observers | |
| CN108417087B (en) | Vehicle safe passing system and method | |
| GB2536549A (en) | Virtual autonomous response testbed | |
| CN105551110A (en) | Traveling vehicle data recording method, device and system | |
| GB2536771A (en) | Autonomous driving refined in virtual environments | |
| US6813554B1 (en) | Method and apparatus for adding commercial value to traffic control systems | |
| CN1833934A (en) | Safety monitoring system for running car and monitoring method | |
| CN104266654A (en) | Vehicle real scene navigation system and method | |
| EP4131200A1 (en) | Method and device for providing road congestion reason | |
| CN201268230Y (en) | Speed limit sign information vehicle auxiliary drive system based on computer vision | |
| CN113808418B (en) | Road condition information display system, method, vehicle, computer device and storage medium | |
| CN105809095B (en) | Determination of traffic intersection passage state | |
| CN110838231A (en) | Pedestrian crossing intelligent detection system and method | |
| CN114394100B (en) | Unmanned patrol car control system and unmanned car | |
| CN110610153A (en) | Lane recognition method and system for automatic driving | |
| CN113837127A (en) | Map and V2V data fusion model, method, system and medium | |
| CN111985373A (en) | Safety early warning method and device based on traffic intersection identification and electronic equipment | |
| DE102019110127A1 (en) | Method and device for providing a lighting function | |
| CN112907979B (en) | System and method for tracking illegal running track of motor vehicle based on multiple cameras | |
| CN114199265A (en) | Auxiliary navigation system and navigation method based on target detection algorithm | |
| CN111917873A (en) | Accident processing system, method, device, terminal and storage medium | |
| CN114842455B (en) | Obstacle detection method, device, equipment, medium, chip and vehicle | |
| JP5082298B2 (en) | Autonomous mobile device | |
| CN112435475B (en) | Traffic state detection method, device, equipment and storage medium | |
| US20210092573A1 (en) | Environment machine interface system |
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 |