786 results about "Traffic scene" patented technology

The invention belongs to an unmanned vehicle test and verification platform and a test method thereof in the field of unmanned vehicles. The platform comprises a driving simulation system, an experimental field, a network system, an upper level management center and an unmanned vehicle. The driving simulation system constructs a driving environment of a virtual vehicle according to the information collected in a natural scene; the traffic scene of the actual site of the experimental field coincides with the scene modeled by the driving simulation system; and the upper level management center is used for establishing a simulator driving environment, controlling the driving simulation system and processing data; the unmanned vehicle is a test vehicle and is automatically driven in the experimental field; and the driving information of the unmanned vehicle is transmitted to the upper level management center through a network system and then transmitted to the driving simulator of the driving simulation system. According to the invention, function verification and performance evaluation of the unmanned vehicle can be studied, and meanwhile, influences by the unmanned vehicle on an actual traffic flow can be studied and evaluated through the platform.

The invention discloses an intelligent test system and an intelligent test method for an autonomous vehicle. The intelligent test system comprises a decision-making layer, a display layer, a control layer and a vehicle side layer, wherein the decision-making layer decides a test scheme to use according to test requirements; the display layer displays a high-precision map of a test field by means of a plurality of display screens, monitors the whole process of a test scene, and displays driving paths of a main vehicle and test vehicles; the control layer realizes cooperative control of a plurality of the test vehicles in a driving simulation mode during the test process, so as to simulate a real and complicated traffic scene, carries out various types of test information acquisition simultaneously, and transmits acquired data to a cloud server to realize the cooperative control of the plurality of the test vehicles; and the vehicle side layer comprises the vehicles and equipment directly participating in the test in the test field, and comprises the main vehicle, the test vehicles and a remote driving control unit. The intelligent test system replaces various types of manually operated test equipment, realizes the complicated traffic scene through multi-vehicle cooperative control, avoids the safety problem of manual driving tests, and can carry out high-precision repeated test.

The invention discloses a vehicle-road collaborative vehicle-mounted terminal application system based on LTE-V. In the system, the vehicle maintains wireless short-wave communication with other vehicles and roadside equipment in real time by utilizing the characteristics of low delay and high coverage range of LTE-V communication, and interacts information with a vehicle-road collaborative background server through a 4G network. The method comprises the following steps: receiving real-time state information of other vehicles; The method comprises the following steps of: sensing surrounding traffic conditions by traffic participant real-time information sent by roadside equipment, and judging and analyzing a map road relationship, a road distance and collision time between a vehicle and other traffic participants, traffic signs or events according to the vehicle state information and local lane track point map information; According to different traffic scenes judged by the system, a driver is reminded of different traffic events and early warning information through visualization, so that the driving safety is enhanced, and meanwhile, the traffic efficiency is improved.

The invention, which relates to the technical field of an intelligent vehicle automatic driving and traffic simulation, relates to an artificial intelligent training platform for an intelligent networking vehicle plan decision-making module and aims at improving the intelligent level of the intelligent vehicle plan decision-making module based on enriched and vivid traffic scenes. The artificial intelligent training platform comprises a simulation environment layer, a data transmission layer, and a plan decision-making layer. The simulation environment layer is used for generating a true traffic scene based on a traffic simulation module and simulating sensing and reaction situations to the environment by an intelligent vehicle, thereby realizing multi-scene loading. The plan decision-making layer outputs a decision-making behavior of the intelligent vehicle by using environment sensing information as an input based on a deep reinforcement learning algorithm, thereby realizing training optimization of network parameters. And the data transmission layer connects the traffic environment module with a deep reinforcement learning frame based on a TCP / IP protocol to realize transmission of sensing information and vehicle control information between the simulated environment layer and the plan decision-making layer.

A vehicle lighting system for illuminating ahead of a vehicle is provided with a plurality of light sources each adapted to illuminate a different illumination area to thereby form a predetermined light distribution pattern as a whole and a control unit for determining a traffic scene in which the vehicle is running and independently changing at least either an orientations or the light quantities of the plurality of light sources so as to form a target light distribution pattern which has been set in advance for the traffic scene.

The invention discloses a traffic scene multi-target classification method, to be specific, discloses a traffic scene classification method based on a multi-scale convolution neural network. The traffic scene classification method is characterized in that recessive characteristics based on the multi-scale convolution neural network are extracted; and an optimal covering segmentation tree is acquired. During the realizing of the traffic scene classification, the multi-scale convolution neural network is adopted, and the excellent recessive characteristics having the invariance property are effectively extracted from an original image in different scales, and by comparing with the single-scale convolution neural network, the acquisition of the abundant and effective characteristic information of the image is realized. The effective information extracted by the convolution neural network is combined with the original segmentation tree of the image to form an optimal purity price tree, and the covering having the optimal purity is carried out, and therefore a clearer target contour is acquired, and the classification accuracy is increased. The RGB-D is used as the convolution neural network input, and by comparing with the conventional RGB convolution neural network input, the training characteristic is additionally provided with the depth information, and the classification of the input image is more accurate.

The invention provides an intelligent connected vehicle in-the-loop simulation and test and verification system and method. Simulation of the intelligent trolley operation scene is performed mainly byIPG CarMaker, CarSIM and other tools. The system is composed of the functions of Simulink data simulation calculation, transmission of the simulation scene parameters to the intelligent trolley for simulation of the real scene, intelligent trolley image acquisition and display and test analysis. The intelligent trolley side is mainly responsible for environmental perception and control algorithmverification of the intelligent connected vehicle and is composed of a trolley and a control module, a driving module, a communication module and a sensor module to act as the carrier of loading the test algorithm software and the sensor and the communication module required to be tested. According to the system, the cost and the dangerousness of the intelligent connected vehicle test can be reduced, onsite verification of the intelligent vehicle system can be performed in the real site, the research work of modeling, simulation and control and other aspects for vehicle-road coordination and other application in different simulation traffic scenes and thus a lot of data support can be provided for the research of the intelligent connected vehicle.

The invention provides a method for detecting and tracking vehicles through processing a video image of a traffic scene, comprising the following steps of: collecting images of the traffic scene; deblocking a traffic image according to 8 multiplied by 8 pixels; initiating and continuously updating a background image; setting a virtual detection line with a width of 8 pixels vertical to a direction of a traffic lane at the entrance of the traffic lane; detecting the vehicles through calculating the variation of a gray scale of blocks on the virtual detection line relative to the background image; labeling the blocks belonging to the regions of the vehicles by using a sequence that the vehicles reach the detection line; receiving initial positions of labels of the vehicles; calculating motion vectors of the vehicles by combining a three-step method and a full search method and constantly updating the positions of the labels of the vehicles by using the motion vectors so as to implement the tracking of vehicles. The method also performs detection and splitting against an occlusion occurrence between vehicles in a complicated traffic scene, thereby having the advantages of high accuracy of detection and tracking and good real-time performance.

The invention provides a real-time road traffic signal coordination optimization control method and a control system thereof and solves problems that real-time optimization for control characteristic parameters such as a signal period and a green signal ratio according road traffic flow change can be not realized in a traffic signal control system mainly in a multi-period fixed signal timing mode; traffic pass efficiency and pass capability are not high; traffic congestion or even traffic jam can not be effectively ameliorated. According to an active optimization real-time driving system based on traffic data, macroscopic control strategies and microscopic tactical control strategies are made according to different dynamic and static traffic scenes, main-line coordination optimization control and real-time dynamic crossing optimization control are actively carried out, the control characteristic parameters such as the signal period and the green signal ratio are dynamically adjusted and generated, dynamic benefit evaluation is controlled through controlling operation characteristic indexes, and real-time coordination control is carried out through gradual operation monitoring and repeated optimization searching.

The invention discloses a method and a device for deeply analyzing a traffic scene. The method comprises the steps of using a data set of original images in a plurality of traffic scene databases and road areas corresponding to the original images as training samples; through a Laplacian Pyramid transform mode, respectively resizing the original images into different scales, and inputting into neural networks respectively corresponding to the different scales, wherein each neural network is composed of a convolutional neural network part and a deconvolutional neural network part; outputting a one-dimensional array having the same pixel with the original image through a fully connected layer connected with each neural network, and restoring into a result image having the same size with the original image, wherein different types of roads are marked in the result image; processing the result image by a preset standard to restore the segmentation result of the roads; and inputting the image to be detected into the successfully trained neural network, and thereby obtaining the result image in which the road segmentation is completed and corresponding to the image to be detected. According to the method provided by the invention, the accuracy of analyzing the traffic scene can be improved.

The invention discloses a driving scene classification method based on a convolution neural network, and the method comprises the following steps: collecting a road environment video image; carrying out the classification of a traffic scene, and building a traffic scene recognition database; extracting sample images of different driving scenes from the traffic scene recognition database, carryingout the feature extraction and multiple convolution training of the sample images through a deep convolution neural network, carrying out the rasterization of pixels, connecting the pixels to form a vector, inputting the vector into a conventional neural network, obtaining convolution neural network output, and achieving the deep learning of different driving scenes; carrying out the parameter optimization of a network structure of the built convolution neural network, obtaining a trained convolution neural network classifier, carrying out the adjustment of a traffic scene recognition model, and selecting an optimal mode as the standard of the traffic scene recognition model; carrying out the real-time collection of the image of a to-be-detected traffic scene, and inputting the image intothe traffic scene recognition model for the recognition of a road environment scene.

The invention provides an image information and radar information fusion method and system for a traffic scene, and the method comprises the following steps: preprocessing the image information, obtained by a camera, of the front of a vehicle; extracting the feature information in the image information, and comparing the feature information with the pre-stored traffic scene information; classifying the current traffic scene according to a comparison judgment result, performing a corresponding fusion algorithm according to a preset fusion method of image information and radar information whichis adapted to the current traffic scene category, and outputting the result of the fusion algorithm. The method judges a scene according to the collected image information, switches between differentfusion algorithms, effectively utilizes resources, and improves scene adaptability. The method fully utilizes redundancy and complementary characteristics between different sensor data to improve system robustness and reliability. The method adopts a deep learning algorithm in image information processing, which has higher real-time performance and is more accurate in target recognition.

The invention provides an automobile automatic driving test case generation method based on scenes and tasks, and the method comprises the following steps: 1), selecting typical traffic behaviors in acomplex and multi-element road traffic scene, and constructing a test task library; 2) defining terms of a traffic scene and related elements thereof, describing concepts, attributes and hierarchicalrelationships of scene elements, and constructing an ontology model of the road traffic scene; 3) generating a traffic scene based on the traffic scene model and the traffic behavior task library, the traffic scene being generally composed of a plurality of traffic participants, a specific time-space road environment and a traffic scene, and generally including a plurality of driving tasks; and 4) further describing a traffic scene in detail by defining a parameter range and correlation of variables in the state space, and selecting representative discrete values for combination by applying boundary value analysis, equivalence class division and other methods in software testing to generate a plurality of test cases.

This invention discloses a mixed traffic flow parameters detection method based on video. The method is the application of computer program on the basis of existing hardware to realize the video detection method, including the following steps: installing hardware equipment, acquiring traffic video frame, initialization, background model, shadow detection, object recognition, motion tracking, accessing the mixed traffic flow parameters and processing the received parameters getting the accurate control parameters to manage and control traffic through hardware device. Among them, the initialization step includes detection area selection, traffic scene location mapping and background initialization. The invention can access the micro-traffic behavior according to the action tracking, such as: accelerating, changing lane, stop, turning and reverse.

The invention discloses a vehicle-mounted video target detection method based on deep learning, and the method comprises the steps: employing an improved Faster R-CNN algorithm to realize target detection in a complex traffic environment, to provide a driving safety auxiliary function. An existing target tracking algorithm has a serious problem of small target missed detection. The depth information channel is added, the depth information channel is connected with an original color image channel in parallel, fusion is carried out in the channel dimension, candidate frame extraction and targetdetection are carried out on the fused feature image, the detection rate of a small target is improved, in addition, training of a difficult sample is added in training, and the overall target recognition rate of the algorithm is improved. According to the method, a problem of small target missed detection by using the Faster R-CNN algorithm is fully considered. The accuracy of vehicle recognitionin a complex traffic scene is improved through depth image feature fusion and a difficult sample mining method.

The invention provides a 4D real traffic scene simulation based severe weather early-warning management system and method. The system comprises a multi-factor omnibearing tracking detection radar sensor, a data collection and analysis processor, a cloud processing server and main equipment of a 4D simulation management work station; a radar tracks vehicles and pedestrians on the road, obtains position information of the vehicles and pedestrians, and triggers a license plate snapshot camera to carry out real-time snapshot on vehicles synchronously; the system obtains node weather information ofthe present road to simulate the traffic road condition environment in a 4D manner and start an early warning rule, the road which the vehicles / pedestrians are passing is determined according to theposition information of the vehicles and pedestrians; and the alarm information is sent to the vehicles and pedestrians to pass the road in a point to point or broadcasting manner to make alarm prompt. According to 4D real traffic scene simulation, a road manager can use the system to make an emergency plan, carry out road control and rescue of vehicles and pedestrians in an accident, informationis pushed to avoid the planned route, and the risk of high-speed travel in the severe weather is reduced.

The invention discloses a night vehicle video detection method based on illumination visibility identification, which comprises the following steps of: 1) collecting a night traffic scene video image: using a camera lens to compress the collected traffic scene video image into MPEG (moving picture experts group) format and further transmitting to a computer for storage; 2) identifying the night illumination mode: determining whether the mode is the night mode with street lamps or the night mode without the street lamps; 3) carrying out vehicle detection under the night mode without the street lamps or carrying out the vehicle detection under the night mode with the street lamps; 4) tracking the motion of a night vehicle: utilizing the kalman filtering algorithm to carry out motion tracking on a matched vehicle headlamp (under the night mode without the street lamps) or the whole vehicle (under the night mode with the street lamps), obtaining the motion state of the vehicle and realizing the continuous and fast tracking of the motion of the vehicle; and 5) extracting a traffic parameter of the night vehicle: and adopting the two-dimensional reconstruction algorithm based on black box calibration to realize the extraction of a running speed parameter of the vehicle according to a projection relationship model between an image coordinate and a world coordinate.

The city traffic flow forecast and traffic information guidance system comprises the traffic information gathering subsystem, the communication subsystem, the traffic information integrated processing platform, the traffic information issuing subsystem and closed-circuit television monitoring subsystem. The traffic information gathering subsystem collects the traffic information of the scene, and transmits to the said traffic information integrated processing platform through the said communication subsystem, and the said traffic information integrated processing platform analyzes, integrates, and forecasts the traffic data, and then the real-time traffic information is directly transmitted to the traffic scene faced vehicles through the traffic information issuing subsystem. The invention can improve the passing efficiency of traffic system, and can significantly increase the drivers' understanding of the road traffic conditions, to adjust the arrived value of traffic volume, and make the arrived value of each trunk road to a balanced distribution, and avoid vehicles number exceeding over the actual capacity, thereby improving traffic jams, increase traffic smooth and traffic system operating efficiency.

The invention discloses a traffic video intelligent analysis method based on target detection and tracking. The method specifically comprises the following steps: using highway videos acquired by using monitoring cameras; detecting vehicle target detection by using a deep learning method, acquiring target tracks by using a multi-target tracking method according to a detection result, automaticallyanalyzing the target tracks acquiring traffic flow and vehicle speed, detecting traffic congestion and parking traffic abnormal events and completing intelligent analysis of traffic videos. Vehiclescan be detected and tracked for a long time in a visual field range, so that traffic parameters are accurately obtained, and traffic events are detected. The invention provided by the invention has high stability when being used in various traffic scenes, and has certain practical value and wide market potential.

The invention relates to an urban road traffic real-time state distinguishing and guidance system, consisting of a traffic information acquisition subsystem, a communication subsystem, a traffic information integrated processing platform, a traffic information release subsystem and a closed circuit television monitoring subsystem. The traffic information acquisition subsystem acquires the on-the-spot traffic information and transmits the traffic information to the traffic information integrated processing platform through the communication subsystem; the traffic information integrated processing platform adopts a traffic real-time state distinguishing algorithm to determine the traffic state of a specified section of the road according to a traffic flow, a speed, an occupancy and a periodic data packet, and transmits the determined result real-timely to the information release subsystem which then sends the result to a vehicle-oriented traffic scene. The urban road traffic real-time state distinguishing and guidance system of the invention can increase traffic efficiency of a traffic system, and improve the problem of the traffic jams.

The invention discloses a single camera calibration method based on road surface collinear three points and parallel line thereof, comprising the following steps: (1) building an image two-dimensional coordinate system and a ground two-dimensional world coordinate system; (2) in a traffic scene, selecting three points in straight roadway demarcation lines and parallel roadway edge straight lines as a calibration target, measuring the intervals among three points and the interval between two parallel straight lines, and obtaining the coordinate of the corresponding pixel of the calibrated target in an image; (3) calculating the angle of parallel lines in a ground world coordinate system; (4) calculating the focal distance and the depression angle parameter of the camera; and (5) calculating the coordinate conversion relation of each corresponding point in the image coordinate system and the ground world coordinate system to finally finish calibrating the camera. The invention does not need to put a regular calibration rectangle in a traffic scene, but only needs three points and the parallel lines thereof on a traffic marking. The method is simple and practical and has favorable pertinence and popularity.

The invention discloses a traffic scene simulation platform and method based on virtual reality. The traffic scene simulation platform comprises: a simulation scene control module (1), a man-machine interaction mechanism decision module (2), a network communication simulation module (3) and a manipulation input interface (4), wherein the simulation scene control module (1), the network communication simulation module (3) and the manipulation input interface (4) are connected with the man-machine interaction mechanism decision module (2), and the man-machine interaction mechanism decision module (2) has pre-packaged virtual driver characteristics; and the man-machine interaction mechanism decision module (2) simulates behavior decisions of a real driver or a virtual driver and other trafficelements for various scenes, carries out human-vehicle-road combined operation virtual simulation, and controls the virtual vehicle to carry out acceleration, deceleration, braking, overtaking and / orlane changing in the virtual environment based on the dynamic attributes of the vehicle in the real environment.

The invention discloses a method and device for testing the automatic driving performance. The method comprises the steps of receiving some data which is collected by a sensor installed on a manual driving vehicle and related to the actual traffic scene; constructing a test scene in allusion to the automatic driving performance based on the data; and testing the automatic driving performance in the constructed test scene. According to the scheme in various embodiments of the invention, the test scene is easy to be constructed and can reflect the real traffic environment and cover various complex traffic conditions, so that a test result for the automatic driving performance is enabled to be more accurate.

The invention discloses a traffic safety sensing network based on mobile information for road intelligent traffic management. The traffic safety sensing network takes a vehicle as a mobile intelligent sensing node and forms into a sensing network with a mobile communication network and a traffic monitoring center. A vehicle-mounted positioning system and a vehicle-mounted sensor immediately acquire the position and the speed of the vehicle as well as the image information in and out of the vehicle, thereby providing real-time road traffic condition information to a traffic monitoring center; and a monitoring vehicle and the traffic monitoring center can perform various information interaction through a mobile communication network. With a vision sensor and a vehicle driving status identifying system, the traffic safety sensing network has the functions for acquiring and processing the image, can realize the functions for monitoring the illegal driving and reappearing the traffic scene, assisting the safe driving of the vehicle, provides a new traffic and vehicle supervising route for the traffic management department, can improve the traffic capability of the road, can improve the driving safety of drivers, and effectively avoids and reduces the personal injury and the economic loss caused by traffic accident.

The invention discloses an automatic license plate recognition method and an automatic license plate recognition system. The system comprises a high-resolution camera (300), a detection and tracking module (312), a hazardous event detection module (316), an image extraction module (322) and a license plate recognition module (324), wherein the high-resolution camera (300) is used for acquiring a video image of a traffic scene; the detection and tracking module (312) is used for detecting and tracking a vehicle in the video image; the hazardous event detection module (316) is used for judging whether the tracked vehicle causes a hazardous event or not according to a set traffic event alarm rule; the image extraction module (322) is used for extracting a video image of the vehicle when the hazardous event detection module (316) judges the occurrence of the hazardous event of the vehicle; and the license plate recognition module (324) is used for detecting and recognizing the license plate of a traffic violation vehicle in the video image of the vehicle. The automatic license plate recognition method and the automatic license plate recognition system can be applied to more complex traffic scenes; simultaneously, the types of recognizable traffic violation events of traffic vehicles are expanded greatly.

The invention relates to the technical field of automatic driving, in particular to an intelligent control platform and method for an automatic driving closed test field. The intelligent control platform for the automatic driving closed test field comprises a display layer, a control layer and a data layer; the display layer is used for displaying a complete test process and test data in the testfield; the control layer is used for realizing an automatic driving test and performing control and management on the display layer and the data layer; and the data layer is used for storing traffic scene data and test data, analyzing and evaluating the test data, and generating a test report. The intelligent control platform and method can be directly applied to the design and implementation of acontrol center of the automatic driving closed test field, and an intelligent and unmanned solution is provided for a control system of the automatic driving test field.

The invention relates to a test system for an automatic driving vehicle. A real vehicle data collection module is used for collecting control data; a test bench is used for generating a reaction forceaccording to the control data and scene information and obtaining state data of an automatic driving vehicle according to the reaction force; a scene model establishing module is used for establishing a traffic scene model, a traffic flow model, an automatic driving vehicle dynamics model and an environment perception sensor model and obtaining position data of the automatic driving vehicle; andan evaluation module is used for calculating a return value for evaluating the running state of the automatic driving vehicle according to the position data and the state data. By combining the test bench with a scene model comprising traffic flow data, the environment recognition capability, the dangerous condition recognition response time and the processing mode of the automatic driving vehiclecan be more truly tested, and the safety and the stability of the autonomous recognition capability under a complex traffic environment can be more truly tested.

A vehicle lighting system for illuminating ahead of a vehicle is provided with a plurality of light sources each adapted to illuminate a different illumination area to thereby form a predetermined light distribution pattern as a whole and a control unit for determining a traffic scene in which the vehicle is running and independently changing at least either an orientations or the light quantities of the plurality of light sources so as to form a target light distribution pattern which has been set in advance for the traffic scene.

The invention discloses a drivable region detection method based on a laser radar. The method includes the steps that the laser radar is used for scanning an actual traffic scene, and flat and rugged conditions of a current scene are obtained; scanning data returned by the laser radar are divided into a plurality of fragments through the height variation condition, an iterative linear fitting algorithm is used for extracting appropriate linear features from all the segments, flat regions are found out from the fragments which meet constraint conditions, and drivable regions are determined from the flat regions according to the requirements of the vehicle width; if a multi-line laser exists, corresponding drivable regions can be determined from the flat regions reflected by the line data on the basis of a drivable region of former line data from the near to the distant. According to the method, detection of the drivable regions in the actual traffic scene can be effectively completed, safe passage regions of an automobile are described accurately, and the method has good real-time performance, usability and robustness.

The invention relates to a ship intelligent control system test platform with virtual and real fusion. The ship intelligent control system test platform with virtual and real fusion comprises a real navigation database, a real-time simulation system for generating a numerical twinning fused traffic scene according to data in the real navigation database, a ship simulation system for simulating a ship in the test scene, and a tested intelligent control system for controlling and simulating the ship according to the data provided by the ship simulation system. The experimental environment constructed by the ship intelligent control system test platform with virtual and real fusion disclosed by the invention comes from real data collection, which ensures the validity of the test; and the based object experimental ship and experimental process are still based on numerical simulation, which reduces the risk and cost of the test; that is, the experimental object ship is based on numerical simulation, and the experimental scene completely comes from the perception fusion of the real world.