CN117253365A - Automatic detection method and related device for vehicle traffic condition - Google Patents

Automatic detection method and related device for vehicle traffic condition Download PDF

Info

Publication number
CN117253365A
CN117253365A CN202311534676.4A CN202311534676A CN117253365A CN 117253365 A CN117253365 A CN 117253365A CN 202311534676 A CN202311534676 A CN 202311534676A CN 117253365 A CN117253365 A CN 117253365A
Authority
CN
China
Prior art keywords
detector
combined
training
output value
combination
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.)
Granted
Application number
CN202311534676.4A
Other languages
Chinese (zh)
Other versions
CN117253365B (en
Inventor
杨扬
胡心怡
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shanghai Boonray Intelligent Technology Co Ltd
Original Assignee
Shanghai Boonray Intelligent Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shanghai Boonray Intelligent Technology Co Ltd filed Critical Shanghai Boonray Intelligent Technology Co Ltd
Priority to CN202311534676.4A priority Critical patent/CN117253365B/en
Publication of CN117253365A publication Critical patent/CN117253365A/en
Application granted granted Critical
Publication of CN117253365B publication Critical patent/CN117253365B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0125—Traffic data processing
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0125—Traffic data processing
    • G08G1/0129—Traffic data processing for creating historical data or processing based on historical data
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0137—Measuring and analyzing of parameters relative to traffic conditions for specific applications

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Traffic Control Systems (AREA)

Abstract

A method and a related device for automatically detecting traffic conditions of vehicles. The method comprises the following steps: a pre-training detector acquisition step of acquiring a first number of pre-training detectors for various types of vehicle traffic condition detection; a traffic condition detection step of inputting vehicle sensor data into the pre-training detector to obtain an output value of an output node of the pre-training detector; judging whether the pre-training detector needs combined detection or not based on the output value; a detector combination step of acquiring a ground area where the vehicle is located; acquiring a terrain point set of the ground area; a combined detector determining step, namely determining a detector combination corresponding to a combined node as a new combined detector after a plurality of times of combined node judgment meeting preset conditions; and a traffic condition combination detection step, namely inputting the sensor data to be detected into the combination detector to obtain a vehicle traffic condition detection result.

Description

Automatic detection method and related device for vehicle traffic condition
Technical Field
The present disclosure relates to the field of autopilot, and in particular to a vehicle traffic condition automatic detection method and apparatus, an electronic device, a storage medium, a program product, and an autopilot vehicle.
Background
The automatic driving technology is a technology for realizing auxiliary driving or unmanned driving through a computer, and the automatic driving technology depends on a visible light camera, a millimeter wave radar, a laser radar, an inertial navigation system, a global positioning system and other sensing systems, so that the computer can partially or completely replace a human driver to automatically and safely operate the vehicle.
In the prior art, the automatic driving technology is mainly applied to standard road scenes. However, in non-standard road scenarios such as mining areas, traffic conditions are variable and it is difficult to adapt to new conditions using pre-trained traffic condition detection detectors. While the prior art can accommodate new situations by frequently upgrading traffic condition detection detectors, frequent upgrades require higher costs and delay before upgrades.
It is therefore desirable that the traffic condition detector can adapt to new traffic conditions when detecting traffic conditions.
Disclosure of Invention
The disclosure provides a vehicle traffic condition automatic detection method and a related device, a storage medium and a vehicle.
According to a first aspect of the present disclosure, there is provided a vehicle traffic condition automatic detection method, including:
a pre-training detector acquisition step of acquiring a first number of pre-training detectors for various types of vehicle traffic condition detection;
A traffic condition detection step of inputting vehicle sensor data into the pre-training detector to obtain an output value of an output node of the pre-training detector; judging whether the pre-training detector needs combined detection or not based on the output value;
a detector combining step, namely inputting an output value of a pre-training detector to be combined and detected into a newly added combining node; the combined node judges whether the current output value is similar to the historical output value or not;
a combined detector determining step, namely determining a detector combination corresponding to a combined node as a new combined detector after a plurality of times of combined node judgment meeting preset conditions;
and a traffic condition combination detection step, namely inputting the sensor data to be detected into the combination detector to obtain a vehicle traffic condition detection result.
According to a second aspect of the present disclosure, there is provided an automatic vehicle traffic condition detection apparatus including:
a pre-training detector acquisition module that acquires a first number of pre-training detectors for use in various types of vehicle traffic condition detection;
the traffic condition detection module inputs the vehicle sensor data into the pre-training detector to obtain an output value of an output node of the pre-training detector; judging whether the pre-training detector needs combined detection or not based on the output value;
The detector combination module inputs the output value of the pre-training detector to be detected in a combined way into a newly added combination node; the combined node judges whether the current output value is similar to the historical output value or not;
the combined detector determining module determines the detector combination corresponding to the combined node as a new combined detector after a plurality of times of combined node judgment meeting the preset condition;
and the traffic condition combination detection module inputs the sensor data to be detected into the combination detector to obtain a vehicle traffic condition detection result.
According to a third aspect of the present disclosure, there is provided an electronic device comprising:
at least one processor, a memory, and a communication interface to communicate with other electronic devices;
the memory stores instructions executable by the at least one processor to enable the electronic device to perform the vehicle traffic condition automatic detection method of the first aspect.
According to a fourth aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions for causing the computer to execute the vehicle traffic condition automatic detection method according to the first aspect.
According to a fifth aspect of the present disclosure, there is provided a computer program product comprising a computer program which, when executed by a processor, implements the vehicle traffic condition automatic detection method according to the first aspect.
According to a sixth aspect of the present disclosure, there is provided an autonomous vehicle comprising an electronic device according to the third aspect.
It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the disclosure, nor is it intended to be used to limit the scope of the disclosure. Other features of the present disclosure will become apparent from the following specification.
Compared with the prior art, the invention has the beneficial effects that:
when the pre-training traffic condition detection method and apparatus are applied to a new scene new environment, the pre-training method and apparatus may not be adaptable to all situations in the new scene new environment. The method and the device can effectively expand the pre-training method and the device, and adapt to the new situation under the new environment of the new scene by combining the pre-training detectors. The method for adapting to the new environment of the new scene has better adaptability, thereby reducing the cost of manually modifying the method and returning the device to the factory for adjustment and greatly improving the working efficiency.
Drawings
The drawings are for a better understanding of the present solution and are not to be construed as limiting the present disclosure. Wherein:
FIG. 1 illustrates a schematic diagram of a nonstandard road scene provided in accordance with one embodiment of the present disclosure;
FIG. 2 illustrates a schematic diagram of a vehicle traffic condition automatic detection method provided in accordance with one embodiment of the present disclosure;
FIG. 3 illustrates a schematic diagram of a vehicle traffic condition automatic detection device provided in accordance with one embodiment of the present disclosure;
fig. 4 shows a schematic diagram of an electronic device provided according to one embodiment of the present disclosure.
Detailed Description
Exemplary embodiments of the present disclosure are described below in conjunction with the accompanying drawings, which include various details of the embodiments of the present disclosure to facilitate understanding, and should be considered as merely exemplary. Accordingly, one of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.
The technical terms involved include:
standard road scene: the standard road is a road which accords with the rules and regulations related to road traffic and has information such as specific marking marks and the like, and the road surface of the road is subjected to leveling hardening. Under a standard road scene, the automatic driving technology can acquire accurate road information through information such as standard road marks, marks and the like, and can also perform standard information exchange with traffic infrastructure, so that environment information necessary for automatic driving is acquired. The automatic driving technology under the standard road scene has stronger universality in different road scenes and environments.
Non-standard road scene: the nonstandard road scene is a road without information such as a marking mark specified by the laws and regulations related to road traffic. Non-standard road scenes generally have more uncertain traffic conditions than standard road scenes. For example, an unanticipated traffic participant may be present in an unanticipated scene. These uncertain traffic conditions result in a failure to pre-exhaust all possible traffic conditions. So that the pre-trained traffic condition detector cannot be used directly for judging the traffic condition of the nonstandard road scene.
Traffic conditions: traffic conditions include the condition of the traffic facility and the condition of the traffic participant. Traffic facilities such as routes and obstacles. Traffic participants such as vehicles, pedestrians, non-motor vehicles. Traffic conditions include not only the conditions of the traffic facility and the traffic participants themselves, such as location and speed, but also some conditions that affect each other, such as traffic facility restrictions on the traffic participants, and so forth.
Pre-training the detector: a detector for detecting traffic conditions. Typically pre-trained, can accommodate general standard road scenarios and/or non-standard road scenarios. Typically pre-deployed on a vehicle. However, when the vehicle is running in a specific traffic scenario, the pre-training detector cannot extract the unexpected traffic condition, and thus retrains or adaptively detects according to the scene condition, because the unexpected traffic condition may occur.
A combination detector: for assisting the pre-training detector to detect traffic conditions. When the output values of the pre-trained detectors do not meet the requirements of confidence, accuracy, or integrity, etc., a combination of multiple pre-trained detectors may still characterize a particular traffic condition. For example, when the traffic light detector only judges that the green light is not on from the incomplete red-green image, but the vehicle detector detects that the front vehicle is waiting, the combined detector can still judge that the vehicle is in the current red light state.
Example 1
Fig. 1 is a schematic diagram of a non-standard road scenario.
In the prior art, the automatic driving technology is mainly applied to standard road scenes. Under standard road scenarios, traffic facilities and traffic participants are typically limited and controllable, and thus pre-trained detectors that detect traffic conditions can generally meet the general needs under standard road scenarios.
Under non-standard road scenes, traffic facilities or traffic participants are generally uncontrollable, for example, natural scenes such as field environments, agricultural environment scenes such as rural soil roads, specific operation scenes such as mine shafts, and the like, and the landforms, road routes, obstacles and the like of the traffic facilities are easy to change; unexpected traffic participants are prone to occur in various traffic scenarios.
In summary, there are at least the following difficulties in the research of automatic detection of traffic conditions in non-standard road scenes such as mining areas: unexpected traffic conditions are prone to occur; it is difficult for the pre-trained detector to pre-exhaust all possible traffic conditions.
The present embodiment performs traffic condition detection more accurately and in real time based on real-time adaptive reconstruction of the pre-trained detector.
Fig. 2 shows a schematic diagram of an automatic detection method of a vehicle traffic condition.
The automatic detection method for the traffic condition of the vehicle provided by the embodiment of the disclosure comprises the following steps:
s110, a pre-training detector obtaining step, namely obtaining a first number of pre-training detectors, wherein the pre-training detectors are used for detecting various types of traffic conditions of vehicles;
s120, traffic condition detection, namely inputting vehicle sensor data into the pre-training detector to obtain an output value of an output node of the pre-training detector; judging whether the pre-training detector needs combined detection or not based on the output value;
s130, a detector combining step, namely inputting an output value of a pre-training detector to be combined and detected into a newly added combining node; the combined node judges whether the current output value is similar to the historical output value or not;
S140, determining a combination detector, namely determining a detector combination corresponding to a combination node as a new combination detector after a plurality of times of combination node judgment meeting preset conditions;
s150, a traffic condition combination detection step, namely inputting sensor data to be detected into the combination detector to obtain a vehicle traffic condition detection result.
In this embodiment, in the pre-training detector obtaining step S110: a first number of pre-trained detectors is acquired for various types of vehicle traffic condition detection.
In one embodiment, the pre-training detector comprises a detector that detects a traffic facility condition and a detector that detects a traffic participant condition in the pre-training detector acquisition step.
In one embodiment, in the pre-training detector acquiring step, the pre-training sensor receives sensor data in the same format; alternatively, the pre-trained sensor receives sensor data in a different format.
Specifically, the pre-training detector has an input node that accepts input of vehicle sensing data and an output node. The output value of the output node depends on the type of detector. For example, the classifier may output a class label (and confidence) and the object detector may output an object name (and probability).
In one embodiment, the pre-trained sensor receives sensor data in the same format. At this time, the pretrained sensor receives sensor data in the same format, i.e., various sensor data acquired by the vehicle sensor, and can be simultaneously provided to a plurality of pretrained sensors.
In one embodiment, the pre-trained sensor receives sensor data in a different format. In this case, the types of sensors to which the pre-training detector is directed are different, and for example, the detectors are directed to image data, and the detectors are directed to laser radar data. The detector acquires sensor data input and then independently processes the sensor data input.
The beneficial effects of step S110 include:
(1) The pre-training detectors can be flexibly combined; the preset can be performed according to different traffic scenes.
(2) The pre-trained detector is compatible with the same format or different formats of sensor data. In the face of a detector that can only detect one traffic condition, other pre-trained detectors can be conveniently added, and the newly added detector and the original pre-trained detector can share the same format of sensor data, and can also respectively use different formats of sensor data. The newly added detectors may use sensor data in the same format or in different formats among the newly added detectors in the face of newly added detectors that can detect various traffic conditions. It is therefore convenient to add complex detectors to the pre-trained detector that can detect a variety of traffic conditions tasks.
In this embodiment, in the traffic condition detection step S120: inputting vehicle sensor data into the pre-training detector, and obtaining an output value of an output node of the pre-training detector; and judging whether the pre-training detector needs combined detection or not based on the output value.
In one embodiment, in the traffic condition detection step, the sensor data local to the vehicle includes traffic facility conditions and traffic participant conditions.
In particular, traffic facilities include road terrain, traffic markings, traffic signs, obstacles, and the like. Traffic participants include vehicles, pedestrians, non-motor vehicles, and the like.
Specifically, the sensor data of the vehicle local area is acquired based on an on-vehicle sensor; or, the sensor data of the vehicle local is obtained based on the vehicle-road cooperation.
In one embodiment, in the traffic condition detection step, the determining whether the pre-training detector needs combination detection based on the output value includes: if any of the pre-trained sensors or existing combination detectors can detect the sensor data, then no combination detection is required; and/or if any of the pre-trained sensors or existing combination detectors can detect the sensor data, then no combination detection is required; and/or if either the pre-trained detector or an existing combination detector is unable to detect the sensor data, combination detection is required.
Specifically, the determination criteria for the combination detection may be not the above criteria, but the detectors may be listed as required for the combination detection according to specific needs. When a detector itself is required to be used in combination with other detectors, for example, the detector can correctly detect a part of an entire object, or the detector can be directly incorporated into the range of detection required to be combined, for example, the detector can detect the entire object together with other detectors.
Specifically, mapping each sensor data into a vector of a predetermined length includes: mapping each traffic facility condition to a vector of a predetermined length; each traffic participant condition is mapped to a vector of predetermined length.
Specifically, mapping each traffic facility condition into a vector of a predetermined length includes: mapping the terrain condition to a first vector of a first predetermined length; mapping traffic marking conditions to a second vector of a second predetermined length; mapping the traffic sign condition to a third vector of a third predetermined length; the obstacle condition is mapped to a fourth vector of a fourth predetermined length.
Specifically, mapping each traffic participant condition to a vector of predetermined length includes: mapping the vehicle condition to a fifth vector of a fifth predetermined length; mapping the pedestrian condition to a sixth vector of a sixth predetermined length; the non-motor vehicle condition is mapped to a seventh vector of a seventh predetermined length.
In particular, the various mappings described above may be dimensionality reduction algorithms. The various mappings described above may also be implemented based on a mapping matrix, i.e., the sensor data vector is multiplied by the mapping matrix to obtain a vector of predetermined length.
Specifically, combining the vectors of the predetermined length together according to a predetermined embedding matrix includes: and multiplying each vector reflecting different traffic conditions by the corresponding embedding matrix, and embedding the vector into the feature vector.
Specifically, the embedding matrix may embed the respective corresponding vectors into predetermined positions of the feature vectors. And performing matrix multiplication on the respective corresponding vectors and the embedding matrix to obtain the vector to be embedded with the feature vector.
In particular, the combination detection is required, i.e. in combination with other detectors, to obtain detection results.
Specifically, the output value is mapped to whether or not the combination detection is required, and the output value can be distinguished according to a threshold value. When the output value of the detector is greater than (or equal to) the detection threshold, the detection result is "detection". When the output value of the detector is less than (or equal to) the undetected threshold value, the detection result is "undetected". When the output value of the detector is between the detection threshold and the non-detection threshold, the detection result is "combination detection is required". Wherein the detection threshold is greater than the undetected threshold.
The beneficial effects of step S120 include:
(1) The obtained comprehensive sensor data of the vehicle can provide the information as comprehensive as possible for detecting traffic conditions.
(2) The sensor data is mapped into a vector with a preset length according to a preset mapping method and embedded into the feature vector, so that the comprehensive information can be converted into the feature vector with a preset format, and the feature vector can be input into the detector.
In this embodiment, in the detector combining step S130: inputting the output value of the pre-training detector to be combined and detected into a newly added combined node; the combining node determines whether the current output value is similar to the historical output value.
In one embodiment, in the step of combining the detectors, the combining node determines whether the current output value is similar to the historical output value, and the criteria include: the average distance between the vector of current output values and the vector of historical output values is less than a predetermined threshold.
Specifically, during operation of the detector, it is necessary to save the historical output value and compare the current output value vector with the historical output value vector. The historical output value vectors can be organized in time sequence, or all the historical output value vectors can be organized into an array or a collection structure. By comparing and analyzing the historical output value vector set and the current output value vector, whether the current output value vector is similar to the historical output value vector or not can be known. In order to express the similarity, the set of historical output value vectors can be regarded as a cluster, and when the current output value vector can be classified into the cluster, the two are similar. That is to say,
In one embodiment, in the step of combining the detectors, the combining node determines whether the current output value is similar to the historical output value, and the criteria include: the distance between the vector formed by the current output value and the clustering center vector of the vector formed by the historical output value is smaller than a preset threshold value.
In particular, the input of the combining node is the output node of the pre-trained detector that needs to combine the detection.
Specifically, not all pre-trained detectors participate in the combination when performing the combination detection. Wherein the inputs of part of the detectors may not be functional in the combined node. I.e. the only possible part of the pre-trained detectors involved in the actual combination.
In one embodiment, the combined detector is adapted to express traffic conditions that cannot be expressed by any one of the pre-trained detectors, but can be expressed by a combination of multiple detectors.
Specifically, when a traffic condition cannot be represented by any one of the existing pre-trained detectors, it means that the detection result of each detector does not reach the degree of independent detection.
For example, in the process of detecting a certain feature vector input by the pre-training detector, all detectors cannot detect the feature vector alone (i.e., the output results of all the detector nodes are not "detected"), so that detection by a combination of detectors is required (i.e., the output results of some detectors are "undetected", but the output results of some detectors are "required to be detected).
The combination detector in this case is a combination of detectors that output the result "combination detection is required". A combination detector embodied at a combination node other than the detector output node. The combination detector also needs to hit the input information by multiple rounds to become a confirmed combination detector.
The beneficial effects of step S130 include:
(1) For an undetected feature vector, it is explained that the current pre-trained detector fails to perform an effective individual detection of the feature vector. However, if the detection is terminated in this way, the task of detection cannot be completed. If the partial detector with output result can be used further, for example, by the combination of the existing detectors, the task of detection can still be completed if effective detection is possible. I.e. the detection capability of the pre-trained detector is extended.
(2) For the combination of detectors, the result of one detection is not simply used for the combination. But is a formal combination detector after confirmation by the result of the multiple inputs. However, the validation process is not a formal and time-consuming training process, but is performed during use, similar to the process of unsupervised learning to develop clusters. The combined detector formally holds only when multiple inputs of information can support the conclusion that the combined detector is gradually formed (i.e., the gradual determination of the combination of detectors can be confirmed). I.e. the stability and reliability of the output result of the combined detector is improved.
In this embodiment, in the step S140 of determining the combined detector, after the combined nodes satisfying the preset condition are determined for several times, the detector combination corresponding to the combined nodes is determined as a new combined detector.
In one embodiment, the several detections are that the detection results obtained by different input information hit a certain combination detector, and the whole detection process meets a preset condition, and the hit times are greater than a preset time threshold.
In one embodiment, the preset condition includes: the similarity of the current and the historical sensor data satisfies a first value condition and/or the output value of the current and the historical pre-trained detector satisfies a second value condition.
Specifically, the first value condition is that the similarity is greater than a first threshold; the second value condition is that the output value is larger than a second threshold value.
Specifically, the detectors in the combined detector include a pre-trained detector that outputs a result of "combined detection is required". Further, after a plurality of times of detection meeting the preset condition, the output results are all the detectors requiring combined detection. The technical scheme has the beneficial effects that the stability of the combined detector is better, and the detection result actually represented by the input information can be represented better.
In one embodiment, the combination of detectors is based on weights. The detectors in the combined detector include detectors in which the weights in the combination are greater than a predetermined weight threshold.
Specifically, the combination detector is a combination after excluding an impossible combination of detectors in advance. The technical scheme has the beneficial effect that the calculation efficiency of the combined detector is higher.
Specifically, the combination detector is a pre-specified detector combination.
In one embodiment, the preset condition includes: when the current and the historical sensor data meet the first time sequence condition, and/or the current and the historical output values of the pre-training detector meet the second time sequence condition.
Specifically, the first timing condition is that the variation of the feature vector in time sequence accords with a first specific mode; the second timing condition is that the change in timing of the output value conforms to a second particular pattern.
The beneficial effects of step S140 include:
(1) Whether the combination detector is established is determined by judging whether the output value of the detector in the combination detector satisfies a specific pattern. So that input information that would otherwise not be detectable by a single detector can be detected by a combination of multiple detectors.
(2) Determining whether the particular pattern is established includes spatially conforming the output value of the detector to the particular pattern and/or temporally conforming the output value of the detector to the particular pattern. So that the spatial combination of detectors, as well as the temporal combination of detectors, can be compatible.
(3) While the output value of the detector corresponds to a particular pattern, the input value of the detector (feature vector) also corresponds to a similar pattern, so that the sub-combination pattern can be determined more efficiently.
In this embodiment, in the step S150 of detecting the traffic condition combination, the sensor data to be detected is input to the combination detector, so as to obtain a detection result of the traffic condition of the vehicle.
In one embodiment, suitable combination detectors are found for those sensing information that are "to be detected in combination" through multiple rounds of learning. By inputting the above-described sensing information into the newly found combination detector, a certain detection result can be obtained.
In one embodiment, the detection result output by the combination detector is a combination of detection results of a plurality of detectors. The combination is not arbitrary, but a stable detection result obtained through multiple rounds of learning.
In one embodiment, the sensor data to be detected is input to all of the pre-trained detectors and all of the validated combined detectors; and obtaining a vehicle traffic condition detection result according to the output results of all the pre-training detectors and the combined detector. The technical effect of the above technical solution is that after a period of time of processing and learning of input sensing information, a plurality of combined detectors may have been acquired; for the sensing information to be detected, the most suitable combination detector needs to be obtained from a plurality of combination detectors.
In one embodiment, the pre-training detector is set to have a higher priority than the combined detector, and the combined detector having a higher similarity to the historical output value is set to have a higher priority than the combined detector having a lower similarity to the historical output value; the vehicle traffic condition detection result is output according to the priority of the detector.
In one embodiment, the output results of all of the detectors of all of the plurality of combined detectors are calculated; determining whether the output result of each detector is valid (e.g., determining that the calculation of the confidence level above the confidence level threshold is valid); the combined detector with the largest number of valid detectors is determined to be the best combined detector.
The embodiment is not limited to an application scenario and a specific implementation, and may be determined according to an actual situation, which is not described herein.
The beneficial effects of step S150 include:
(1) After a period of learning, the pre-trained detector has extended a number of combined detectors that can be used to detect the input information. I.e. the ability of the pre-trained detector is extended.
(2) The combination mode of each detector in the combined detector is stable, so that a stable detection result can be provided for detecting input information.
This embodiment may be implemented alone or in combination with other embodiments.
Example two
The embodiment of the disclosure provides an automatic detection device for vehicle traffic conditions. As shown in fig. 3, includes:
a pre-training detector acquisition module 110 that acquires a first number of pre-training detectors for various types of vehicle traffic condition detection;
the traffic condition detection module 120 inputs vehicle sensor data into the pre-training detector to obtain an output value of an output node of the pre-training detector; judging whether the pre-training detector needs combined detection or not based on the output value;
the detector combining module 130 inputs the output value of the pre-training detector to be detected by combining into a newly added combining node; the combined node judges whether the current output value is similar to the historical output value or not;
the combined detector determining module 140 determines the detector combination corresponding to the combined node as a new combined detector after a plurality of times of combined node judgment meeting the preset condition;
the traffic condition combination detection module 150 inputs the sensor data to be detected into the combination detector to obtain a detection result of the traffic condition of the vehicle.
The beneficial effects of each module of the above device are as in the previous embodiment, and will not be described here again.
It should be noted that, the embodiment of the present disclosure is not limited to specific implementation of the application scenario of the apparatus, which may be determined according to actual situations, and will not be described herein.
It should be noted that, it should be understood that the division of the modules of the above apparatus is merely a division of a logic function, and may be fully or partially integrated into a physical entity or may be physically separated. And these modules may all be implemented in software in the form of calls by the processing element; or can be realized in hardware; the method can also be realized in a form of calling software by a processing element, and the method can be realized in a form of hardware by a part of modules. For example, the processing module may be a processing element that is set up separately, may be implemented in a chip of the above-mentioned apparatus, or may be stored in a memory of the above-mentioned apparatus in the form of program codes, and the functions of the above processing module may be called and executed by a processing element of the above-mentioned apparatus. The implementation of the other modules is similar. In addition, all or part of the modules can be integrated together or can be independently implemented. The processing element described herein may be an integrated circuit having signal processing capabilities. In implementation, each step of the above method or each module above may be implemented by an integrated logic circuit of hardware in a processor element or an instruction in a software form.
This embodiment may be implemented alone or in combination with other embodiments.
Example III
As shown in fig. 4, in the present embodiment, an electronic device 600 includes:
at least one processor 601, a memory 608, and a communication interface 609 to communicate with other electronic devices; the memory 608 stores instructions executable by the at least one processor to enable the electronic device to perform the vehicle traffic condition automatic detection method of the previous embodiments.
Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the disclosure described and/or claimed herein. The electronic device may be the first device described above, or may be a vehicle control device, or a control center on a vehicle, which is not limited in this aspect.
As shown in fig. 4, the electronic device further includes: one or more ROM602, RAM603, bus 604, I/O interface 605, input unit 606, output unit 607, etc., as well as interfaces for connecting components, including high-speed and low-speed interfaces, and communication interfaces for communicating with other electronic devices. The various components are interconnected using different buses and may be mounted on a common motherboard or in other manners as desired. The processor may process instructions executing within the electronic device, including instructions stored in or on memory to display graphical information of the GUI on an external input/output device, such as a display device coupled to the interface. In other embodiments, multiple processors and/or multiple buses may be used, if desired, along with multiple memories and multiple memories. Also, multiple electronic devices may be connected, each providing a portion of the necessary operations (e.g., as a server array, a set of blade servers, or a multiprocessor system). In this embodiment, a processor 601 is taken as an example.
Memory 608 is a non-transitory computer-readable storage medium provided by the present disclosure. Wherein the memory stores instructions executable by the at least one processor to cause the at least one processor to perform the methods provided by the present disclosure. The non-transitory computer readable storage medium of the present disclosure stores computer instructions for causing a computer to perform the methods provided by the present disclosure. Memory 608, as a non-transitory computer readable storage medium, may be used to store non-transitory software programs, non-transitory computer-executable programs, and modules, such as program instructions/modules, corresponding to the methods in embodiments of the present disclosure. The processor 601 executes various functional applications of the server and data processing, i.e. implements the methods of the method embodiments described above, by running non-transitory software programs, instructions, and modules stored in the memory 608.
Memory 608 may include a storage program area that may store an operating system, at least one application program required for functionality, and a storage data area; the storage data area may store data created according to the use of electronic devices controlled by the autonomous vehicle, and the like. In addition, memory 608 may include high-speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory 608 may optionally include memory located remotely from processor 601, which may be connected to the data processing electronics via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The various components of the electronic device may be connected by a bus or other means, in this embodiment by way of example.
The input unit 606 may receive input numeric or character information and generate key signal inputs related to user settings and function control of the data processing electronic device, such as a touch screen, a keypad, a mouse, a track pad, a touch pad, a pointer stick, one or more mouse buttons, a track ball, a joystick, and the like. The output unit 607 may include a display device, auxiliary lighting means (e.g., LEDs), tactile feedback means (e.g., vibration motors), and the like. The display device may include, but is not limited to, a Liquid Crystal Display (LCD), a Light Emitting Diode (LED) display, and a plasma display. In some implementations, the display device may be a touch screen.
This embodiment may be implemented alone or in combination with other embodiments.
Example IV
According to the present embodiment, there is provided a non-transitory computer-readable storage medium storing computer instructions for causing the computer to execute the automatic vehicle traffic condition detection method according to the foregoing embodiment.
This embodiment may be implemented alone or in combination with other embodiments.
Example five
According to the present embodiment a computer program product is provided, comprising a computer program which, when being executed by a processor, implements the vehicle traffic condition automatic detection method according to the previous embodiment.
The computer-readable storage medium and computer program product of the above embodiments storing a computer program (also referred to as a program, software application, or code) includes machine instructions of a programmable processor, and may be implemented using high-level procedural and/or object-oriented programming languages, and/or assembly/machine languages. As used herein, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, apparatus, and/or device (e.g., magnetic discs, optical disks, memory, programmable Logic Devices (PLDs)) used to provide machine instructions and/or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal used to provide machine instructions and/or data to a programmable processor. The present embodiment is not particularly limited thereto.
This embodiment may be implemented alone or in combination with other embodiments.
Example six
According to the present embodiment there is provided an autonomous vehicle comprising an apparatus according to the above embodiments.
It will be appreciated that the present embodiment is equally applicable to a manned vehicle which may assist in controlling the operation of the vehicle in the form of a prompt or automatic control provided to the driver based on the acquired road information. Some vehicles are equipped with a car-driving computer or an On Board Unit (OBU), and some vehicles are equipped with a user terminal such as a mobile phone, a user holding the user terminal, and the like. A cell phone, a driving computer or an OBU in the vehicle may be used as an electronic device for performing detector training or driving assistance.
It will be appreciated that the present embodiment is equally applicable to an intelligent traffic network, where a plurality of vehicles capable of wireless communication, and traffic control devices, remote servers, road side devices, and base stations for wireless communication with respective vehicles may be included in the intelligent traffic network, where the remote servers or the traffic control devices may also control traffic facilities, and so on.
The present embodiment does not limit the types, the number, and the application scenario of the vehicles.
This embodiment may be implemented alone or in combination with other embodiments.
It should be appreciated that various implementations of the systems and techniques described in this disclosure may be implemented in digital electronic circuitry, integrated circuitry, special purpose ASICs (application specific integrated circuits), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device. The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), and the internet. The computing system may include clients and servers. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.
It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps described in the present disclosure may be performed in parallel, sequentially, or in a different order, so long as the desired result of the technical solution of the present disclosure can be achieved, and the present disclosure is not limited herein.
The above detailed description should not be taken as limiting the scope of the present disclosure. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present disclosure are intended to be included within the scope of the present disclosure.

Claims (16)

1. An automatic detection method for traffic conditions of vehicles, comprising:
a pre-training detector acquisition step of acquiring a first number of pre-training detectors for various types of vehicle traffic condition detection;
a traffic condition detection step of inputting vehicle sensor data into the pre-training detector to obtain an output value of an output node of the pre-training detector; judging whether the pre-training detector needs combined detection or not based on the output value;
A detector combining step, namely inputting an output value of a pre-training detector to be combined and detected into a newly added combining node; the combined node judges whether the current output value is similar to the historical output value or not;
a combined detector determining step, namely determining a detector combination corresponding to a combined node as a new combined detector after a plurality of times of combined node judgment meeting preset conditions;
and a traffic condition combination detection step, namely inputting the sensor data to be detected into the combination detector to obtain a vehicle traffic condition detection result.
2. The method of claim 1, wherein in the pre-training detector acquisition step, the pre-training detector comprises a detector that detects traffic conditions and a detector that detects traffic participant conditions.
3. The method of claim 2, wherein in the pre-training detector acquiring step, the pre-training sensor receives sensor data in the same format; alternatively, the pre-trained sensor receives sensor data in a different format.
4. The method of claim 1, wherein in the traffic condition detecting step, the sensor data includes traffic facility conditions and traffic participant conditions.
5. The method of claim 4, wherein in the traffic condition detecting step, the determining whether the pre-training detector requires combined detection based on the output value comprises: if any of the pre-trained sensors or existing combination detectors can detect the sensor data, then no combination detection is required; and/or if any of the pre-trained sensors or existing combination detectors can detect the sensor data, then no combination detection is required; and/or if either the pre-trained detector or an existing combination detector is unable to detect the sensor data, combination detection is required.
6. The method of claim 1, wherein in the detector combining step, the combining node determines whether the current output value is similar to the historical output value, and criteria include: the average distance between the vector of current output values and the vector of historical output values is less than a predetermined threshold.
7. The method of claim 6, wherein in the detector combining step, the combining node determines whether the current output value is similar to the historical output value, and criteria include: the distance between the vector formed by the current output value and the clustering center vector of the vector formed by the historical output value is smaller than a preset threshold value.
8. The method of claim 1, the preset conditions comprising: the similarity of the current and the historical sensor data satisfies a first value condition and/or the output value of the current and the historical pre-trained detector satisfies a second value condition.
9. The method of claim 1 or 8, the preset conditions comprising: when the current and the historical sensor data meet the first time sequence condition, and/or the current and the historical output values of the pre-training detector meet the second time sequence condition.
10. The method of claim 1, the traffic condition combination detection step comprising:
inputting sensor data to be detected into all the pre-trained detectors and all confirmed combined detectors; and obtaining a vehicle traffic condition detection result according to the output results of all the pre-training detectors and the combined detector.
11. The method of claim 10, wherein the obtaining the vehicle traffic condition detection result from the output results of all the pre-trained detectors and the combined detector comprises:
setting the priority of the pre-training detector higher than that of the combined detector, and setting the priority of the combined detector with high similarity to the historical output value higher than that of the combined detector with low similarity to the historical output value;
The vehicle traffic condition detection result is output according to the priority of the detector.
12. An automatic vehicle traffic condition detection device comprising:
a pre-training detector acquisition module that acquires a first number of pre-training detectors for use in various types of vehicle traffic condition detection;
the traffic condition detection module inputs the vehicle sensor data into the pre-training detector to obtain an output value of an output node of the pre-training detector; judging whether the pre-training detector needs combined detection or not based on the output value;
the detector combination module inputs the output value of the pre-training detector to be detected in a combined way into a newly added combination node; the combined node judges whether the current output value is similar to the historical output value or not;
the combined detector determining module determines the detector combination corresponding to the combined node as a new combined detector after a plurality of times of combined node judgment meeting the preset condition;
and the traffic condition combination detection module inputs the sensor data to be detected into the combination detector to obtain a vehicle traffic condition detection result.
13. An electronic device, comprising:
At least one processor, a memory, and a communication interface to communicate with other electronic devices;
the memory stores instructions executable by the at least one processor to enable the electronic device to perform the method of any one of claims 1-11.
14. A non-transitory computer readable storage medium storing computer instructions for causing the computer to perform the method of any one of claims 1-11.
15. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any of claims 1-11.
16. An autonomous vehicle comprising the electronic device of claim 13.
CN202311534676.4A 2023-11-17 2023-11-17 Automatic detection method and related device for vehicle traffic condition Active CN117253365B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311534676.4A CN117253365B (en) 2023-11-17 2023-11-17 Automatic detection method and related device for vehicle traffic condition

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311534676.4A CN117253365B (en) 2023-11-17 2023-11-17 Automatic detection method and related device for vehicle traffic condition

Publications (2)

Publication Number Publication Date
CN117253365A true CN117253365A (en) 2023-12-19
CN117253365B CN117253365B (en) 2024-02-02

Family

ID=89131746

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311534676.4A Active CN117253365B (en) 2023-11-17 2023-11-17 Automatic detection method and related device for vehicle traffic condition

Country Status (1)

Country Link
CN (1) CN117253365B (en)

Citations (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103559508A (en) * 2013-11-05 2014-02-05 福建省视通光电网络有限公司 Video vehicle detection method based on continuous Adaboost
CN105160924A (en) * 2015-08-25 2015-12-16 公安部第三研究所 Video processing-based intelligent signal lamp state detection method and detection system
CN108154118A (en) * 2017-12-25 2018-06-12 北京航空航天大学 A kind of target detection system and method based on adaptive combined filter with multistage detection
US10140855B1 (en) * 2018-08-24 2018-11-27 Iteris, Inc. Enhanced traffic detection by fusing multiple sensor data
US20180349782A1 (en) * 2017-06-06 2018-12-06 PlusAI Corp Method and system for close loop perception in autonomous driving vehicles
CN110111335A (en) * 2019-05-08 2019-08-09 南昌航空大学 A kind of the urban transportation Scene Semantics dividing method and system of adaptive confrontation study
CN110400274A (en) * 2019-07-19 2019-11-01 西安科技大学 A kind of vehicle mounted infrared pedestrian detection infrared image enhancing method
CN111353413A (en) * 2020-02-25 2020-06-30 武汉大学 Low-missing-report-rate defect identification method for power transmission equipment
CN111386559A (en) * 2018-10-25 2020-07-07 北京嘀嘀无限科技发展有限公司 Method and system for judging whether target road facilities exist at intersection
US20200219386A1 (en) * 2019-01-09 2020-07-09 Volkswagen Aktiengesellschaft Method, apparatus, and computer program for determining a plurality of traffic situations
WO2021128365A1 (en) * 2019-12-28 2021-07-01 Beijing Didi Infinity Technology And Development Co., Ltd. Systems and methods for controlling traffic signals
CN113748315A (en) * 2018-12-27 2021-12-03 北京嘀嘀无限科技发展有限公司 System for automatic lane marking
CN114972731A (en) * 2022-05-16 2022-08-30 武汉智行者科技有限公司 Traffic light detection and identification method and device, moving tool and storage medium
CN116721537A (en) * 2023-04-21 2023-09-08 淮阴工学院 Urban short-time traffic flow prediction method based on GCN-IPSO-LSTM combination model
US20230331235A1 (en) * 2022-04-18 2023-10-19 Qualcomm Incorporated Systems and methods of collaborative enhanced sensing

Patent Citations (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103559508A (en) * 2013-11-05 2014-02-05 福建省视通光电网络有限公司 Video vehicle detection method based on continuous Adaboost
CN105160924A (en) * 2015-08-25 2015-12-16 公安部第三研究所 Video processing-based intelligent signal lamp state detection method and detection system
US20180349782A1 (en) * 2017-06-06 2018-12-06 PlusAI Corp Method and system for close loop perception in autonomous driving vehicles
CN108154118A (en) * 2017-12-25 2018-06-12 北京航空航天大学 A kind of target detection system and method based on adaptive combined filter with multistage detection
US10140855B1 (en) * 2018-08-24 2018-11-27 Iteris, Inc. Enhanced traffic detection by fusing multiple sensor data
CN111386559A (en) * 2018-10-25 2020-07-07 北京嘀嘀无限科技发展有限公司 Method and system for judging whether target road facilities exist at intersection
CN113748315A (en) * 2018-12-27 2021-12-03 北京嘀嘀无限科技发展有限公司 System for automatic lane marking
US20200219386A1 (en) * 2019-01-09 2020-07-09 Volkswagen Aktiengesellschaft Method, apparatus, and computer program for determining a plurality of traffic situations
CN110111335A (en) * 2019-05-08 2019-08-09 南昌航空大学 A kind of the urban transportation Scene Semantics dividing method and system of adaptive confrontation study
CN110400274A (en) * 2019-07-19 2019-11-01 西安科技大学 A kind of vehicle mounted infrared pedestrian detection infrared image enhancing method
WO2021128365A1 (en) * 2019-12-28 2021-07-01 Beijing Didi Infinity Technology And Development Co., Ltd. Systems and methods for controlling traffic signals
CN111353413A (en) * 2020-02-25 2020-06-30 武汉大学 Low-missing-report-rate defect identification method for power transmission equipment
US20230331235A1 (en) * 2022-04-18 2023-10-19 Qualcomm Incorporated Systems and methods of collaborative enhanced sensing
CN114972731A (en) * 2022-05-16 2022-08-30 武汉智行者科技有限公司 Traffic light detection and identification method and device, moving tool and storage medium
CN116721537A (en) * 2023-04-21 2023-09-08 淮阴工学院 Urban short-time traffic flow prediction method based on GCN-IPSO-LSTM combination model

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
周亚峰 等: "基于深度学习的光伏图像透视变换算法优化", 新型工业化, vol. 11, no. 2, pages 70 - 72 *

Also Published As

Publication number Publication date
CN117253365B (en) 2024-02-02

Similar Documents

Publication Publication Date Title
US20230245449A1 (en) Terminal device, information processing device, object identifying method, program, and object identifying system
CN111723768B (en) Method, device, equipment and storage medium for vehicle re-identification
CN111797187B (en) Map data updating method and device, electronic equipment and storage medium
CN111220164A (en) Positioning method, device, equipment and storage medium
CN110806215B (en) Vehicle positioning method, device, equipment and storage medium
CN111274343A (en) Vehicle positioning method and device, electronic equipment and storage medium
CN113723141B (en) Vehicle positioning method and device, electronic equipment, vehicle and storage medium
US11288887B2 (en) Object tracking method and apparatus
CN111079079B (en) Data correction method, device, electronic equipment and computer readable storage medium
CN110968718B (en) Target detection model negative sample mining method and device and electronic equipment
CN111275011B (en) Mobile traffic light detection method and device, electronic equipment and storage medium
CN111523471B (en) Method, device, equipment and storage medium for determining lane where vehicle is located
CN111950537B (en) Zebra crossing information acquisition method, map updating method, device and system
CN112802325B (en) Vehicle queuing length detection method and device
CN111767831B (en) Method, apparatus, device and storage medium for processing image
CN113673281B (en) Speed limit information determining method, device, equipment and storage medium
CN113844463B (en) Vehicle control method and device based on automatic driving system and vehicle
CN111667706A (en) Lane-level road surface condition recognition method, road condition prompting method and device
CN117253365B (en) Automatic detection method and related device for vehicle traffic condition
CN112577524A (en) Information correction method and device
CN114429631B (en) Three-dimensional object detection method, device, equipment and storage medium
CN114708565A (en) Intelligent driving scene recognition model creating method, device, equipment and storage medium
CN114677660A (en) Model training and road detection method and device
CN110826448A (en) Indoor positioning method with automatic updating function
CN113587916B (en) Real-time sparse vision odometer, navigation method and 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
GR01 Patent grant
GR01 Patent grant