CN114578344A - Target sensing method, device and system suitable for rainy environment - Google Patents
Target sensing method, device and system suitable for rainy environment Download PDFInfo
- Publication number
- CN114578344A CN114578344A CN202210042864.4A CN202210042864A CN114578344A CN 114578344 A CN114578344 A CN 114578344A CN 202210042864 A CN202210042864 A CN 202210042864A CN 114578344 A CN114578344 A CN 114578344A
- Authority
- CN
- China
- Prior art keywords
- target
- rainfall
- sensing
- range
- setting
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/86—Combinations of radar systems with non-radar systems, e.g. sonar, direction finder
- G01S13/865—Combination of radar systems with lidar systems
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S17/00—Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
- G01S17/86—Combinations of lidar systems with systems other than lidar, radar or sonar, e.g. with direction finders
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- Computer Networks & Wireless Communication (AREA)
- General Physics & Mathematics (AREA)
- Electromagnetism (AREA)
- Traffic Control Systems (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention discloses a target sensing method, a device and a system suitable for a rainy environment, wherein the method comprises the following steps: acquiring the rainfall in a rainy environment; adjusting configuration parameters of the laser radar and the millimeter wave radar according to the rainfall; and respectively acquiring output results of the radars under the corresponding configuration parameters, and fusing the output results to obtain target information. The device comprises a rainfall detection unit, a sensing strategy generation unit, a target sensing unit and a target level fusion unit; the rainfall detection unit acquires a rainfall measurement value and converts the rainfall measurement value into a rainfall grade; the sensing strategy generation unit generates a sensing strategy suitable for the current scene according to the rainfall level; the target sensing unit extracts dynamic and static obstacle targets based on a sensing strategy; the target level fusion unit performs target level fusion based on the target result and outputs target information. The system of the invention corresponds to the method described above. The invention has the advantages of high target sensing precision, simple configuration, wide applicability and the like.
Description
Technical Field
The invention mainly relates to the technical field of unmanned driving, in particular to a target sensing method, device and system suitable for rainy environments.
Background
Environmental awareness is a basic technology for unmanned realization, and the technology is directly related to the safety of automobile driving, so that great attention is always paid to the industry. One mainstream configuration of realizing the unmanned perception function at present is to use laser radar as the core of perception surrounding environment, this is because laser radar generally has the precision height, the field of vision is wide, not influenced by the illumination condition, the better advantage of stability, can be applicable to the operation demand of most scenes. Although the advantages of the laser radar are obvious, the laser has the disadvantages of short wavelength and weak penetrating power, and when the laser meets a medium which has an influence on the propagation of light, such as raindrops, the phenomenon of turning back can occur, so that the raindrops are mistakenly measured as obstacles, false alarm point clouds are brought to a perception algorithm, and the unmanned vehicle cannot smoothly run. In addition, the electromagnetic wave of the millimeter wave radar can penetrate through raindrops and is not interfered by the size of raindrop particles, all-weather work can be realized, but due to the fact that the resolution ratio is too low, the modeling precision is not high, and more noise points exist, the static obstacle target is easy to detect by mistake, and if pedestrians cannot be identified accurately. Therefore, in an environment such as rainy days, the industry generally combines the advantages of the two to use, and makes up for the deficiencies of the two to achieve the autonomous driving target of the unmanned vehicle.
The invention discloses an invention patent application with the publication number of CN110406544A and the invention name of a vehicle perception system and a method under a rain and fog scene, and provides a vehicle perception system and a method under the rain and fog scene. The invention is provided with a laser sensor, a millimeter wave sensor, a binocular sensor, a rainfall sensor, a temperature sensor, a haze sensor, a fog sensor and an illumination sensor. The specific implementation method comprises the following steps: firstly, an illumination sensor is used for carrying out illumination measurement on the environment so as to detect whether the environment is in the daytime or not; judging whether the snowing days exist or not through the measurement values of the temperature sensors, and if the snowing days exist, directly entering obstacle detection; otherwise, continuously checking whether the rainfall sensor is in rainy days, and entering obstacle detection if the rainfall sensor is in rainy days; otherwise, checking whether the measurement condition of the fog quantity sensor is in fog days, and entering obstacle detection if the measurement condition of the fog quantity sensor is in fog days; if the measurement condition of the haze-dust sensor is the haze-dust weather, directly entering obstacle detection if the measurement condition of the haze-dust sensor is the haze-dust weather; and if the weather is the non-haze weather, the obstacle detection is also carried out. The obstacle detection implementation steps are as follows: detecting obstacles around the vehicle according to the laser sensor and the millimeter wave sensor, calculating confidence coefficient formed by the laser radar, the millimeter wave and the vision sensor if the obstacles exist, and then giving information of the size, the direction and the running speed of the obstacles according to the confidence coefficient. And finally, deciding the running speed and the advancing route of the vehicle according to the conditions of different types of sensors respectively, and conducting the decision to an execution unit. In the whole process, the invention is completely provided with various types of sensors to measure different climate environment information, and the obtained measurement result is used for restricting the traveling speed of the vehicle, but does not act on the level of perception capability improvement. Therefore, the invention does not improve the perception result and only brings auxiliary effect.
The invention discloses an invention patent application with the publication number of CN112147615A and the invention name of an unmanned perception method based on an all-weather environment monitoring system, and provides a perception scheme with a weather monitoring system. The weather monitoring system of the scheme is composed of a temperature sensor, a rainfall sensor, a light sensor and a dust sensor; and the environment perception system comprises a millimeter wave radar, a laser radar, a camera, an ultrasonic radar and a vision sensor. The main operation of the invention is as follows: and selecting and deciding the sensor information according to different results of weather monitoring in a decision module. The invention only selects the results of different sensors according to different weather conditions to carry out fusion output, but even in the same weather, such as light rain, medium rain, heavy rain or heavy rain, the influenced condition of the main sensing equipment is different, so that the scheme lacks of quantification operation on the influenced condition of the same weather.
In conclusion, how to combine the results of the laser radar and the millimeter wave radar to perform accurate target identification is a difficult problem in the sensing field.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: aiming at the problems in the prior art, the invention provides a target sensing method, a device and a system with high target sensing precision, which are suitable for a rainy environment.
In order to solve the technical problems, the technical scheme provided by the invention is as follows:
a method for sensing a target in a rainy environment, comprising:
acquiring the rainfall in a rainy environment;
adjusting configuration parameters of the laser radar and the millimeter wave radar according to the rainfall; the configuration parameter comprises a perception range;
and respectively acquiring output results of the laser radar and the millimeter wave radar under corresponding configuration parameters, and fusing the output results to obtain target information.
As a further improvement of the above technical solution:
after the rainfall in the rainy environment is obtained, the rainfall is converted into the corresponding rainfall level, and the configuration parameters of the laser radar and the millimeter wave radar are adjusted according to the rainfall level.
The rainfall grade is divided into five grades, and the five grades are no rain, light rain, medium rain, heavy rain and extra heavy rainstorm in sequence from small to large.
The configuration parameters of the laser radar also comprise filtering parameters for processing the point cloud target; the configuration parameters of the millimeter wave radar further comprise one or more of a maximum effective speed, a minimum effective speed, or an effective target class of the target.
The specific process of adjusting the configuration parameters of the laser radar and the millimeter wave radar according to the rainfall level comprises the following steps:
when the rainfall level is no rain, setting a target sensing range of the millimeter wave radar as a standard FOV range, setting the minimum effective speed and the maximum effective speed as standard measuring ranges, and setting the effective target category as a full category; the effective configuration range of the laser radar point cloud is not set (the range which can be measured by the point cloud), and the obstacle checking switch is set to be closed;
when the rainfall level is light rain, setting the target sensing range of the millimeter wave radar as a standard FOV range, setting the minimum effective speed and the maximum effective speed as standard measuring ranges, and setting the effective target category as a full category; setting the effective configuration range of the laser radar point cloud, and setting a switch to start a laser point cloud filtering function to remove false alarms caused by returning isolated raindrops from the statistical characteristics of the laser point cloud;
when the rainfall level is medium rain, setting a target sensing range of the millimeter wave radar as a standard FOV range, setting the minimum effective speed and the maximum effective speed as standard measuring ranges, and setting the effective target category as a full category; setting a secondary check range for a target of the laser point cloud without setting the effective configuration range of the laser radar point cloud;
when the rainfall level is heavy rain, setting the target sensing range of the millimeter wave radar as a standard FOV range, setting the minimum effective speed and the maximum effective speed as standard measuring ranges, and limiting the types of effective targets to vehicles and pedestrians; the detection range of the laser radar point cloud is configured to be within a1 m in front of a vehicle and b1 m on the left and right of the vehicle, and secondary verification ranges are set for all targets of the laser point cloud;
when the rainfall level is extra-heavy rainstorm, setting the sensing range of the millimeter wave target as a standard FOV range, setting the minimum effective speed and the maximum effective speed as standard measuring ranges, and limiting the types of effective targets to vehicles and pedestrians; the detection range of the laser radar point cloud is configured to be a2 m in front of the vehicle and b2 m left and right, and secondary verification ranges are set for all targets of the laser point cloud.
Before the rainfall is converted into the corresponding rainfall level, the historical numerical value sequence is cached, the abnormal numerical value is eliminated by using smooth filtering, and then the state is kept and output; the abnormal numerical value elimination is to perform operation on the input rainfall level numerical sequence by using a sliding window, and replace the rainfall number with the largest occurrence frequency with the numerical value of the target position; and the state keeping output is to continuously output the grade number of the current rainfall for at least a period of time and then decide whether to change the grade number of the output rainfall.
The invention also discloses a target sensing device suitable for rainy environments, which comprises a rainfall detection unit, a sensing strategy generation unit, a target sensing unit and a target level fusion unit;
the rainfall detection unit acquires a rainfall measurement value and converts the rainfall measurement value into a rainfall grade;
the sensing strategy generation unit adjusts the configuration parameters of the laser radar and the millimeter wave radar according to the rainfall level to generate a sensing strategy suitable for the current scene;
the target sensing unit extracts dynamic barrier and static barrier targets from real-time sensing data based on the generated sensing strategy;
and the target-level fusion unit performs target-level fusion based on the target result and outputs target information to perform path navigation.
As a further improvement of the above technical solution:
the rainfall detection unit includes a rainfall sensor.
The target sensing unit comprises a millimeter wave radar and a laser radar.
The invention further discloses a target sensing system suitable for rainy environment, which comprises:
the first program module is used for acquiring rainfall in a rainy day environment;
the second program module is used for adjusting configuration parameters of the laser radar and the millimeter wave radar according to the rainfall; the configuration parameter comprises a perception range;
and the third program module is used for respectively acquiring output results of the laser radar and the millimeter wave radar under corresponding configuration parameters, and fusing the output results to obtain target information.
Compared with the prior art, the invention has the advantages that:
the invention detects the rainfall in the rainy day environment, and then adjusts the configuration parameters of the laser radar and the millimeter wave radar according to the rainfall, thereby generating a sensing strategy corresponding to the rainfall, realizing the dynamic adjustment of the sensing parameters of each radar, keeping the most reliable part of the result, effectively overcoming the problems of too many false alarm obstacle points of the laser radar, unstable sensing result and low reliability in the rainy day environment, leading the whole sensing result to still have the same reliability as the ideal weather without entering the false alarm information, and providing technical guarantee for the unmanned automobile to realize the safe driving in the rainy day environment.
Drawings
FIG. 1 is a flow chart of an embodiment of the method of the present invention.
FIG. 2 is a block diagram of an embodiment of the system of the present invention.
FIG. 3 is a schematic diagram of the lidar point cloud under no rain (level 0) conditions in accordance with the present invention.
FIG. 4 is a schematic diagram of the lidar point cloud under light rain (level 1) conditions of the present invention.
Fig. 5 is a schematic diagram of the lidar point cloud under rain (level 2) conditions in the present invention.
FIG. 6 is a schematic diagram of the lidar point cloud under heavy rain (level 3) conditions of the present invention.
FIG. 7 is a schematic diagram of the lidar point cloud for a very heavy storm (level 4) condition of the invention.
Detailed Description
The invention is further described below with reference to the figures and the specific embodiments of the description.
As shown in fig. 1, a target sensing method suitable for use in a rainy environment according to an embodiment of the present invention includes:
acquiring the rainfall in a rainy environment;
adjusting configuration parameters of the laser radar and the millimeter wave radar according to the rainfall, and generating a sensing strategy corresponding to the current rainy environment; the configuration parameter comprises a perception range;
and respectively acquiring output results of the laser radar and the millimeter wave radar under corresponding configuration parameters, and fusing the output results to obtain target information.
In a rainy environment, when the rainfall of the laser radar reaches a certain level, more raindrop false alarms exist in the air at a far place, and the imaging is good at a near place; the millimeter waves are less interfered by raindrops, so that the sensing result of the laser radar can be compensated by the sensing result of the millimeter waves in a long distance under the scene that the rainfall exceeds a certain threshold value; the method of keeping the results of the two types of heterogeneous results in a close range fuses the two types of heterogeneous results and outputs an accurate perception target; and for the rainfall in the range that the two sensors are not influenced, the results of the two sensors are still adopted for fusion.
Based on the facts, the invention detects the rainfall in the rainy day environment, and then adjusts the configuration parameters of the laser radar and the millimeter wave radar according to the rainfall, so as to generate the sensing strategy corresponding to the rainfall, realize the dynamic adjustment of the sensing parameters of each radar, ensure that the result keeps the part with the highest reliability, effectively overcome the problems of excessive false alarm obstacle points of the laser radar, unstable sensing result and low reliability in the rainy day environment, ensure that the whole sensing result still has the same reliability as the ideal weather, does not carry false alarm information, and provide technical support for the unmanned automobile to realize the safe driving in the rainy day environment.
In a specific embodiment, the configuration parameters of the laser radar specifically include a sensing range and a filtering parameter for processing a point cloud target; configuration parameters of the millimeter wave radar include a sensing range, a maximum effective speed of the target, a minimum effective speed, and an effective target class.
In a specific embodiment, the configuration parameters of the laser radar and the millimeter wave radar can be adjusted in real time through rainfall, or after the rainfall in a rainy environment is obtained, the rainfall is converted into corresponding rainfall levels, and then the configuration parameters of the laser radar and the millimeter wave radar are adjusted according to the rainfall levels. Specifically, the rainfall level is divided into five levels, and from small to large, no rain, light rain, medium rain, heavy rain and extra heavy rain are sequentially performed (of course, in other embodiments, three, four, six or more levels may be performed according to actual conditions). If the rainfall sensing result is obtained, the rainfall grade value of 0-4 is output outwards and corresponds to no rain (corresponding to the value of 0), light rain (corresponding to the value of 1), medium rain (corresponding to the value of 2), heavy rain (corresponding to the value of 3) and extra heavy rain (corresponding to the value of 4). The radar point cloud data for these five cases are shown in fig. 3-7. The cuboid is a vehicle body boundary, the radar is installed in front of the vehicle body, and the circle is false alarm point cloud (raindrop).
Through analysis, the following results are obtained: the laser radar under the condition of no rain is relatively accurate in measurement, false alarm point clouds do not exist, and the available range of the point clouds is the whole measurement domain; under the condition of light rain, the laser radar still has certain anti-interference capability, the point cloud imaging is good, extremely individual false alarm point clouds exist, and the false alarm can be directly removed through a filtering algorithm, so that the credible range of the point clouds is all the measured point clouds, and a millimeter wave radar is not required to assist in identifying a target; under the condition of medium rain, a small amount of point cloud false alarms exist in the front, the credible range of the point cloud is still all measured point clouds, but millimeter wave radar is used for assisting to identify true and false targets; under the heavy rain condition, the credible range of the point cloud is limited (the distance is about within 4 meters), and a millimeter wave radar is used for assisting to identify true and false targets; and finally, under the condition of extra heavy rainstorm, the credible range of the point cloud is extremely limited (the distance is within about 2 meters), and a millimeter wave radar target is required to assist in identifying the true and false target.
As shown in fig. 2, an embodiment of the present invention further discloses a target sensing device suitable for use in a rainy environment, including a rainfall detection unit, a sensing policy generation unit, a target sensing unit, and a target level fusion unit;
the rainfall detection unit acquires a rainfall measurement value and converts the rainfall measurement value into a rainfall grade;
the sensing strategy generation unit adjusts the configuration parameters of the laser radar and the millimeter wave radar according to the rainfall level to generate a sensing strategy suitable for the current scene;
the target sensing unit extracts dynamic obstacle and static obstacle targets from real-time sensing data based on the generated sensing strategy;
and the target level fusion unit performs target level fusion based on the target result and outputs target information to perform path navigation.
The target sensing device suitable for the rainy environment is used for executing the target sensing method, and has the advantages of the target sensing method.
In a specific embodiment, the rainfall detection unit comprises at least one group of rainfall sensors which are arranged on the vehicle body; the target sensing unit comprises at least one group of laser radar and millimeter wave radar and is installed on the vehicle body. The sensor of the system only comprises the laser radar, the millimeter wave radar and the rainfall sensor, and is simple in configuration, wide in applicability and very suitable for practical application and popularization.
The invention also discloses a target sensing system suitable for rainy environment, which comprises:
the first program module is used for acquiring rainfall in a rainy day environment;
the second program module is used for adjusting configuration parameters of the laser radar and the millimeter wave radar according to the rainfall; the configuration parameter comprises a perception range;
and the third program module is used for respectively acquiring output results of the laser radar and the millimeter wave radar under corresponding configuration parameters, and fusing the output results to obtain target information.
The object sensing system of the present invention, corresponding to the above object sensing method, also has the advantages as described above for the object sensing method.
The invention will be further described with reference to a full embodiment:
under the condition of normal weather (sunny days), the measurement point cloud of the object by the laser radar does not have a false alarm, namely, a corresponding obstacle usually exists near the measurement value position of the laser radar. However, when a medium such as raindrops in the air changes the optical path, a part of the optical path is refracted and returned, so that an echo response is received at a receiving end, and a receiver is deceived to encounter the surface of an obstacle, but the obstacle does not exist actually, so that the point cloud is called as a false alarm point cloud. Tests show that the occurrence degree of false-alarm point cloud is in direct proportion to the size and distance of raindrop particles: the larger the raindrop particles are, the more easily false-alarm point cloud occurs; the more likely a false alarm point cloud occurs, the distance is near the location of the radar optical center. In normal weather or rainy days, the measurement of the laser radar at the near position is accurate, so that the point cloud of the laser radar at the near position does not need to be processed too much. Therefore, in rainy days, the point cloud of the laser radar is screened from the two aspects to obtain more accurate point cloud, the point cloud which is greatly influenced is discarded, and the millimeter wave radar is adopted for making up (the millimeter wave radar measures more accurately at a distance and just can be matched with the defect of the laser radar).
Under the actual operation condition, the vehicle is when the rainy day environment, and is difficult accurate control to the raindrop size of perception, so adopt the rainfall sensor to infer the raindrop size through estimating the rainfall indirectly, combine the experience can instruct the radar can the normal use under the current raindrop influence circumstances cloud scope, specifically can summarize:
when the usable range of the laser radar completely covers the millimeter wave range, setting the range of the millimeter wave radar as a far distance, and fusing the results of the two;
and when the available range of the laser radar is near the vehicle body, setting the sensing range of the millimeter wave radar as an area uncovered by the laser radar, and fusing the results of the two areas.
When the scheme is applied, the rainfall data of the rainfall sensor is processed in real time, filtering processing is carried out, abnormal values are discarded, and rainfall values are converted into five grades: no rain, light rain, medium rain, heavy rain and extra heavy rainstorm. Specifically, the rainfall sensor is configured to output values of 0 to 4 corresponding to no rain, light rain, medium rain, heavy rain, and heavy rain, respectively, based on the sensed rainfall energy. And caching the historical numerical value sequence before the rainfall level is output, eliminating abnormal numerical values by using a smooth filtering processing technology, and then carrying out state keeping output. The abnormal value eliminating algorithm is to operate the input rainfall level number sequence by using a sliding window and replace the rainfall number with the largest occurrence frequency with the value of the target position. The state holding operation is to continuously output the level number of the current rainfall for at least a certain period of time (for example, 20 seconds) before deciding whether to change the output rainfall level number.
When the rainfall sensor outputs a reading of 0, the current rainfall is in a no-rain state. Setting the sensing range of the millimeter wave target as a standard FOV range, setting the minimum maximum effective speed as a standard measurement range, and setting the effective target category as a full category; the effective configuration range of the laser radar point cloud is not set (the range which can be measured by the point cloud), and the obstacle checking switch is set to be closed. The target result of the millimeter wave is directly fused with the laser detection result at the fusion end, and the fusion target can be output outwards.
When the rainfall sensor outputs a reading 1 to the outside, the current rainfall is in a light rain state. Setting the sensing range of the millimeter wave target as a standard FOV range, setting the minimum maximum effective speed as a standard measurement range, and setting the effective target category as a full category; the effective configuration range of the laser radar point cloud is not set (the default is the range which can be measured by the point cloud), and a switch is set to start the laser point cloud filtering function so as to remove false alarms caused by returning of isolated raindrops from the statistical characteristics of the laser point cloud. And fusing the millimeter wave target at the fusion end with the target with the false alarm eliminated by using the statistical characteristics of the laser point cloud, and outputting the fused target externally.
When the rainfall sensor outputs a reading 2, the current rainfall is in a medium rain state. Setting the sensing range of the millimeter wave target as a standard FOV range, setting the minimum maximum effective speed as a standard measurement range, and setting the effective target category as a full category; the effective configuration range of the laser radar point cloud is not set (the range which can be measured by the point cloud), but a secondary check range is set for the target of the laser point cloud: and carrying out secondary inspection on the isolated target within the range of 10 meters in front of the laser radar and 2 meters left and right by using a target result of the millimeter wave to discriminate whether the target detected by the laser detection module is a false target, and removing the false target (the specific process is that if the isolated target within the range of 10 meters in front of the laser radar and 2 meters left and right does not have a millimeter wave target within 1 meter and the number of point clouds is less than 5, the isolated target is indicated as raindrop, and the target is removed). And fusing the millimeter wave target and the laser target after the raindrop target is removed at the fusion end, and outputting the fused target outwards.
When the rainfall sensor outputs a reading 3 to the outside, the current rainfall is in a heavy rain state. Setting the sensing range of the millimeter wave target as a standard FOV range, setting the minimum maximum effective speed as a standard measuring range, and limiting the class of effective targets to vehicles and pedestrians; configuring the detection range of the laser radar point cloud into 4 meters in front of the vehicle and 2 meters at left and right, and setting secondary check ranges for all targets of the laser point cloud: and judging whether an effective millimeter wave target exists nearby a target detected by the laser radar in the configuration range, if so, indicating that the target is a real target, and otherwise, indicating that the target is a false alarm. And the fusion end fuses all effective targets of the millimeter waves, fuses radar obstacle targets near the limited vehicle body, and outputs the fused targets of the two sensors to the outside.
When the rainfall sensor outputs a reading 4 to the outside, the current rainfall is in a rainstorm state, and the laser point cloud is influenced by very large raindrops. Setting the sensing range of the millimeter wave target as a standard FOV range, setting the minimum maximum effective speed as a standard measuring range, and limiting the class of effective targets to vehicles and pedestrians; configuring the detection range of the laser radar point cloud into 2 meters in front of the vehicle and 2 meters in left and right, and setting secondary check ranges for all targets of the laser point cloud: and judging whether an effective millimeter wave target exists nearby a target detected by the laser radar in the configuration range, if so, indicating that the target is a real target, and otherwise, indicating that the target is a false alarm. And the fusion end fuses all effective targets of the millimeter waves, fuses radar obstacle targets near the limited vehicle body, and outputs the fused targets of the two sensors to the outside.
And through the configuration adjustment of the laser radar and the millimeter wave radar, a sensing strategy corresponding to the rainfall level is generated, and dynamic obstacle and static obstacle targets are respectively extracted from the real-time sensing data of the millimeter wave radar and the laser radar. The method specifically comprises the following steps: receiving original sensing data from the millimeter waves and the laser radar respectively, and dynamically adjusting parameters for extracting targets according to corresponding configuration parameter values to obtain two kinds of heterogeneous target information;
and performing target-level fusion in the fusion system based on the target result extracted by the multi-source sensor, and sending the fusion result to an upper-layer application for path navigation.
Specifically, a fusion algorithm is operated from two heterogeneous target results, the same kind of target is merged, and target motion information is supplemented. Wherein the fusion process is as follows: the laser radar extracts static target information (static obstacle) according to the allocated configuration parameters; the static obstacle superposes the multi-frame detection result to carry out motion tracking, so that the dyskinesia information (dyskinesia) can be further obtained; the millimeter wave radar extracts real-time moving target information according to the distributed configuration parameters; and performing position association on the movement obstacle of the laser radar and the movement target information extracted by the millimeter wave radar, filling the movement information of the millimeter wave radar to obtain a relatively perfect movement target, and outputting the perfect movement target and the static target to an upper-layer plan for path planning.
The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above-mentioned embodiments, and all technical solutions belonging to the idea of the present invention belong to the protection scope of the present invention. It should be noted that modifications and embellishments within the scope of the invention may be made by those skilled in the art without departing from the principle of the invention.
Claims (10)
1. A target perception method suitable for a rainy environment, comprising:
acquiring the rainfall in a rainy environment;
adjusting configuration parameters of the laser radar and the millimeter wave radar according to the rainfall; the configuration parameter comprises a perception range;
and respectively acquiring output results of the laser radar and the millimeter wave radar under corresponding configuration parameters, and fusing the output results to obtain target information.
2. The method for sensing the target in the rainy environment according to claim 1, wherein after the rainfall in the rainy environment is obtained, the rainfall is converted into a corresponding rainfall level, and the configuration parameters of the laser radar and the millimeter wave radar are adjusted according to the rainfall level.
3. The method as claimed in claim 2, wherein the rain amount is divided into five levels, from small to large, no rain, light rain, medium rain, heavy rain, and heavy rainstorm.
4. The method as claimed in claim 3, wherein the configuration parameters of the lidar further comprise filtering parameters for processing point cloud targets; the configuration parameters of the millimeter wave radar further comprise one or more of a maximum effective speed, a minimum effective speed, or an effective target class of the target.
5. The method for sensing the target in the rainy environment according to claim 4, wherein the specific process of adjusting the configuration parameters of the lidar and the millimeter wave radar according to the rainfall level comprises:
when the rainfall level is no rain, setting the target sensing range of the millimeter wave radar as a standard FOV range, setting the minimum effective speed and the maximum effective speed as standard measuring ranges, and setting the effective target category as a full category; the effective configuration range of the laser radar point cloud is not set, and an obstacle check switch is set to be closed;
when the rainfall level is light rain, setting the target sensing range of the millimeter wave radar as a standard FOV range, setting the minimum effective speed and the maximum effective speed as standard measuring ranges, and setting the effective target category as a full category; setting the effective configuration range of the laser radar point cloud, and setting a switch to start a laser point cloud filtering function to remove false alarms caused by returning isolated raindrops from the statistical characteristics of the laser point cloud;
when the rainfall level is medium rain, setting the target sensing range of the millimeter wave radar as a standard FOV range, setting the minimum effective speed and the maximum effective speed as standard measurement ranges, and setting the effective target category as a full category; setting a secondary check range for a target of the laser point cloud without setting the effective configuration range of the laser radar point cloud;
when the rainfall level is heavy rain, setting the target sensing range of the millimeter wave radar as a standard FOV range, setting the minimum effective speed and the maximum effective speed as standard measuring ranges, and limiting the types of effective targets to vehicles and pedestrians; configuring the detection range of the laser radar point cloud to be within a1 m in front of the vehicle and b1 m left and right of the vehicle, and setting secondary check ranges for all targets of the laser point cloud;
when the rainfall level is extra-heavy rainstorm, setting the sensing range of the millimeter wave target as a standard FOV range, setting the minimum effective speed and the maximum effective speed as standard measuring ranges, and limiting the types of effective targets to vehicles and pedestrians; the detection range of the laser radar point cloud is configured to be within a2 m in front of a vehicle and b2 m on the left and right of the vehicle, and secondary verification ranges are set for all targets of the laser point cloud.
6. The method for sensing the target in the rainy environment according to any one of claims 3 to 5, wherein before the rainfall is converted into the corresponding rainfall level, the historical numerical value sequence is buffered, and an abnormal numerical value is removed by using smooth filtering, and then state keeping output is performed; the abnormal numerical value elimination is to perform operation on the input rainfall level numerical sequence by using a sliding window, and replace the rainfall number with the largest occurrence frequency with the numerical value of the target position; and the state keeping output is to continuously output the grade number of the current rainfall for at least a period of time and then decide whether to change the grade number of the output rainfall.
7. A target sensing device suitable for rainy environments is characterized by comprising a rainfall detection unit, a sensing strategy generation unit, a target sensing unit and a target level fusion unit;
the rainfall detection unit acquires a rainfall measurement value and converts the rainfall measurement value into a rainfall grade;
the sensing strategy generation unit adjusts the configuration parameters of the laser radar and the millimeter wave radar according to the rainfall level to generate a sensing strategy suitable for the current scene;
the target sensing unit extracts dynamic barrier and static barrier targets from real-time sensing data based on the generated sensing strategy;
and the target-level fusion unit performs target-level fusion based on the target result and outputs target information to perform path navigation.
8. The apparatus as claimed in claim 7, wherein the rainfall detection unit comprises a rainfall sensor.
9. The apparatus as claimed in claim 7 or 8, wherein the target sensing unit comprises millimeter wave radar and laser radar.
10. A target perception system adapted for use in a rainy day environment, comprising:
the first program module is used for acquiring rainfall in a rainy day environment;
the second program module is used for adjusting configuration parameters of the laser radar and the millimeter wave radar according to the rainfall; the configuration parameter comprises a perception range;
and the third program module is used for respectively acquiring output results of the laser radar and the millimeter wave radar under corresponding configuration parameters, and fusing the output results to obtain target information.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210042864.4A CN114578344A (en) | 2022-01-14 | 2022-01-14 | Target sensing method, device and system suitable for rainy environment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210042864.4A CN114578344A (en) | 2022-01-14 | 2022-01-14 | Target sensing method, device and system suitable for rainy environment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114578344A true CN114578344A (en) | 2022-06-03 |
Family
ID=81769692
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210042864.4A Pending CN114578344A (en) | 2022-01-14 | 2022-01-14 | Target sensing method, device and system suitable for rainy environment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114578344A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115825982A (en) * | 2023-02-02 | 2023-03-21 | 深圳煜炜光学科技有限公司 | Method and system for scanning point cloud data of unmanned aerial vehicle in rainy environment |
| CN116884250A (en) * | 2023-07-12 | 2023-10-13 | 凉山州交通运输应急指挥中心 | Early warning method based on laser radar and expressway early warning system |
| WO2024045178A1 (en) * | 2022-09-02 | 2024-03-07 | 华为技术有限公司 | Sensing method, apparatus, and system |
-
2022
- 2022-01-14 CN CN202210042864.4A patent/CN114578344A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2024045178A1 (en) * | 2022-09-02 | 2024-03-07 | 华为技术有限公司 | Sensing method, apparatus, and system |
| CN115825982A (en) * | 2023-02-02 | 2023-03-21 | 深圳煜炜光学科技有限公司 | Method and system for scanning point cloud data of unmanned aerial vehicle in rainy environment |
| CN116884250A (en) * | 2023-07-12 | 2023-10-13 | 凉山州交通运输应急指挥中心 | Early warning method based on laser radar and expressway early warning system |
| CN116884250B (en) * | 2023-07-12 | 2024-01-26 | 凉山州交通运输应急指挥中心 | Early warning method based on laser radar and expressway early warning system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114578344A (en) | Target sensing method, device and system suitable for rainy environment | |
| US5351044A (en) | Vehicle lane position detection system | |
| Dannheim et al. | Weather detection in vehicles by means of camera and LIDAR systems | |
| CN112693466A (en) | System and method for evaluating performance of vehicle environment perception sensor | |
| CN112562405A (en) | Radar video intelligent fusion and early warning method and system | |
| CN110406544A (en) | Vehicle sensory perceptual system and method under misty rain scene | |
| CN110356339B (en) | Lane change blind area monitoring method and system and vehicle | |
| CN111880196A (en) | Unmanned mine car anti-interference method, system and computer equipment | |
| CN111016903B (en) | System and method for guaranteeing driving safety of automatic driving vehicle in rainy days | |
| US11292481B2 (en) | Method and apparatus for multi vehicle sensor suite diagnosis | |
| CN112101316B (en) | Target detection method and system | |
| CN108263389A (en) | A kind of vehicle front false target device for eliminating and method | |
| CN111582130A (en) | Traffic behavior perception fusion system and method based on multi-source heterogeneous information | |
| US11423661B2 (en) | Object recognition apparatus | |
| CN115034324A (en) | Multi-sensor fusion perception efficiency enhancement method | |
| CN113741388A (en) | Safety diagnosis system and method based on automatic driving perception failure | |
| US20200110419A1 (en) | Method of processing data, apparatus for processing data, and system for controlling vehicle | |
| CN112801024B (en) | Detection information processing method and device | |
| CN114296095A (en) | Method, device, vehicle and medium for extracting effective target of automatic driving vehicle | |
| CN114084129A (en) | Fusion-based vehicle automatic driving control method and system | |
| CN117173666A (en) | Automatic driving target identification method and system for unstructured road | |
| CN111959515A (en) | Forward target selection method, device and system based on visual detection | |
| CN116935551A (en) | Perimeter intrusion detection method, system, equipment and storage medium | |
| CN112505704A (en) | Method for improving safety of train autonomous intelligent sensing system and train | |
| CN114675295A (en) | Method, device and equipment for judging obstacle and storage medium |
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 |