CN116626668A - Target detection method, system and storage medium - Google Patents
Target detection method, system and storage medium Download PDFInfo
- Publication number
- CN116626668A CN116626668A CN202310633537.0A CN202310633537A CN116626668A CN 116626668 A CN116626668 A CN 116626668A CN 202310633537 A CN202310633537 A CN 202310633537A CN 116626668 A CN116626668 A CN 116626668A
- Authority
- CN
- China
- Prior art keywords
- target
- particle
- cloud data
- pedestrian
- trackid
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 26
- 239000002245 particle Substances 0.000 claims abstract description 76
- 238000001914 filtration Methods 0.000 claims abstract description 36
- 238000007781 pre-processing Methods 0.000 claims abstract description 13
- 238000000034 method Methods 0.000 claims description 11
- 238000012795 verification Methods 0.000 claims description 4
- 230000000007 visual effect Effects 0.000 abstract description 5
- 230000009191 jumping Effects 0.000 abstract description 4
- 102100034112 Alkyldihydroxyacetonephosphate synthase, peroxisomal Human genes 0.000 description 3
- 101000799143 Homo sapiens Alkyldihydroxyacetonephosphate synthase, peroxisomal Proteins 0.000 description 3
- 238000000848 angular dependent Auger electron spectroscopy Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000003491 array Methods 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 239000012535 impurity Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000005303 weighing Methods 0.000 description 2
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
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
- G01S13/66—Radar-tracking systems; Analogous systems
- G01S13/72—Radar-tracking systems; Analogous systems for two-dimensional tracking, e.g. combination of angle and range tracking, track-while-scan radar
- G01S13/723—Radar-tracking systems; Analogous systems for two-dimensional tracking, e.g. combination of angle and range tracking, track-while-scan radar by using numerical data
-
- 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
- G01S13/88—Radar or analogous systems specially adapted for specific applications
-
- 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
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
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)
- Radar Systems Or Details Thereof (AREA)
Abstract
The application discloses a target detection method, a target detection system and a storage medium, wherein the target detection method comprises the following steps: preprocessing millimeter wave Lei Dadian cloud data to obtain point cloud data of a pedestrian target; setting an initial value of a particle filtering algorithm; initializing a dirac particle array in a particle filtering algorithm by using radar precision, following frame number and serial number; predicting a pedestrian target and outputting a predicted value; performing target matching by using particle filtering, updating particle weight and normalizing; updating the target pool output target. The application solves the problem of jumping of the visual target under the condition of insufficient light and the problem of false recognition or missing recognition caused by excessive clutter of the millimeter wave radar in a relatively closed space, and can obviously improve the recognition precision of the pedestrian target, thereby improving the driving safety.
Description
Technical Field
The present application relates to the field of target detection technologies, and in particular, to a target detection method, system, and storage medium.
Background
Intelligent three-electricity (battery, motor, electric control), intelligent cabin, and automatic driving have represented the trend of future automobiles. Unmanned software systems are generally defined as six modules, sensing, prediction, high-precision mapping, positioning, decision planning, control. The sensing module generally refers to obstacle and traffic light recognition, and the sensors mainly comprise a laser radar, a millimeter wave radar and a camera. These three sensors have different advantages in different environments, respectively.
The camera has low cost, has the advantages of target detection and classification, is widely applied to ADAS automatic driving systems, but is sensitive to light variation, depends on a deep learning network model and a training data set, and has the possibility of failure. The laser radar is used as an active sensor, has the advantages of high precision, strong three-dimensional sensing capability and the like, but has high cost and is easily influenced by weather. The millimeter wave radar has relatively low cost, has three-dimensional point cloud and speed sensing capability, can work all-weather at the same time, and has high reliability, so that the millimeter wave radar is widely applied to ADAS automatic driving systems of various grades.
The ADAS autopilot system has higher requirements on the perception system, when the pedestrian target is tracked in the underground garage, the autopilot system needs to accurately perceive the position and track of the pedestrian, so that a decision module and a planning module can conveniently calculate a reasonable decision and path, and the possible safety risk is avoided. Because the light is dim, the visual pedestrian target detection is easy to generate the phenomena of false detection, false identification or large-amplitude jumping and the like, and the tracking matching of the visual sensor to the target of the person in the garage is not satisfactory. For another example, the Chinese patent application No. CN202010995314.5 discloses a short-distance personnel target tracking method based on millimeter wave radar, and the kalman millimeter wave radar target processing method used in the scheme can be used for performing conventional tracking on the target, so that the problem of jumping of the visual target can be effectively solved, but the linear tracking method is poor in effect due to instability of the moving track of the pedestrian target.
Disclosure of Invention
The present application is directed to overcoming the problems in the prior art, and providing a target detection method, system and storage medium.
The primary purpose of the application is to solve the technical problems, and the technical scheme of the application is as follows:
the first aspect of the present application provides a target detection method, comprising the steps of:
preprocessing millimeter wave Lei Dadian cloud data to obtain point cloud data of a pedestrian target;
setting an initial value of a particle filtering algorithm;
initializing a dirac particle array in a particle filtering algorithm by using radar precision, following frame number and serial number;
predicting a pedestrian target and outputting a predicted value;
performing target matching by using particle filtering, updating particle weight and normalizing;
updating the target pool output target.
Further, the preprocessing millimeter wave Lei Dadian cloud data specifically includes:
filtering the first frame of millimeter wave Lei Dadian cloud data by utilizing a preset first threshold, and eliminating false identification points;
calculating the position and the speed of each point in the point cloud data;
and clustering, RCS man-vehicle distinguishing and second threshold filtering are sequentially carried out on the point cloud data according to the positions and the speeds of the points to obtain the point cloud data of the pedestrian target.
Further, the setting of the initial value of the particle filter algorithm specifically includes: and the tracking frame number of the pedestrian target is recorded as TrackFrame, trackFrame and is set as a set value, the number is recorded as a TrackID, and the pedestrian target is stored in a target pool TargetPool to be used as an initial value of a particle filtering algorithm.
Further, initializing a dirac particle array in the particle filter algorithm by using the radar precision, the following frame number and the serial number, specifically includes:
constructing a two-dimensional array with the number m of lines and the number n of columns as preset values for the numbered TrackID of each tracking frame number;
the first item of the two-dimensional array is TrackID, trackFrame, the distance of each grid in the two-dimensional array is recorded as Dmdn, and the value of Dmdn is the distance precision of the radar.
Further, when the trackID is 1, the initial value of the particle weight is 0.01.
Further, the pedestrian target is predicted and a predicted value is output, specifically:
the state equations of the points in the point cloud in the X direction and the Y direction are respectively as follows:
X k+1 frame interval time (frametime) +x k
Y k+1 Frame interval time (frametime) +y k
Where k is the number of iterations, where the value of k is equal to the tracking frame number TrackFrame, and the updated target is noted as PredictTarget, trackID and other attributes along with the initial pedestrian target.
Further, the particle filtering is utilized to carry out target matching, the particle weight is updated and normalized, and the method specifically comprises the following steps:
taking a pedestrian target obtained by preprocessing the point cloud data as an observation value of a new round;
calculating Euclidean distance between the new target corresponding to the new round of observation value and each target in the target pool, taking the target in the target pool corresponding to the minimum Euclidean distance as a first new target,
the first new target passes through an energy threshold to be subjected to matching verification to obtain a second new target;
and modifying the particle weight of the two-dimensional array in the target TrackID by using a second new target, and normalizing the modified weight.
Further, updating the target pool output target specifically includes:
and updating the TrackID target position by using the updated particle weight, and outputting the target when the tracking frame number TrackFrame is larger than the set threshold TrackFrameThresh.
In a second aspect, the present application provides an object detection system comprising: the device comprises a memory and a processor, wherein the memory comprises an object detection method program, and the execution of the object detection method program by the processor realizes the following steps:
preprocessing millimeter wave Lei Dadian cloud data to obtain point cloud data of a pedestrian target;
setting an initial value of a particle filtering algorithm;
initializing a dirac particle array in a particle filtering algorithm by using radar precision, following frame number and serial number;
predicting a pedestrian target and outputting a predicted value;
performing target matching by using particle filtering, updating particle weight and normalizing;
updating the target pool output target.
A third aspect of the present application provides a computer-readable storage medium having embodied therein an object detection method program which, when executed by a processor, implements the steps of the object detection method.
Compared with the prior art, the technical scheme of the application has the beneficial effects that:
the application initializes the dirac particle array based on the particle filtering algorithm by utilizing the radar precision, the following frame number and the serial number, and updates the weight in the particle array by prediction iteration, thereby effectively reducing the uncertainty of the randomness of particle generation of the particle filtering algorithm on the detection of the pedestrian target, leading the detection of the state of the pedestrian not to depend on the linear behavior prediction any more, solving the problem of the jumping of the visual target under the condition of insufficient light and the problem of false recognition or missing recognition caused by excessive clutter of the millimeter wave radar in a relatively closed space, and obviously improving the recognition precision of the pedestrian target, thereby improving the driving safety.
Drawings
Fig. 1 is a flowchart of a target detection method according to an embodiment of the present application.
Fig. 2 is a flowchart of initializing a dirac particle array according to an embodiment of the present application.
FIG. 3 is a flow chart of performing object matching for particle filtering according to an embodiment of the present application.
Detailed Description
In order that the above-recited objects, features and advantages of the present application will be more clearly understood, a more particular description of the application will be rendered by reference to the appended drawings and appended detailed description. It should be noted that, without conflict, the embodiments of the present application and features in the embodiments may be combined with each other.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application, however, the present application may be practiced in other ways than those described herein, and therefore the scope of the present application is not limited to the specific embodiments disclosed below.
Example 1
As shown in fig. 1, the first aspect of the present application provides a target detection method, which includes the following steps:
s1, preprocessing millimeter wave Lei Dadian cloud data to obtain point cloud data of a pedestrian target;
in the application, the point cloud data of the millimeter wave radar can be received through the Ethernet, the first frame of point cloud data is firstly subjected to physical value conversion, the distance and the azimuth angle are converted into X, Y coordinates according to the definition of a vehicle coordinate system and the trigonometric function relation, and then the coordinate values are converted into the vehicle coordinate system and stored in the global variable. Then filtering by using a preset first threshold, and eliminating false recognition points; the first threshold is a coarse threshold, for example, false identification points are removed by setting an energy threshold, and impurity points can be removed by signal-to-noise ratio and X, Y value, so that the quality of point cloud is improved. And clustering, RCS (radar cross section) pedestrian and vehicle distinguishing and second threshold filtering are sequentially carried out on the point cloud data according to the positions and the speeds of the points after the first threshold filtering, so as to obtain the point cloud data of the pedestrian target. The second threshold in the present application may be distance, speed, angle, etc.
S2, setting an initial value of a particle filter algorithm;
note that, for the first frame point cloud data, the tracking frame number of the pedestrian target is recorded as TrackFrame, trackFrame and is set as a set value, the number is recorded as TrackID, and the pedestrian target is stored in the target pool TargetPool as an initial value of a particle filtering algorithm.
S3, initializing a Dirac particle array in a particle filter algorithm by using radar precision, following frame number and serial number;
as shown in fig. 2, in a specific embodiment, the initialization is as follows:
s301, constructing a two-dimensional array with a row number m and a column number n as preset values for the numbered TrackID of each tracking frame number;
s302, the first items of the two-dimensional arrays are TrackID, trackFrame respectively, the distance of each grid in the two-dimensional arrays is recorded as Dmdn, and the value of Dmdn is the distance precision of the radar.
It should be noted that, considering that an excessively large array would have a large demand on the computational load, and an excessively small array would result in inaccurate prediction of the target, after weighing, a two-dimensional array DiracArray, diracArray with m and n values of 101 (this term is an adjustable parameter) is built for each trackID, the m and n values of the first term are trackID and TrackFrame respectively, so as to facilitate matching of the target and the particle, considering that a stable particle method is required to reduce the computational load and randomness, the distance DmDn of each grid is the distance precision Dmin of the radar, and since for initialization, the trackID is 1, the weight of each particle is set to be 1/100=0.01;
s4, predicting the pedestrian target and outputting a predicted value;
in a specific embodiment, the state equations of the X direction and the Y direction of the points in the set point cloud are respectively:
X k+1 frame interval time (frametime) +x k
Y k+1 Frame interval time (frametime) +y k
Where k is the number of iterations, where the value of k is equal to the tracking frame number TrackFrame, and the updated target is noted as PredictTarget, trackID and other attributes along with the initial pedestrian target.
S5, performing target matching by utilizing particle filtering, updating particle weight and normalizing;
as shown in fig. 3, in a specific embodiment, the specific steps are:
s501, taking a pedestrian target obtained by preprocessing point cloud data as an observation value of a new round;
s502, calculating Euclidean distance between the new target corresponding to the new round of observation value and each target in the target pool, taking the target in the target pool corresponding to the minimum Euclidean distance as a first new target,
s503, carrying out matching verification on the first new target passing the energy threshold to obtain a second new target;
s504, modifying the particle weight of the two-dimensional array in the target TrackID by using the second new target, and normalizing the modified weight.
The particle weight is modified by traversing the two-dimensional array, taking the X direction as an example, and normalizing the particle weight by a simple likelihood probability formula to achieve the effect of updating;
the simple version likelihood probability formula expression is as follows:
DiracArray(i)＝exp(-((NewTarget.x-(PredictTarget.x)2)2/(2*0.001))
s6, updating the target pool output target.
In a specific embodiment the trackID target position is updated with the updated particle weights and the target is output when the tracking frame number, trackFrame, is greater than a set threshold TrackFrameThresh.
In a second aspect, the present application provides an object detection system comprising: the device comprises a memory and a processor, wherein the memory comprises an object detection method program, and the execution of the object detection method program by the processor realizes the following steps:
s1, preprocessing millimeter wave Lei Dadian cloud data to obtain point cloud data of a pedestrian target;
in the application, the point cloud data of the millimeter wave radar can be received through the Ethernet, the first frame of point cloud data is firstly subjected to physical value conversion, the distance and the azimuth angle are converted into X, Y coordinates according to the definition of a vehicle coordinate system and the trigonometric function relation, and then the coordinate values are converted into the vehicle coordinate system and stored in the global variable. Then filtering by using a preset first threshold, and eliminating false recognition points; the first threshold is a coarse threshold, if false identification points are removed by setting an energy threshold, the impurity points can be filtered through a signal-to-noise ratio and an XY value, and the quality of the point cloud is improved. And clustering, RCS man-vehicle distinguishing and second threshold filtering are sequentially carried out on the point cloud data according to the positions and the speeds of the points after the first threshold filtering, so as to obtain the point cloud data of the pedestrian target. The second threshold in the present application may be distance, speed, angle, etc.
S2, setting an initial value of a particle filter algorithm;
note that, for the first frame point cloud data, the tracking frame number of the pedestrian target is recorded as TrackFrame, trackFrame and is set as a set value, the number is recorded as TrackID, and the pedestrian target is stored in the target pool TargetPool as an initial value of a particle filtering algorithm.
S3, initializing a Dirac particle array in a particle filter algorithm by using radar precision, following frame number and serial number;
in a specific embodiment, a two-dimensional array of X, Y is constructed for each tracking frame number numbered TrackID, wherein X represents the number of distance points in the X direction and Y represents the number of distance points in the Y direction;
the first item of the two-dimensional array is TrackID, trackFrame, the distance of each grid in the two-dimensional array is recorded as DxDy, and the value of the DxDy is the distance precision of the radar.
It should be noted that, considering that an excessively large array would have a large demand on the computational load, and an excessively small array would result in inaccurate prediction of the target, after weighing, a two-dimensional array DiracArray, diracArray with m and n values of 101 (this term is an adjustable parameter) is built for each trackID, the m and n values of the first term are trackID and TrackFrame respectively, so as to facilitate matching of the target and the particle, considering that a stable particle method is required to reduce the computational load and randomness, the distance DmDn of each grid is the distance precision Dmin of the radar, and since for initialization, the trackID is 1, the weight of each particle is set to be 1/100=0.01;
s4, predicting the pedestrian target and outputting a predicted value;
in a specific embodiment, the state equations of the X direction and the Y direction of the points in the set point cloud are respectively:
X k+1 frame interval time (frametime) +x k
Y k+1 Frame interval time (frametime) +y k
Where k is the number of iterations, where the value of k is equal to the tracking frame number TrackFrame, and the updated target is noted as PredictTarget, trackID and other attributes along with the initial pedestrian target.
S5, performing target matching by utilizing particle filtering, updating particle weight and normalizing;
in a specific embodiment, taking a pedestrian target obtained by preprocessing point cloud data as a new round of observation value;
calculating Euclidean distance between the new target corresponding to the new round of observation value and each target in the target pool, taking the target in the target pool corresponding to the minimum Euclidean distance as a first new target,
the first new target passes through an energy threshold to be subjected to matching verification to obtain a second new target;
and modifying the particle weight of the two-dimensional array in the target TrackID by using a second new target, and normalizing the modified weight.
The particle weight is modified by traversing the two-dimensional array, taking the X direction as an example, and normalizing the particle weight by a simple likelihood probability formula to achieve the effect of updating;
the simple version likelihood probability formula expression is as follows:
DiracArray(i)＝exp(-((NewTarget.x-(PredictTarget.x)2)2/(2*0.001))
s6, updating the target pool output target.
In a specific embodiment, the trackID target position is updated with the updated particle weights and the target is output when the tracking frame number, trackFrame, is greater than a set threshold TrackFrameThresh.
A third aspect of the present application provides a computer-readable storage medium having embodied therein an object detection method program which, when executed by a processor, implements the steps of the object detection method.
It is to be understood that the above examples of the present application are provided by way of illustration only and not by way of limitation of the embodiments of the present application. Other variations or modifications of the above teachings will be apparent to those of ordinary skill in the art. It is not necessary here nor is it exhaustive of all embodiments. Any modification, equivalent replacement, improvement, etc. which come within the spirit and principles of the application are desired to be protected by the following claims.
Claims (10)
1. A method of target detection comprising the steps of:
preprocessing millimeter wave Lei Dadian cloud data to obtain point cloud data of a pedestrian target;
setting an initial value of a particle filtering algorithm;
initializing a dirac particle array in a particle filtering algorithm by using radar precision, following frame number and serial number;
predicting a pedestrian target and outputting a predicted value;
performing target matching by using particle filtering, updating particle weight and normalizing;
updating the target pool output target.
2. The target detection method according to claim 1, wherein the preprocessing millimeter wave Lei Dadian cloud data specifically includes:
filtering the first frame of millimeter wave Lei Dadian cloud data by utilizing a preset first threshold, and eliminating false identification points;
calculating the position and the speed of each point in the point cloud data;
and clustering, RCS man-vehicle distinguishing and second threshold filtering are sequentially carried out on the point cloud data according to the positions and the speeds of the points to obtain the point cloud data of the pedestrian target.
3. The method for detecting an object according to claim 1, wherein the setting of the initial value of the particle filter algorithm is specifically: and the tracking frame number of the pedestrian target is recorded as TrackFrame, trackFrame and is set as a set value, the number is recorded as a TrackID, and the pedestrian target is stored in a target pool TargetPool to be used as an initial value of a particle filtering algorithm.
4. The method for detecting a target according to claim 3, wherein initializing the dirac particle array in the particle filter algorithm by using the radar precision, the following frame number and the number is specifically:
constructing a two-dimensional array with the number m of lines and the number n of columns as preset values for the numbered TrackID of each tracking frame number;
the first item of the two-dimensional array is TrackID, trackFrame, the distance of each grid in the two-dimensional array is recorded as Dmdn, and the value of Dmdn is the distance precision of the radar.
5. A method of detecting an object according to claim 3, wherein when the TrackID is 1, the initial value of the particle weight is 0.01.
6. A target detection method according to claim 3, wherein the pedestrian target is predicted and a predicted value is output, specifically:
the state equations of the points in the point cloud in the X direction and the Y direction are respectively as follows:
X k+1 frame interval time (frametime) +x k
Y k+1 Frame interval time (frametime) +y k
Where k is the number of iterations, where the value of k is equal to the tracking frame number TrackFrame, and the updated target is noted as PredictTarget, trackID and other attributes along with the initial pedestrian target.
7. The method for detecting an object according to claim 6, wherein the object matching is performed by using particle filtering, and the particle weights are updated and normalized, specifically:
taking a pedestrian target obtained by preprocessing the point cloud data as an observation value of a new round;
calculating Euclidean distance between the new target corresponding to the new round of observation value and each target in the target pool, taking the target in the target pool corresponding to the minimum Euclidean distance as a first new target,
the first new target passes through an energy threshold to be subjected to matching verification to obtain a second new target;
and modifying the particle weight of the two-dimensional array in the target TrackID by using a second new target, and normalizing the modified weight.
8. A method of object detection according to claim 3, wherein updating the object pool output object is specifically:
and updating the TrackID target position by using the updated particle weight, and outputting the target when the tracking frame number TrackFrame is larger than the set threshold TrackFrameThresh.
9. An object detection system, the system comprising: the device comprises a memory and a processor, wherein the memory comprises an object detection method program, and the execution of the object detection method program by the processor realizes the following steps:
preprocessing millimeter wave Lei Dadian cloud data to obtain point cloud data of a pedestrian target;
setting an initial value of a particle filtering algorithm;
initializing a dirac particle array in a particle filtering algorithm by using radar precision, following frame number and serial number;
predicting a pedestrian target and outputting a predicted value;
performing target matching by using particle filtering, updating particle weight and normalizing;
updating the target pool output target.
10. A computer-readable storage medium, characterized in that the computer-readable storage medium comprises therein an object detection method program, which, when executed by a processor, implements the steps of an object detection method according to any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310633537.0A CN116626668A (en) | 2023-05-31 | 2023-05-31 | Target detection method, system and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310633537.0A CN116626668A (en) | 2023-05-31 | 2023-05-31 | Target detection method, system and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116626668A true CN116626668A (en) | 2023-08-22 |
Family
ID=87602357
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310633537.0A Pending CN116626668A (en) | 2023-05-31 | 2023-05-31 | Target detection method, system and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116626668A (en) |
-
2023
- 2023-05-31 CN CN202310633537.0A patent/CN116626668A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Nabati et al. | Rrpn: Radar region proposal network for object detection in autonomous vehicles | |
| CN114384920B (en) | Dynamic obstacle avoidance method based on real-time construction of local grid map | |
| Yu et al. | Vehicle detection and localization on bird's eye view elevation images using convolutional neural network | |
| US9563808B2 (en) | Target grouping techniques for object fusion | |
| US8818702B2 (en) | System and method for tracking objects | |
| Weon et al. | Object Recognition based interpolation with 3d lidar and vision for autonomous driving of an intelligent vehicle | |
| CN112639502A (en) | Robot pose estimation | |
| Einsele | Real-time self-localization in unknown indoor environment using a panorama laser range finder | |
| CN103268616A (en) | Multi-feature multi-sensor method for mobile robot to track moving body | |
| CN1940591A (en) | System and method of target tracking using sensor fusion | |
| US10379542B2 (en) | Location and mapping device and method | |
| CN114998276B (en) | Robot dynamic obstacle real-time detection method based on three-dimensional point cloud | |
| CN112052802A (en) | Front vehicle behavior identification method based on machine vision | |
| CN112578673B (en) | Perception decision and tracking control method for multi-sensor fusion of formula-free racing car | |
| US11618480B2 (en) | Kurtosis based pruning for sensor-fusion systems | |
| Shan et al. | LiDAR-based stable navigable region detection for unmanned surface vehicles | |
| CN111508282A (en) | Low-altitude unmanned farmland operation flight obstacle conflict detection method | |
| Wen et al. | Research on 3D point cloud de-distortion algorithm and its application on Euclidean clustering | |
| CN113759928B (en) | Mobile robot high-precision positioning method for complex large-scale indoor scene | |
| CN114879207A (en) | Ultrasonic obstacle avoidance method for L4-level automatic driving vehicle | |
| CN116626668A (en) | Target detection method, system and storage medium | |
| CN113483769B (en) | Vehicle self-positioning method, system, equipment and medium based on particle filter | |
| Tamas et al. | Lidar and vision based people detection and tracking | |
| CN113947636B (en) | Laser SLAM positioning system and method based on deep learning | |
| US11948367B2 (en) | Multi-object tracking for autonomous vehicles |
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 |