CN109977885A - A kind of people's vehicle automatic identifying method and device based on Doppler Feature - Google Patents
A kind of people's vehicle automatic identifying method and device based on Doppler Feature Download PDFInfo
- Publication number
- CN109977885A CN109977885A CN201910248992.2A CN201910248992A CN109977885A CN 109977885 A CN109977885 A CN 109977885A CN 201910248992 A CN201910248992 A CN 201910248992A CN 109977885 A CN109977885 A CN 109977885A
- Authority
- CN
- China
- Prior art keywords
- target
- fluctuation
- sliding window
- vehicle
- velocity
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
- G06V20/584—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of vehicle lights or traffic lights
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Human Computer Interaction (AREA)
- Radar Systems Or Details Thereof (AREA)
- Traffic Control Systems (AREA)
Abstract
The present invention discloses a kind of people's vehicle automatic identifying method and device based on Doppler Feature, and this method step includes: that S1. uses radar to carry out target detection to investigative range inner region, and one group of object detection results is obtained when detecting target every time;S2. the continuous multiple groups object detection results of radar are obtained and extract the velocity information of target in every group of object detection results, obtain the continuous multiple groups velocity information of target;S3. the velocity perturbation trend that target is obtained by the continuous multiple groups velocity information that step S2 is obtained identifies that target is vehicle or pedestrian according to velocity perturbation trend.The present invention can automatic identification pedestrian and vehicle, and have many advantages, such as that implementation method is simple, complexity is low, accuracy of identification and recognition efficiency are high.
Description
Technical field
The present invention relates to Radar Targets'Detection technical fields more particularly to a kind of people's vehicle based on Doppler Feature to know automatically
Other method and device.
Background technique
Radar has outstanding moving-target detectability, is widely used in such as perimeter region security protection field, still
Radar itself is only able to detect target and sounds an alarm, and does not identify the classification of target, and such as perimeter region security protection etc.
In application scenarios, it usually needs can quickly distinguish mobile vehicle or pedestrian, targetedly be handled with executing.
The type identification of target may be implemented in classification of radar targets, and classification of radar targets is to know mode in the prior art
It is not applied in radar target acquisition with the relevant knowledge of machine learning, is usually all first to use the radar echo signal of target
Short Time Fourier Transform, extracts the characteristic information (such as micro-Doppler feature) that can embody target property, recycle support to
The feature of extraction is substituted into classifier, makes classification to the target data of input and sentence by the classifiers such as amount machine, convolutional neural networks
It is fixed, but such mode needs to extract the characteristic information of such as micro-Doppler feature, and micro-Doppler feature be extract it is more
It further extracts and obtains on the basis of general Le feature, extraction process is complex, can extend the identifying processing time, while using branch
It is high to hold the computationally intensive of the classifiers such as vector machine, convolutional neural networks, computation complexity, thus recognition efficiency is not high, it is difficult to
Quickly distinguish vehicle and pedestrian.And if considering direct extraction Doppler Feature to realize the differentiation of vehicle and pedestrian, although can
To reduce feature extraction complexity, but accuracy of identification is excessively poor, easily generation erroneous detection, and the speed such as people is usually relatively slow, but vehicle
It may also slow in some cases, simple operating speed is just difficult to differentiate between out vehicle and pedestrian.
Summary of the invention
The technical problem to be solved in the present invention is that, for technical problem of the existing technology, the present invention provides one
Kind of implementation method is simple, complexity is low, accuracy of identification and the high people's vehicle automatic identification based on Doppler Feature of recognition efficiency
Method and device.
In order to solve the above technical problems, technical solution proposed by the present invention are as follows:
A kind of people's vehicle automatic identifying method based on Doppler Feature, step include:
S1. target detection is carried out to investigative range inner region using radar, one group of target is obtained when detecting target every time
Testing result;
S2. the continuous multiple groups object detection results of radar are obtained and extract the speed letter of target in every group of object detection results
Breath, obtains the continuous multiple groups velocity information of target;
S3. the velocity perturbation trend that target is obtained by the continuous multiple groups velocity information that the step S2 is obtained, according to described
Velocity perturbation trend identifies that target is vehicle or pedestrian.
As a further improvement of the present invention: the speed letter of target in every group of object detection results is extracted in the step S2
When breath, further includes that the velocity information that will be extracted is converted using blind speed, obtain final velocity information step.
As a further improvement of the present invention: it is described using blind speed converted when, specifically according to formula [| vk/vm1|] carry out
Conversion, wherein the blind speed section of radar is-vm1~vm1, vkThe velocity amplitude of target is detected for kth time, [] is to divide exactly to take the remainder
Operation.
As a further improvement of the present invention: when identifying target according to the velocity perturbation trend in the step S3, if
The speed for determining target is in the fluctuation tendency for continuing to increase or continuing to decline in default first duration, or determines target
Speed, in the fluctuation tendency for continuing to increase or continuing to decline, is determined as vehicle, otherwise in more than two default second durations
It is determined as pedestrian.
As a further improvement of the present invention: when obtaining the velocity perturbation trend of target in the step S3, using default
The sliding window of size successively slips over each group velocity amplitude of target according to preset step-length, by counting in each sliding window former and later two
Fluctuation status between velocity amplitude judges fluctuation tendency of the speed of target in the sliding window.
As a further improvement of the present invention: when identifying target in the step S3, first judged using the first sliding window,
If determining there are being in the fluctuation tendency for continuing to increase or continuing to decline in the first sliding window described at least one, it is judged as vehicle,
Otherwise it continues to use the second sliding window to be judged, if determining, there are be in continue to increase or hold at least two second sliding windows
The fluctuation tendency of continuous decline, then be judged as vehicle, be otherwise determined as pedestrian, wherein the length of second sliding window is less than described the
The length of one sliding window.
As a further improvement of the present invention: when fluctuation tendency of the speed in the sliding window of the judgement target, tool
Body is by calculating separately the difference in the sliding window between former and later two velocity amplitudes, if the previous value >=0 of current value-, is determined as
Increase fluctuation status, and is correspondingly arranged the first fluctuation mark;If the previous value=0 of current value-, is determined as no fluctuation status, and
It is correspondingly arranged the second fluctuation mark;If the previous value < 0 of current value-, is judged to declining fluctuation status, and be correspondingly arranged third wave
Dynamic mark respectively fluctuates the quantity of mark by counting, determines wave of the speed of target in the sliding window in each sliding window
Dynamic trend.
As a further improvement of the present invention: the specific steps of the step S3 include:
S31. the 1st velocity amplitude { v detected to n-th of radar is obtained1,...vn};
It S32. the use of length is n1The first sliding window successively slip over each velocity amplitude { v1,...vn, it unites respectively when sliding every time
It counts the fluctuation status in first sliding window between each front and back velocity amplitude and corresponding fluctuation mark is set, it is sliding to obtain multiple groups first
The fluctuation statistical result of window, judging whether there is first sliding window is to be all the first fluctuation mark and second fluctuation
The combination of mark or the combination for being all the third fluctuation mark and the second fluctuation mark are moved back if so, being determined as vehicle
It identifies out, is otherwise transferred to and executes step S33;
It S33. the use of length is n2The second sliding window successively slip over each velocity amplitude { v1,...vn, it unites respectively when sliding every time
It counts the fluctuation status in second sliding window between each front and back velocity amplitude and corresponding fluctuation mark is set, it is sliding to obtain multiple groups second
The fluctuation statistical result of window, judge whether there is at least two second sliding windows be all it is described first fluctuation mark with it is described
The combination of second fluctuation mark or the combination for being all the third fluctuation mark with the second fluctuation mark, if so, determining
For vehicle, otherwise it is determined as pedestrian, exits identification.
A kind of people's vehicle automatic identification equipment based on Doppler Feature, comprising:
Radar Detection module detects target for carrying out target detection to investigative range inner region using radar every time
When obtain one group of object detection results;
Doppler Feature extraction module, for obtaining the continuous multiple groups object detection results of radar and extracting every group of target inspection
The velocity information for surveying target in result, obtains the continuous multiple groups velocity information of target;
Automatic identification module, the continuous multiple groups velocity information for being obtained according to the Doppler Feature extraction module obtain
The velocity perturbation trend of target identifies that target is vehicle or pedestrian according to the velocity perturbation trend.
A kind of pedestrian's vehicle automatic identification equipment based on Doppler Feature, the computer including being stored with computer program can
Read storage medium, which is characterized in that the computer program realizes above-mentioned method when executing.
Compared with the prior art, the advantages of the present invention are as follows:
1, it the present invention is based on the people's vehicle automatic identifying method and device of Doppler Feature, fully considers between pedestrian and vehicle
Speed point wave characteristic, pedestrian's vehicle automatic identification is realized using the characteristic, by first extracting in Radar Targets'Detection result
The Doppler Feature of velocity information, feature extraction is simple and efficient, on this basis by continuous velocity acquisition of information velocity perturbation
Trend realizes vehicle and the differentiation of pedestrian using velocity perturbation trend, realizes simply, can quickly distinguish vehicle and row
People, and velocity perturbation characteristic between vehicle and pedestrian is taken full advantage of, accuracy of identification is high, is particularly suitable for such as security protection and only needs
Distinguish the field of mobile vehicle and pedestrian.
2, the present invention is based on the people's vehicle automatic identifying methods and device of Doppler Feature, by judging whether there is default
In the fluctuation tendency for continuing to increase or continuing to decline in first duration, or with the presence or absence of in more than two default second durations
In the fluctuation tendency for continuing to increase or continuing to decline, identified using two kinds of characteristics of car speed fluctuation tendency target whether be
Vehicle can fast and accurately identify vehicle.
3, the present invention is based on the people's vehicle automatic identifying method and device of Doppler Feature, by first using biggish sliding window come
Judge whether the velocity perturbation trend of target meets the first velocity perturbation trend of vehicle, if being unsatisfactory for reusing lesser cunning
Window judges whether the velocity perturbation trend of target meets second of velocity perturbation trend of vehicle, and can quickly find out is
It is no to meet two kinds of velocity perturbation trend of vehicle, to efficiently distinguish vehicle and pedestrian.
Detailed description of the invention
Fig. 1 is the implementation process schematic diagram of people vehicle automatic identifying method of the present embodiment based on Doppler Feature.
Fig. 2 is the people's velocity perturbation curve synoptic diagram obtained in a particular embodiment.
Fig. 3 is the car speed curve of cyclical fluctuations schematic diagram obtained in a particular embodiment.
Fig. 4 is the realization principle schematic diagram that the present embodiment slips over each velocity amplitude using sliding window.
Fig. 5 is the schematic illustration that statistics calculates fluctuating change in sliding window in the present embodiment.
Fig. 6 is the implementation process schematic diagram that step S3 realizes people's vehicle Division identification in concrete application embodiment of the present invention.
Specific embodiment
Below in conjunction with Figure of description and specific preferred embodiment, the invention will be further described, but not therefore and
It limits the scope of the invention.
As shown in Figure 1, people vehicle automatic identifying method step of the present embodiment based on Doppler Feature includes:
S1. detections of radar: target detection is carried out to investigative range inner region using radar, is obtained when detecting target every time
One group of object detection results;
S2. Doppler Feature extracts: obtaining the continuous multiple groups object detection results of radar and extracts every group of target detection knot
The velocity information of target in fruit obtains the continuous multiple groups velocity information of target;
S3. the velocity perturbation trend of target, root automatic identification: are obtained according to the continuous multiple groups velocity information that step S2 is obtained
It is vehicle or pedestrian according to velocity perturbation trend identification target.
The present embodiment uses radar to carry out repeated detection, the speed that will be detected every time to pedestrian and vehicle respectively in advance
Point value is stored after being converted using blind speed, and the speed point curve of cyclical fluctuations of the pedestrian and vehicle that detect are respectively such as Fig. 2 and Fig. 3 institute
Show, wherein the longitudinal axis is represented using the speed after blind speed conversion, and horizontal axis represents the time, and each frame correspondence detects each time, for side
Just it analyzes, Fig. 3 is labelled with the transverse and longitudinal axial coordinate value of wherein five speed points, as can be seen that the speed of pedestrian from Fig. 2 and Fig. 3
Point fluctuates larger, the velocity perturbation very little of vehicle, and variation tendency than more consistent, i.e. the difference of pedestrian and vehicle are the speed of pedestrian
Degree point fluctuation is very big, and variation tendency back and forth is presented, and the variation of the speed point of vehicle is relatively steady.
The present embodiment fully considers the speed point wave characteristic between above-mentioned pedestrian and vehicle, realizes row using the characteristic
People's vehicle automatic identification, by first extracting the Doppler Feature of velocity information in Radar Targets'Detection result, feature extraction it is simple and
Efficiently, while on this basis by continuous velocity acquisition of information velocity perturbation trend, vehicle is realized using velocity perturbation trend
With the differentiation of pedestrian, realizes simply, can quickly distinguish vehicle and pedestrian, and take full advantage of speed between vehicle and pedestrian
Wave characteristic, accuracy of identification is high, is particularly suitable for the field that such as security protection only needs to distinguish mobile vehicle and pedestrian.
It further include that will mention when extracting the velocity information of target in every group of object detection results in step S2 in the present embodiment
The velocity information got is converted using blind speed, obtains final velocity information step.By being surveyed in the measurement period every time
After the speed for measuring target, the speed of target is converted into blind speed area, it being capable of speed point wave clear by the speed point after converting
It is dynamic, make it possible to more intuitive, apparent get speed point fluctuation tendency.
In the present embodiment, when being converted using blind speed, specifically by each velocity amplitude according to formula [| vk/vm1|] rolled over
It calculates, the velocity amplitude after being converted, wherein the blind speed section of radar is-vm1~vm1, vkThe speed of target is detected for kth time
Value, [] are to divide exactly to take the remainder operation.
In the present embodiment, when identifying target according to velocity perturbation trend in step S3, if determining the speed of target pre-
If in the fluctuation tendency for continuing to increase or continuing to decline in the first duration, or determining the speed of target more than two default
In the fluctuation tendency for continuing to increase or continuing to decline in second duration, it is determined as vehicle, is otherwise determined as pedestrian.
The present embodiment sufficiently analyzes the speed point wave characteristic of vehicle in advance, and the speed point fluctuation tendency for obtaining vehicle has
Following characteristic: one, there are have identical (increase) upwards or downwards (decline) variation tendency in continuous one section of longer duration;
Two, there are it is more than two continuous one section shorter when it is long in have identical (increases) upwards or downward (decline) variation tendency, and
There is interval between each group, i.e., each group duration is discontinuous.The present embodiment makes full use of above-mentioned characteristic, by judging target
Whether speed meets above-mentioned two characteristic, i.e., becomes with the presence or absence of in default first duration in the fluctuation for continuing to increase or continuing to decline
Gesture, or preset with the presence or absence of more than two in the fluctuation tendency for continuing to increase or continuing to decline in the second durations, thus according to
Identify whether target is vehicle, and fast and accurately target identification may be implemented according to velocity perturbation trend.
Above-mentioned default first duration, the second duration can specifically be set according to actually required precision, efficiency consideration, be needed
Meet the first duration greater than the second duration.
In the present embodiment, when obtaining the velocity perturbation trend of target in step S3, the sliding window of specifically used default size is pressed
The each group velocity amplitude that target is successively slipped over according to preset step-length, by counting the fluctuation in each sliding window between former and later two velocity amplitudes
State judges fluctuation tendency of the speed of target in sliding window, thus can by the fluctuation status between velocity amplitude each in sliding window
With the fluctuation tendency of speed point in the continuous one section of duration of Efficient Characterization.Each velocity amplitude is slipped over using sliding window as shown in figure 4, wherein every
One hatched box represents a speed point, and first grid stores speed point v1, second grid storage speed point v2, with this
Analogize.
When identifying target in the present embodiment, in step S3, first judged using the first sliding window, if determining in the presence of at least
In the fluctuation tendency for continuing to increase or continuing to decline in one the first sliding window, it is judged as vehicle, otherwise continues to use the second sliding window
Judged, if determining, there are, in the fluctuation tendency for continuing to increase or continuing to decline, judge at least two second sliding windows
For vehicle, otherwise it is determined as pedestrian, wherein length of the length of the second sliding window less than the first sliding window.By first using biggish cunning
Window judges whether the velocity perturbation trend of target meets the first velocity perturbation trend of above-mentioned vehicle, if being unsatisfactory for reusing
Lesser sliding window judges whether the velocity perturbation trend of target meets second of velocity perturbation trend of above-mentioned vehicle, can be fast
Speed accurately finds out whether meet the above two velocity perturbation trend of vehicle, to distinguish vehicle and pedestrian.
In the present embodiment, when judging fluctuation tendency of the speed of target in sliding window, especially by calculating separately in sliding window
Difference between former and later two velocity amplitudes, if the previous value >=0 of current value-, is judged to increasing fluctuation status, as ascending wave
It is dynamic, and it is correspondingly arranged the first fluctuation mark (specifically can be configured to 1);If the previous value=0 of current value-, is determined as no undulating
State as without fluctuation, and is correspondingly arranged the second fluctuation mark (specifically can be configured to 0);If the previous value < 0 of current value-, determines
To decline fluctuation status, as fluctuation downwards, and it is correspondingly arranged third fluctuation mark (specifically can be configured to -1), it is every by counting
The quantity that mark is respectively fluctuated in a sliding window can be convenient, efficiently determine fluctuation tendency of the speed of target in sliding window.
As shown in fig. 6, the specific steps of step S3 include: in the present embodiment
S31. the 1st velocity amplitude { v detected to n-th of radar is obtained1... vn};
It S32. the use of length is n1The first sliding window successively slip over each velocity amplitude { v1... vn, it unites respectively when sliding every time
It counts the fluctuation status in the first sliding window between each front and back velocity amplitude and corresponding fluctuation mark is set, obtain the first sliding window of multiple groups
Statistical result is fluctuated, judging whether there is the first sliding window is to be all the combination of the first fluctuation mark and the second fluctuation mark or be all
Third fluctuation mark and the combination of the second fluctuation mark exit identification if so, being determined as vehicle, are otherwise transferred to execution step
S33;
It S33. the use of length is n2The second sliding window successively slip over each velocity amplitude { v1... vn, n1> n2, every time when sliding
The fluctuation status in the second sliding window between each front and back velocity amplitude is counted respectively and corresponding fluctuation mark is set, and obtains multiple groups second
The fluctuation statistical result of sliding window, judging whether there is at least two second sliding windows is to be all the first fluctuation mark and the second fluctuation mark
The combination of will or be all third fluctuation mark with second fluctuation mark combination be otherwise judged to going if so, being determined as vehicle
People exits identification.
Through the above steps, two kinds of characteristics that car speed fluctuation tendency can be utilized, by judging velocity perturbation trend
In whether there is continuous n1It is secondary to have identical variation tendency (upwardly or downwardly), or there are two groups of continuous n2Secondary having is identical
Variation tendency and two groups between exist interval, vehicle is efficiently identified, within recognition efficiency is up to 3s.
In concrete application embodiment, when each sliding window calculates, if the previous value >=0 of current value-, setting fluctuation mark is
1, if the previous value=0 of current value -, setting fluctuation mark is 0, if the previous value < 0 of current value -, setting fluctuation mark is -1,
Each velocity amplitude { v is successively slipped over using the first sliding window1... vnWhen, if it exists in sliding window for (1,1,1....00...) or (1 ,-
1, -1....00...), then it is determined as vehicle, otherwise continues to use the second sliding window and judged, more than two if it exists, sliding window
Interior is (1,1,1....00...) or (1, -1, -1....00...), then is determined as vehicle, otherwise can determine that as pedestrian.Such as Fig. 5
It is shown, the fluctuation tendency of speed is judged by statistics sliding window 1,0 and -1 number.
The present invention realizes the detailed step of people's vehicle automatic identification in concrete application embodiment are as follows:
Step 1: setting the blind speed section of radar as-vm1~vm1, radar the 1st time is deposited to the velocity amplitude that n-th detects
Storage is got off, and { v is denoted as1... vn, vkThe speed of target is detected for kth time;
Step 2: the speed stored being converted using blind speed, i.e., to { v1... vnIn each detection speed
Carry out [| vk/vm1|] operation, remember that the speed array after operating becomes { v '1... v 'n};
Step 3: being n with length1Larger sliding window count { v '1... v 'nIn front and back speed point variation fluctuation situation,
If >=0, Flag=1 of the previous value of current value-in represent fluctuation upwards, if current value-previous value=0, Flag=0, indicate
Without fluctuation, if the previous value < 0, Flag=-1 of current value-, fluctuation downwards is represented;Count 1 to n1In sliding window grid 1,0 and-
After 1 number, sliding window moves right one, i.e., counts 2 to n again1In+1 grid after 1,0 and -1 number, and so on, until
The number that length is in the entire complete sliding window of grid counting statistics of n 1,0 and -1 is counted, if there is being all (1,1 ... in sliding window
Object judgement is then vehicle, terminates identification, otherwise continue to execute by the case where .0 ...) or being all (- 1, -1 ... .0 ...) or be all 0
Next step 4.
Step 4: changing the length of sliding window, the length for changing rear window is n2(n1- 3 < n2< n1), counting statistics length n2It is sliding
1,0 and -1 number, is next n with length in window2Sliding window end be starting point, be spaced L1(usual value 2,3,4) a point,
Counting statistics length is n again21,0 and -1 number in sliding window;If two length are n2Sliding window in be all (1,1 ... simultaneously
.0 it ...) or is all (- 1, -1 ... .0 ...) or is all 0, be then judged as vehicle, terminate identification.Otherwise continue to execute and sliding window rise
Initial point elapses one backward, repeats step 4 until completing all calculating;
Step 5: if not currently being judged as vehicle yet, by target classification at pedestrian, terminating identification.
People vehicle automatic identification equipment of the present embodiment based on Doppler Feature, comprising:
Radar Detection module detects target for carrying out target detection to investigative range inner region using radar every time
When obtain one group of object detection results;
Doppler Feature extraction module, for obtaining the continuous multiple groups object detection results of radar and extracting every group of target inspection
The velocity information for surveying target in result, obtains the continuous multiple groups velocity information of target;
Automatic identification module, the continuous multiple groups velocity information for being obtained according to Doppler Feature extraction module obtain target
Velocity perturbation trend, according to velocity perturbation trend identify target be vehicle or pedestrian.
Pedestrian vehicle automatic identification equipment of the present embodiment based on Doppler Feature, the calculating including being stored with computer program
Machine readable storage medium storing program for executing realizes such as the above method when computer program executes.
Pedestrian vehicle automatic identification equipment of the present embodiment based on Doppler Feature and the above-mentioned pedestrian based on Doppler Feature
Vehicle automatic identifying method is to correspond, and this is no longer going to repeat them.
Above-mentioned only presently preferred embodiments of the present invention, is not intended to limit the present invention in any form.Although of the invention
It has been disclosed in a preferred embodiment above, however, it is not intended to limit the invention.Therefore, all without departing from technical solution of the present invention
Content, technical spirit any simple modifications, equivalents, and modifications made to the above embodiment, should all fall according to the present invention
In the range of technical solution of the present invention protection.
Claims (10)
1. a kind of people's vehicle automatic identifying method based on Doppler Feature, which is characterized in that step includes:
S1. target detection is carried out to investigative range inner region using radar, one group of target detection is obtained when detecting target every time
As a result;
S2. the continuous multiple groups object detection results of radar are obtained and extract the velocity information of target in every group of object detection results,
Obtain the continuous multiple groups velocity information of target;
S3. the velocity perturbation trend that target is obtained by the continuous multiple groups velocity information that the step S2 is obtained, according to the speed
Fluctuation tendency identifies that target is vehicle or pedestrian.
2. people's vehicle automatic identifying method according to claim 1 based on Doppler Feature, which is characterized in that the step
When extracting the velocity information of target in every group of object detection results in S2, further include by the velocity information extracted using blind speed into
Row conversion, obtains final velocity information step.
3. people's vehicle automatic identifying method according to claim 2 based on Doppler Feature, it is characterised in that: the use
When blind speed is converted, specifically according to formula [| vk/vm1|] converted, wherein the blind speed section of radar is-vm1~vm1, vkIt is
Detect the velocity amplitude of target for k times, [] is to divide exactly to take the remainder operation.
4. people's vehicle automatic identifying method according to claim 1 or 2 or 3 based on Doppler Feature, which is characterized in that institute
When stating in step S3 according to velocity perturbation trend identification target, if the speed for determining target is in default first duration
The fluctuation tendency for continuing to increase or continuing to decline, or determining the speed of target is in hold in more than two default second durations
The continuous fluctuation tendency increased or continue to decline, is determined as vehicle, is otherwise determined as pedestrian.
5. people's vehicle automatic identifying method according to claim 4 based on Doppler Feature, which is characterized in that the step
When obtaining the velocity perturbation trend of target in S3, each group of target is successively slipped over according to preset step-length using the sliding window of default size
Velocity amplitude judges the speed of target in institute by counting the fluctuation status in each sliding window between former and later two velocity amplitudes
State the fluctuation tendency in sliding window.
6. people's vehicle automatic identifying method according to claim 5 based on Doppler Feature, which is characterized in that the step
When identifying target in S3, first judged using the first sliding window, if determining, there are be in hold in the first sliding window described at least one
The continuous fluctuation tendency increased or continue to decline, is judged as vehicle, otherwise continues to use the second sliding window and judged, deposited if determining
In the fluctuation tendency for continuing to increase or continuing to decline at least two second sliding windows, then it is judged as vehicle, otherwise determines
For pedestrian, wherein the length of second sliding window is less than the length of first sliding window.
7. people's vehicle automatic identifying method according to claim 6 based on Doppler Feature, which is characterized in that the judgement
When fluctuation tendency of the speed of target in the sliding window, especially by calculate separately in the sliding window former and later two velocity amplitudes it
Between difference, if current value-it is previous value >=0, be judged to increasing fluctuation status, and be correspondingly arranged the first fluctuation mark;If working as
The preceding previous value=0 of value-, is determined as no fluctuation status, and be correspondingly arranged the second fluctuation mark;If the previous value < 0 of current value-,
It is judged to declining fluctuation status, and is correspondingly arranged third fluctuation mark, respectively fluctuates mark in each sliding window by counting
Quantity determines fluctuation tendency of the speed of target in the sliding window.
8. people's vehicle automatic identifying method according to claim 7 based on Doppler Feature, which is characterized in that the step
The specific steps of S3 include:
S31. the 1st velocity amplitude { v detected to n-th of radar is obtained1,...vn};
It S32. the use of length is n1The first sliding window successively slip over each velocity amplitude { v1,...vn, institute is counted respectively when sliding every time
It states the fluctuation status in the first sliding window between each front and back velocity amplitude and corresponding fluctuation mark is set, obtain the first sliding window of multiple groups
Statistical result is fluctuated, judging whether there is first sliding window is to be all the first fluctuation mark to indicate with second fluctuation
Combination or be all third fluctuation mark and the combination of the second fluctuation mark exits knowledge if so, being determined as vehicle
Not, it is otherwise transferred to and executes step S33;
It S33. the use of length is n2The second sliding window successively slip over each velocity amplitude { v1,...vn, institute is counted respectively when sliding every time
It states the fluctuation status in the second sliding window between each front and back velocity amplitude and corresponding fluctuation mark is set, obtain the second sliding window of multiple groups
Statistical result is fluctuated, judging whether there is at least two second sliding windows is to be all the first fluctuation mark and described second
It fluctuates the combination of mark or is all the combination of the third fluctuation mark with the second fluctuation mark, if so, being determined as vehicle
, otherwise it is determined as pedestrian, exits identification.
9. a kind of people's vehicle automatic identification equipment based on Doppler Feature characterized by comprising
Radar Detection module obtains when detecting target every time for carrying out target detection to investigative range inner region using radar
To one group of object detection results;
Doppler Feature extraction module, for obtaining the continuous multiple groups object detection results of radar and extracting every group of target detection knot
The velocity information of target in fruit obtains the continuous multiple groups velocity information of target;
Automatic identification module, the continuous multiple groups velocity information for being obtained according to the Doppler Feature extraction module obtain target
Velocity perturbation trend, according to the velocity perturbation trend identify target be vehicle or pedestrian.
10. a kind of people's vehicle automatic identification equipment based on Doppler Feature, including being stored with the computer-readable of computer program
Storage medium, which is characterized in that the side as described in any one of claim 1~8 is realized when the computer program executes
Method.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910248992.2A CN109977885A (en) | 2019-03-29 | 2019-03-29 | A kind of people's vehicle automatic identifying method and device based on Doppler Feature |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910248992.2A CN109977885A (en) | 2019-03-29 | 2019-03-29 | A kind of people's vehicle automatic identifying method and device based on Doppler Feature |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109977885A true CN109977885A (en) | 2019-07-05 |
Family
ID=67081636
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910248992.2A Pending CN109977885A (en) | 2019-03-29 | 2019-03-29 | A kind of people's vehicle automatic identifying method and device based on Doppler Feature |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109977885A (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110488279A (en) * | 2019-08-30 | 2019-11-22 | 成都纳雷科技有限公司 | A kind of pedestrian based on radar and trees differentiating method and device |
CN110488258A (en) * | 2019-08-30 | 2019-11-22 | 成都纳雷科技有限公司 | False alarm rejection method and device in a kind of perimeter region security protection |
CN111025288A (en) * | 2019-12-27 | 2020-04-17 | 森思泰克河北科技有限公司 | Security radar monitoring device and system |
CN113495270A (en) * | 2020-04-07 | 2021-10-12 | 富士通株式会社 | Monitoring device and method based on microwave radar |
CN113687351A (en) * | 2021-09-14 | 2021-11-23 | 无锡威孚高科技集团股份有限公司 | Security radar monitoring and early warning system capable of reducing false alarm rate |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2006177907A (en) * | 2004-12-24 | 2006-07-06 | Nec Corp | Interference-type radar |
US20070274566A1 (en) * | 2006-05-24 | 2007-11-29 | Nissan Motor Co., Ltd. | Pedestrian detector and pedestrian detecting method |
US20120007773A1 (en) * | 2010-07-08 | 2012-01-12 | Raytheon Company | Processing Radar Return Signals To Detect Targets |
CN102778680A (en) * | 2012-06-06 | 2012-11-14 | 西安电子科技大学 | Method for imaging uniformly accelerated motion rigid group targets based on parameterization |
CN102928835A (en) * | 2012-10-09 | 2013-02-13 | 北京航空航天大学 | Human body target motion state identifying method based on improved generalized S conversion |
CN105512641A (en) * | 2015-12-31 | 2016-04-20 | 哈尔滨工业大学 | Method for using laser radar scanning method to calibrate dynamic pedestrians and vehicles in video in snowing or raining state |
CN105701479A (en) * | 2016-02-26 | 2016-06-22 | 重庆邮电大学 | Intelligent vehicle multi-laser radar fusion recognition method based on target features |
CN108919195A (en) * | 2018-07-12 | 2018-11-30 | 中国船舶重工集团公司第七二四研究所 | A kind of alternative optimization search irregular middle repetition PD design method of arteries and veins group |
-
2019
- 2019-03-29 CN CN201910248992.2A patent/CN109977885A/en active Pending
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2006177907A (en) * | 2004-12-24 | 2006-07-06 | Nec Corp | Interference-type radar |
US20070274566A1 (en) * | 2006-05-24 | 2007-11-29 | Nissan Motor Co., Ltd. | Pedestrian detector and pedestrian detecting method |
US20120007773A1 (en) * | 2010-07-08 | 2012-01-12 | Raytheon Company | Processing Radar Return Signals To Detect Targets |
CN102778680A (en) * | 2012-06-06 | 2012-11-14 | 西安电子科技大学 | Method for imaging uniformly accelerated motion rigid group targets based on parameterization |
CN102928835A (en) * | 2012-10-09 | 2013-02-13 | 北京航空航天大学 | Human body target motion state identifying method based on improved generalized S conversion |
CN105512641A (en) * | 2015-12-31 | 2016-04-20 | 哈尔滨工业大学 | Method for using laser radar scanning method to calibrate dynamic pedestrians and vehicles in video in snowing or raining state |
CN105701479A (en) * | 2016-02-26 | 2016-06-22 | 重庆邮电大学 | Intelligent vehicle multi-laser radar fusion recognition method based on target features |
CN108919195A (en) * | 2018-07-12 | 2018-11-30 | 中国船舶重工集团公司第七二四研究所 | A kind of alternative optimization search irregular middle repetition PD design method of arteries and veins group |
Non-Patent Citations (3)
Title |
---|
RUEGG, M ET.AL: "Capabilities of dual-frequency millimeter wave SAR with monopulse processing for ground moving target indication", 《IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING》 * |
冀振元 等: "基于小波变换的人车目标特征提取算法", 《微计算机信息》 * |
张云展 译: "MTD雷达速度测量的可靠性", 《火控雷达技术》 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110488279A (en) * | 2019-08-30 | 2019-11-22 | 成都纳雷科技有限公司 | A kind of pedestrian based on radar and trees differentiating method and device |
CN110488258A (en) * | 2019-08-30 | 2019-11-22 | 成都纳雷科技有限公司 | False alarm rejection method and device in a kind of perimeter region security protection |
CN110488279B (en) * | 2019-08-30 | 2022-12-09 | 成都纳雷科技有限公司 | Pedestrian and tree distinguishing method and device based on radar |
CN111025288A (en) * | 2019-12-27 | 2020-04-17 | 森思泰克河北科技有限公司 | Security radar monitoring device and system |
CN111025288B (en) * | 2019-12-27 | 2022-05-31 | 森思泰克河北科技有限公司 | Security radar monitoring device and system |
CN113495270A (en) * | 2020-04-07 | 2021-10-12 | 富士通株式会社 | Monitoring device and method based on microwave radar |
CN113687351A (en) * | 2021-09-14 | 2021-11-23 | 无锡威孚高科技集团股份有限公司 | Security radar monitoring and early warning system capable of reducing false alarm rate |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109977885A (en) | A kind of people's vehicle automatic identifying method and device based on Doppler Feature | |
CN107290741B (en) | Indoor human body posture identification method based on weighted joint distance time-frequency transformation | |
CN102419437B (en) | Track-before-detect method based on flight path inspection | |
CN104299417B (en) | Vehicle identification method based on waveforms detection | |
CN109461095A (en) | Data processing method and equipment for load event detection for non-invasive load identification | |
CN109164450B (en) | Downburst prediction method based on Doppler radar data | |
CN103559508A (en) | Video vehicle detection method based on continuous Adaboost | |
CN104851301B (en) | Vehicle parameter identification method based on deceleration strip sound analysis | |
CN108107086A (en) | A kind of gas detection method and gas sensor based on array gas sensor | |
CN112068120A (en) | micro-Doppler time-frequency plane individual soldier squad identification method based on two-dimensional Fourier transform | |
CN101782964B (en) | Weight loss feather extraction method based on acceleration transducer | |
Ding et al. | A fall detection method based on K-nearest neighbor algorithm with MIMO millimeter-wave radar | |
CN106204641A (en) | Tracking before a kind of dynamic programming based on multihypothesis test detection | |
CN107436363B (en) | A kind of rail traffic vehicles speed dynamic measurement method | |
CN105743756A (en) | Frame detection method based on Adaboost algorithm in Wi-Fi system | |
CN103679183A (en) | Defect identification method for plain weave gray cloth | |
CN109409216A (en) | Speed adaptive indoor human body detection method based on subcarrier dynamic select | |
JP2007333515A (en) | Target detection device | |
CN110488279B (en) | Pedestrian and tree distinguishing method and device based on radar | |
CN105447511B (en) | A kind of SVM object detection method based on Adaboost Haar-Like feature | |
JP5623253B2 (en) | Multi-target tracking device | |
CN103730011B (en) | A kind of vehicle type recognition detection method based on laser instrument and system | |
CN105629210A (en) | Airborne radar space and ground moving target classification and recognition method | |
CN108646227A (en) | A kind of Multi-Target Classification Method based on microwave | |
CN105030231B (en) | A kind of fetal heart monitoring accelerates the extracting method of characteristic parameter |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190705 |