CN108256418A - A kind of pedestrian's method for early warning and system based on infrared imaging - Google Patents
A kind of pedestrian's method for early warning and system based on infrared imaging Download PDFInfo
- Publication number
- CN108256418A CN108256418A CN201711251575.0A CN201711251575A CN108256418A CN 108256418 A CN108256418 A CN 108256418A CN 201711251575 A CN201711251575 A CN 201711251575A CN 108256418 A CN108256418 A CN 108256418A
- Authority
- CN
- China
- Prior art keywords
- pedestrian
- target
- early warning
- pedestrian target
- pressure
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims description 41
- 238000003331 infrared imaging Methods 0.000 title claims description 14
- 238000001514 detection method Methods 0.000 claims abstract description 18
- 230000000007 visual effect Effects 0.000 claims abstract description 13
- 238000006073 displacement reaction Methods 0.000 claims description 17
- 230000003321 amplification Effects 0.000 claims description 8
- 238000013459 approach Methods 0.000 claims description 8
- 238000003199 nucleic acid amplification method Methods 0.000 claims description 8
- 238000005192 partition Methods 0.000 claims description 8
- 230000000717 retained effect Effects 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 2
- 238000009825 accumulation Methods 0.000 claims 1
- 238000005516 engineering process Methods 0.000 abstract description 15
- 230000008859 change Effects 0.000 abstract description 10
- 238000003384 imaging method Methods 0.000 abstract description 5
- 238000002329 infrared spectrum Methods 0.000 abstract description 3
- 238000001931 thermography Methods 0.000 description 11
- 239000012530 fluid Substances 0.000 description 9
- 238000012549 training Methods 0.000 description 5
- 238000006243 chemical reaction Methods 0.000 description 4
- 238000002474 experimental method Methods 0.000 description 4
- 230000004927 fusion Effects 0.000 description 4
- 210000001190 lateral line system Anatomy 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- 241000251468 Actinopterygii Species 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000035484 reaction time Effects 0.000 description 3
- 238000001228 spectrum Methods 0.000 description 3
- 230000001133 acceleration Effects 0.000 description 2
- 230000004888 barrier function Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012827 research and development Methods 0.000 description 1
- 230000009182 swimming Effects 0.000 description 1
Classifications
-
- 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
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W30/00—Purposes of road vehicle drive control systems not related to the control of a particular sub-unit, e.g. of systems using conjoint control of vehicle sub-units
- B60W30/08—Active safety systems predicting or avoiding probable or impending collision or attempting to minimise its consequences
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/20—Scenes; Scene-specific elements in augmented reality scenes
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Human Computer Interaction (AREA)
- Automation & Control Theory (AREA)
- Transportation (AREA)
- Mechanical Engineering (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention discloses a kind of based on pedestrian target prompting of the infrared spectrum imaging with artificial side line technology and collision warning systems, the Dynamic IR image of visual field scene is acquired by infrared assembly, the pedestrian target in the infrared image of each time frame is filtered out using Adaboost algorithm, make early warning more targeted, and further using crossing algorithm and approximate algorithm filters out risk object from pedestrian target, it is aided with distance and speed that artificial side line unit perceives risk object, calculate the collision time of vehicle and risk object, when collision time is less than specified threshold, provide anti-collision warning prompting.Present invention is particularly suitable for the crowd massings such as shopping centre, traffic intersection area, solve pedestrian's feature change it is larger, it is difficult to the problem of detection and early warning.
Description
Technical field
The invention belongs to safe driving field, more particularly, to a kind of safety pre-warning system based on infrared imaging.
Background technology
At present, in safe driving, it is seen that light imaging technique and computer vision technique comparative maturity, but due to can
The characteristics of seeing light imaging causes the imaging effect under the severe driving environment such as night, rainy day, haze poor, and these are severe
Driving environment is often again dangerous driving environment, it is therefore necessary to which research and development are for the driving auxiliary system of the extreme climate environments such as night
System.Infrared thermal imaging can provide a preferable driving environment under the environment such as night and greasy weather, but infrared thermal imaging is deposited
In the distance that can not judge front pedestrian target, it is impossible to provide accurate alarm.
With the development of electronic technology, the pedestrian target detecting system based on image procossing is gradually instead of traditional infrared
Or the modes such as radar, the pedestrian target detection technique based on image procossing have higher reliability, convenience and low cost and get over
More attract attention, and existing Adaboost pedestrian targets detector does not consider that pedestrian target exists near big remote in video image
Small dimensional variation causes detection error larger, it is impossible to make accurate early warning to dynamic pedestrian target.At present, it is logical both at home and abroad
It crosses and has carried out a series of researchs to the artificial lateral-line system of underwater robot loading, be applied to land but without artificial lateral-line system
Or the report of aviation field.
Invention content
For the disadvantages described above or Improvement requirement of the prior art, it is imaged and locates based on infrared spectrum the present invention provides one kind
The pedestrian target prompting of reason technology is with collision warning systems, it can be achieved that boisterous accurate pedestrian's early warning.This system utilizes
Adaboost algorithm filters out the pedestrian target in infrared image, and the pedestrian approached is further detected according to target approaches algorithm
Target provides pedestrian target and approaches prompting, when having detected that pedestrian target is crossed to steering direction side, according to pedestrian target
It crosses algorithm and judges pedestrian target state, and provided for different degrees of danger and different cross early warning.
Meanwhile to solve the problems, such as that infrared distance measurement is inaccurate in the prior art, present invention auxiliary is with artificial side line controller
Front pedestrian target is detected, front pedestrian target position and speed is perceived by pressure change, when with front pedestrian's mesh
When target collision time is less than specified threshold, prompting collision warning prompting is provided.Artificial side line technology is by fish lateral-line system
A kind of system that can perceive ambient condition information for inspiring and being developed into.Fish rely on its own lateral line system togetherness having
Know characteristic of fluid, can know flow direction and the flow intensity of fluid, and then complete the behaviors such as predation, study, cluster.Side line
System is based on being distributed in the neuromast of fish body different parts, during fish swimming, body and flow position
Fluid motion is perceived when generating variation.Artificial side line technology is applied to non-underwater field by the present invention for the first time, is believed by mechanical wave
Number test the distance of barrier in speed, the electromagnetic spectrum signal than usually driving early warning system use is more steady
It is fixed.
This system is compared with traditional DAS (Driver Assistant System), since infrared thermal imaging technique being combined with Adboost algorithms,
Pedestrian target under bad weather can be made and more accurately detected, there is preferably prompting and early warning effect.This system and electricity
Electromagnetic spectrum signal system is compared, and auxiliary is mounted with to receive the artificial side line unit of mechanical wave signals, improves the stability of system,
Driver is made to make effective reaction, avoids the generation of safety accident under bad weather.
To achieve the above object, one side according to the invention provides a kind of based on infrared thermal imaging and artificial side
Pedestrian's method for early warning of line technology is acquired the obstructions chart picture of driving environment with the method for infrared thermal imaging, profit
Pedestrian target is detected with Adaboost cascade classifier detection algorithms, early warning calculation is approached and crossed using pedestrian target
Method carries out approaching judgement and crosses judgement, and make corresponding early warning for different pedestrian target states to pedestrian target, profit
Manually side line technology calculates the relative velocity and distance of pedestrian target, so as to further calculate collision time, realizes
Detection and early warning of the extreme climate environments such as night to the pedestrian target of driving environment.
Pedestrian's method for early warning based on infrared imaging includes the following steps：
S1：Obtain visual field scene Dynamic IR image, and according to the Dynamic IR image obtain pedestrian target and its
Time frame dynamic, the time frame dynamic are that change in location and area of the pedestrian target in the Dynamic IR image different frame become
Change；
S2：Risk object is obtained according to pedestrian target and its time frame dynamic；
S3：The air flow information influenced by risk object is obtained, the air flow information includes more on the straight line in front
A effective pressure obtains the speed and distance of risk object according to the air flow information, according to the speed and distance of risk object
Collision time is obtained, and anti-collision warning signal is obtained according to the collision time.
Preferably, the step S1 includes：
S11：The infrared picture comprising pedestrian target is obtained as positive sample collection；
S12：The infrared picture without pedestrian target is obtained as negative sample collection；
S13：According to positive and negative sample set, training obtains the Adaboost cascade classifiers with pedestrian target feature recognition；
S14：With the Adaboost cascade classifiers detection infrared assembly obtain infrared image, obtain pedestrian target and
Its variation on each frame infrared image.
Preferably, the method for the acquisition pedestrian target is specially：
S141：By the infrared image equal area partition of current time frame in infrared image at least two subimage block；
S142：The subimage block is detected by Adaboost cascade classifiers, is rejected special without pedestrian target
The subimage block of sign retains the subimage block for including pedestrian target feature；
S143：By the further equal area partition of each subimage block of reservation at least two subimage block, pass through
Adaboost cascade classifiers are detected the subimage block finally divided, reject the subgraph without pedestrian target feature
As block, retain the subimage block for including pedestrian target feature；
S144：Judge whether the number divided is more than preset cycle-index, is to stop dividing, what fusion finally retained
Subimage block obtains pedestrian target, otherwise returns to S143.
Preferably, included in the step S2 according to the method that pedestrian target and its time frame dynamic obtain risk object：Institute
The amplification number of master control system detection pedestrian target in the time of 15~20 frames is stated, if amplification number is more than three times,
The pedestrian target is then judged for risk object, otherwise excludes the pedestrian target, the face for being enlarged into image shared by pedestrian target
Product becomes larger.
Preferably, included in the step S2 according to the method that pedestrian target and its time frame dynamic obtain risk object：
S21：When Δ t is equal to first threshold, if | Δ xt| then judge the pedestrian target to cross mesh more than second threshold
Mark, if | Δ yt| more than third threshold value and | kt|>|kt-1|, then judge that the pedestrian target crosses target to approach, obtain the first danger
Dangerous pre-warning signal；Otherwise enter S122, time changes of the Δ t for t frames, Δ xtFor t frame one skilled in the art's target left and right directions
Displacement, Δ ytFor the displacement in the front-back direction of t frame one skilled in the art's targets, the ktFor the displacement slope of t frame one skilled in the art's targets, kt-1For
(t-1) the displacement slope of frame one skilled in the art target, kt=Δ yt/Δxt, kt-1=Δ yt-1/Δxt-1, the Δ xt-1For (t-1) frame
The displacement of one skilled in the art's target left and right directions, Δ yt-1For the displacement in the front-back direction of (t-1) frame one skilled in the art's target；The first threshold
It is 1~6, second threshold is 1~3, and third threshold value is 1~3；
S22：When Δ t is equal to first threshold, if crossing target Δ xtMore than the 4th threshold value, then judge that this crosses target and is
Target is crossed on the right side, obtains the second danger early warning signal；If cross target Δ xtFor negative and | Δ xt| more than the 4th threshold value, then sentence
Disconnected pedestrian target crosses target for a left side, obtains third danger early warning signal；Otherwise the pedestrian target is excluded, the 4th threshold value is
2~6.
It is further included before the step S3：Offline preliminary experiment, the distance for controlling pedestrian is constant, sets the speed of multigroup pedestrian
Degree obtains effective pressure of different location on the straight line of front, according to having for the position of the speed of pedestrian and its correspondence
It imitates pressure and obtains the first relational model；
The speed for controlling pedestrian is constant, and the distance of multigroup pedestrian is set to obtain effective pressure of different location on straight line
Difference, the second relational model is obtained according to the difference of the distance of pedestrian and effective pressure；
The step S3 is specially：In driving conditions, by the difference of effective pressure of different location on the straight line of front
The distance that the second relational model obtains risk object is substituted into, by effective pressure with the position on the straight line of risk object correspondence
The strong speed for substituting into the first relational model and obtaining risk object, the position on the corresponding straight line of the risk object can be by red
Outer image obtains.
The acquisition methods of effective pressure include：The original pressure of different location on the straight line of front is obtained, according to
Natural wind speed obtains the pressure of natural wind generation with Bernoulli equation, and as benchmark wind pressure, base is subtracted from the original pressure
Quasi- wind pressure obtains effective pressure.According to Bernoulli equation, the stagnation pressure of fluid is equal to static the sum of pressure and dynamic pressure：
P0=Pfs+1/2ρv2
Wherein P0To stagnate pressure, PfsFor static pressure, 1/2 ρ v2For dynamic pressure, ρ and v represent respectively fluid density and
Fluid velocity.Due to the P in atmospheric pressure environmentfs=0, ρ and v represent the speed of atmospheric density and air flowing relative to front end respectively
Degree, therefore can obtain influence of the natural wind speed to pressure.The wind speed is with respect to natural wind speed, the acquisition of the natural wind speed
Method is utilizes vane tachymeter, heat-sensitive type anemobiagraph or the survey of one or more of pitot tube type anemobiagraph or pressure sensor
Obtain natural wind speed.
It is pre- to provide a kind of pedestrian based on infrared thermal imaging and artificial side line technology for other side according to the invention
Alert system, including artificial side line unit, infrared assembly, master control system, prior-warning device, the output terminal of the artificial side line unit
Connect the first input end of master control system, the second input terminal of the output terminal connection master control system of the infrared assembly, institute
The input terminal of the output terminal connection prior-warning device of master control system is stated,
The infrared assembly is used to obtain the Dynamic IR image of visual field scene, and the master control system is used for according to the visual field
The Dynamic IR image of scene obtains risk object, and the artificial side line unit is used for according to straight line in Dynamic IR image
The pressure information of upper different location obtains the speed and distance of risk object, and the master control system is further according to the danger
The speed of target with apart from obtain anti-collision warning signal.
The artificial side line unit includes array of pressure sensors, collection plate, IC bus, STM32 controllers, institute
The output terminal for stating array of pressure sensors connects the input terminal of the collection plate by IC bus, the collection plate it is defeated
Outlet connects the input terminal of the STM32 controllers, and the output terminal of the STM32 controllers is as the artificial side line unit
Output terminal,
The array of pressure sensors is used to be obtained according to the pressure information of different pressures sensor different on straight line
The flow direction of the air fluid of position and flow strength, the collection plate are used for the air according to different location on straight line
The flow direction of fluid and flow strength obtain the multiple effective pressure being located on straight line, and the STM32 controllers are used for
Speed, the distance of pedestrian is obtained according to the corresponding effective pressure of different location on straight line, the pedestrian is front pedestrian.
The array of pressure sensors is set along the stream pressure trace in front, including three or more be installed at middle net
Pressure sensor.
The prior-warning device includes the several of vehicle-mounted screen, light bar prior-warning device, one kind in buzzer prior-warning device or parallel connection
Kind.
In general, by the above technical scheme conceived by the present invention compared with prior art, it can obtain down and show
Beneficial effect：
1st, the present invention is compared with traditional infrared auxiliary driving technology, by extracting dangerous mesh from infrared dynamic image
Mark, specific aim is stronger, reduces the reaction time of driver, and the distance and speed of risk object are obtained by the air flow information in front
Degree calculates collision time, and carries out corresponding early warning, and driver is helped to judge and make significantly more efficient reaction, strengthens traveling
Safety.
2nd, for the complexity of pedestrian target motion state, the present invention is also additionally arranged approximate algorithm and crosses algorithm, compared with
Simple Adboost algorithms more fully consider the state and collision probability of target in reality in the prior art, provide reliable
Empirical parameter to extract risk object according to pedestrian target, and obtains corresponding early warning, advanced optimizes information, reduces driver
The judgement time.
3rd, artificial side line technology is applied in driving assistance system by the present invention for the first time, and usual infrared thermal imaging is in the application
It can only judge the orientation of pedestrian target in the picture, there is the speed and distance that can not obtain pedestrian target, it can only be by follow-up
An estimated value is calculated in image procossing, can not realize the problem of accurately judging, thus introduce artificial side line unit with it is red
The mode that outer thermal imaging member is combined more accurately is alarmed with realizing, this system is compared with electromagnetic spectrum signal system, auxiliary
It is mounted with to receive the artificial side line unit of mechanical wave signals, mechanical wave is than electromagnetic wave system more stability.
4th, the present invention measures different positions on the straight line of front by installing artificial side line unit at the pressure trace of front
Corresponding effective pressure is put, the pass of distance and speed for obtaining effective pressure and its difference and pedestrian is probed by test experiment
System so as to build relational model, is conducive to obtain according to effective pressure of different location on straight line is quick in the process of moving
To the distance and speed of pedestrian, collision time is further calculated, a kind of new way is provided to safety traffic.
5th, Adaboost cascade classifiers use subimage block cycle detection method rapid extraction pedestrian target, than it is traditional by
One detection method is faster.
6th, the function that the artificial side line unit of the present invention realizes high-speed computation by STM32 controllers and information is transmitted, is fitted
For driving early warning system, the reaction time is reduced.
7th, artificial side line unit is communicated by IC bus, which is suitable for multisystem and synchronizes sanction
Certainly, the information transmission time difference of different sensors can be reduced, make testing result relatively reliable.
8th, pressure sensor uses CPS131 sensors, sensitiveer than general sensor, ensures to obtain air flow information
Reliability supports IIC communication functions, can guarantee the fast reading and writing of data, suitable for driving early warning system, when reducing reaction
Between.
Description of the drawings
Fig. 1 is the system construction drawing of the present invention.
Specific embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, with reference to the accompanying drawings and embodiments, it is right
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.
The present invention provides a kind of based on infrared spectrum imaging and the pedestrian target prompting for the treatment of technology and anti-collision warning system
System is, it can be achieved that boisterous accurate pedestrian's early warning.This system filters out the pedestrian in infrared image using Adaboost algorithm
Target further detects the pedestrian target approached according to target approaches algorithm, provides pedestrian target and approaches prompting, when detecting
Have pedestrian target to steering direction side cross when, algorithm is crossed according to pedestrian target and judges pedestrian target state, and for not
With degree of danger provide and different cross early warning.
The pedestrian target that the Adaboost algorithm is filtered out in infrared image includes the following steps：
(1) training process
Collect positive sample collection and negative sample collection, the positive sample collection is the pictures comprising pedestrian target, the negative sample
This collection is the pictures not comprising pedestrian target；Extract Feature Descriptor respectively to sample each in sample set, and these are special
Sign description is used as training data, is trained classification by the method for minimal error rate, obtains several Weak Classifiers, according to
AdaBoost algorithm combinations Weak Classifier obtains strong classifier, then obtains cascade classifier, the cascade by multiple strong classifiers
Grader can effectively identify the Feature Descriptor of pedestrian's target signature, the Feature Descriptor include Haar-like rectangular characteristics,
The combination of one or more of EOH features, HOG features.
(2) detection process
Subimage block to be detected is proposed from infrared image；Calculate corresponding Feature Descriptor in subimage block；Utilize grade
Connection grader screens the Feature Descriptor, obtains including the subimage block of pedestrian target feature；Pedestrian's mesh will be included
The subimage block for marking feature merges fusion, obtains pedestrian target.
Algorithm and approximate algorithm are being crossed based on Adboost algorithms, the first approximate algorithm is：Detection becomes in the time
The amplification number of pedestrian target during first threshold is turned to, if amplification number is more than three times, judges the pedestrian target for danger
Otherwise dangerous target excludes the pedestrian target, the area for being enlarged into image shared by pedestrian target becomes larger, and wherein first threshold is
15~20 frames fully consider the state and collision probability of target in reality, can reduce first threshold when speed is very fast, speed compared with
Increase first threshold when slow, if first threshold is excessive may to lead to early warning not in time, if too small may cause early warning inaccurate.
Second of approximate algorithm and its subsequent algorithm that crosses are：When Δ t is equal to first threshold, if | Δ xt| more than
Two threshold values then judge the pedestrian target to cross target, if | Δ yt| more than third threshold value and | kt|>|kt-1|, then judge the pedestrian
Target crosses target to approach, and obtains the first danger early warning signal；Otherwise entering S122, the Δ t is the time change of t frames,
ΔxtFor the displacement of t frame one skilled in the art's target left and right directions, Δ ytFor the displacement in the front-back direction of t frame one skilled in the art's targets, the ktFor t
The displacement slope of frame one skilled in the art's target, kt-1For the displacement slope of (t-1) frame one skilled in the art's target, kt=Δ yt/Δxt, kt-1=Δ
yt-1/Δxt-1, the Δ xt-1For the displacement of (t-1) frame one skilled in the art's target left and right directions, Δ yt-1For (t-1) frame one skilled in the art's target
Displacement in the front-back direction；
When Δ t is equal to first threshold, if crossing target Δ xtMore than the 4th threshold value, then judge that this crosses target as right horizontal stroke
Target is worn, obtains the second danger early warning signal；If cross target Δ xtFor negative and | Δ xt| more than the 4th threshold value, then judge to go
People's target crosses target for a left side, obtains third danger early warning signal；Otherwise the pedestrian target is excluded.
Wherein first threshold is 1~6, and second threshold is 1~3, and third threshold value is 1~3, and the 4th threshold value is 2~6, in vehicle
It can reduce first threshold when fast very fast, increase first threshold when speed is slower, early warning may be caused too late if first threshold is excessive
When, if too small may cause warning information excessive, inaccurately, the second to four threshold value is bigger, and early warning is stringenter reliable, early warning number
It is fewer, unnecessary pre-warning signal is advantageously reduced, suitable for region known to road conditions or veteran driver, only tight
Start early warning in the case of urgency, the second to four threshold value is bigger, and safety is stronger, but early warning number increases simultaneously, will suitable for safety
It asks higher region or drives new hand, which fully considers the state and collision probability of target in reality, user Ke Gen
It needs to be adjusted according to itself.
Specific embodiment 1
It is pre- to provide a kind of pedestrian based on infrared thermal imaging and artificial side line technology for one embodiment according to the invention
Alert system, as shown in Figure 1, being acquired with infrared assembly to the infrared image of vehicle-periphery, master control system profit
Infrared image is detected with Adaboost cascade classifier detection algorithms, obtains pedestrian target, approached using pedestrian and
It crosses warning algorithm pedestrian is carried out approaching judgement and crosses judgement, obtains dangerous pedestrian target, and for different dangerous rows
People's target makes corresponding early warning, utilizes the air flow information in front of the array of pressure sensors collecting cart of artificial side line unit, institute
It states air flow information and includes the corresponding effective pressure of different location on the straight line of front, straight line is one in Dynamic IR image
Straight line, the array of pressure sensors include at least first pressure sensor, second pressure sensor, third pressure sensor,
Summarize the pressure information of different pressures sensor collection by collection plate, obtain multiple effective pressure on the straight line in front,
STM32 controllers according to the corresponding effective pressure of different location obtains dangerous pedestrian's mesh on straight line in Dynamic IR image
Target relative velocity and distance, by master control system according to the relative velocity and distance of dangerous pedestrian target, so as to further count
Collision time is calculated, realizes vehicle in detection and early warning of the extreme climate environments such as night to pedestrian.
Pedestrian's early warning system includes artificial side line unit, infrared assembly, master control system, prior-warning device, the people
The first input end of the output terminal connection master control system of work side line unit, the output terminal connection master control of the artificial side line unit
Second input terminal of system processed, the input terminal of the output terminal connection prior-warning device of the master control system,
The artificial side line unit is used to obtain pedestrian information according to the pressure information of different location on straight line, described
Infrared assembly is used to obtain the Dynamic IR image of visual field scene, and the master control system is used for the dynamic red according to visual field scene
Outer image obtains dangerous pedestrian target, and anti-collision warning signal is obtained according to the dangerous pedestrian target and pedestrian information, described pre-
Alarm device is used to carry out early warning according to pre-warning signal.
The artificial side line unit includes array of pressure sensors, collection plate, IC bus, STM32 controllers, institute
The output terminal for stating array of pressure sensors is connected by the input terminal of Analogous Integrated Electronic Circuits bus and collection plate, the collection plate
Output terminal connects the input terminal of the STM32 controllers, and the output terminal of the STM32 controllers is as the artificial side line unit
Output terminal,
The array of pressure sensors is used to be obtained according to the pressure information of different sensors multiple on the straight line of front
The pressure information of different location, the pressure information include flow direction and the flow strength of air fluid, and the collection plate is used
In summarizing the pressure information, the corresponding effective pressure of different location on straight line is obtained, the STM32 controllers are used for root
Obtain pedestrian information according to the corresponding effective pressure of different location on straight line, the pedestrian information include pedestrian speed, away from
From the pedestrian is front side pedestrian, and the speed is the relative velocity of front side pedestrian and vehicle, artificial side line unit of the invention
The function that high-speed computation and information transmits is realized by STM32 controllers, suitable for vehicle-mounted early warning system, when reducing reaction
Between, it is communicated by IC bus, which is suitable for multisystem and synchronizes ruling, can reduce different sensors
Information transmission time difference makes testing result relatively reliable, and pressure sensor uses CPS131 sensors, than general sensor more
It is sensitive, ensure to obtain the reliability of air flow information, support IIC communication functions, the fast reading and writing of data is can guarantee, suitable for vehicle
Early warning system is carried, reduces the reaction time.
The array of pressure sensors is set along vehicle airflow pressure trace, including at least three be installed at automobile grille
A pressure above sensor.
The prior-warning device includes the several of vehicle-mounted screen, light bar prior-warning device, one kind in buzzer prior-warning device or parallel connection
Kind.
The method that the infrared assembly obtains the Dynamic IR image of visual field scene：Utilize multiple infrared camera collecting vehicles
Surrounding scene image, and image is pre-processed, obtain the Dynamic IR image of visual field scene.
Preprocess method：According to the needs of data format, the image of acquisition is converted for infrared image around collection vehicle
Histogram equalization operation is carried out into single channel gray level image, and to image, reduces illumination and the influence of background, and to image
Size is suitably adjusted, the estimation of pedestrian may collide in the realtime graphic based on acquisition location information, to partly not
It is likely to occur pedestrian and pedestrian is not at the image of danger zone and is not counted in detection, pedestrian is only detected in smaller range, is subtracted
Few image procossing area so as to reduce data processing amount, improves algorithm practicability.It provides a kind of based on infrared thermal imaging and people
Pedestrian's method for early warning of work side line technology adopts the obstructions chart picture of vehicle periphery with the method for infrared thermal imaging
Collection, is detected pedestrian using Adaboost cascade classifier detection algorithms, warning algorithm is approached and crossed using pedestrian
Pedestrian is carried out approaching judgement and crosses judgement, and corresponding early warning is made for different pedestrian's states, utilizes artificial side line
Technology calculates the relative velocity and distance of pedestrian, so as to further calculate collision time, vehicle is realized in evils such as nights
Detection and early warning of the bad climatic environment to pedestrian.
The method of work of pedestrian's early warning system based on infrared imaging, includes the following steps：
S1：Obtain visual field scene Dynamic IR image, and according to the Dynamic IR image obtain pedestrian target and its
Time frame dynamic, the time frame dynamic are that change in location and area of the pedestrian target in the Dynamic IR image different frame become
Change；
S2：Dangerous pedestrian target is obtained according to pedestrian target and its time frame dynamic；
S3：The air flow information influenced by risk object is obtained, the air flow information includes being located at multiple on straight line
Effective pressure obtains the speed and distance of risk object according to the air flow information, is obtained according to the speed of risk object and distance
Collision time is obtained, and anti-collision warning signal is obtained according to the collision time.
The step S1 includes：
S11：The infrared picture comprising pedestrian target is obtained as positive sample collection；
S12：The infrared picture without pedestrian target is obtained as negative sample collection；
S13：According to positive and negative sample set, training obtains the Adaboost cascade classifiers with pedestrian's feature recognition；
S14：The infrared image that Adaboost cascade classifiers are obtained according to infrared assembly obtains pedestrian target and its every
Variation on one frame infrared image.
The method of acquisition pedestrian target is specially in S14：
S141：By the infrared image equal area partition of current time frame in infrared image at least two subimage block；
S142：The subimage block is detected by Adaboost cascade classifiers, is rejected special without pedestrian target
The subimage block of sign retains the subimage block for including pedestrian target feature；
S143：By the further equal area partition of each subimage block of reservation at least two subimage block, pass through
Adaboost cascade classifiers are detected the subimage block finally divided, reject the subgraph without pedestrian target feature
As block, retain the subimage block for including pedestrian target feature；
S144：Judge whether the number divided is more than preset cycle-index, is to stop dividing, what fusion finally retained
Subimage block obtains pedestrian target, otherwise returns to S143.
The step S2 includes：The amplification time of master control system detection pedestrian target in the time of 20 frames
Number if amplification number is more than three times, judges that the pedestrian target for dangerous pedestrian target, otherwise excludes the pedestrian target, described
The area for being enlarged into image shared by pedestrian target becomes larger.
Preliminary experiment is further included before the step S3：The distance for controlling pedestrian is constant, and the speed of multigroup pedestrian is set to obtain institute
Effective pressure of the position of pedestrian's correspondence is stated, the first relational model is obtained according to the speed of pedestrian and effective pressure；
The speed for controlling pedestrian is constant, and the distance of multigroup pedestrian is set to obtain the effective of different location on the straight line
The difference of pressure obtains the second relational model according to the distance of pedestrian and the difference of effective pressure；
The method that effective pressure is obtained in step S3 is：
S31：The original pressure of different location and benchmark wind pressure on the straight line of front are obtained, the benchmark wind pressure is certainly
The pressure that right wind generates；
S32：Benchmark wind pressure is subtracted from the original pressure, obtains effective pressure.
The dangerous speed of pedestrian target is obtained according to the air flow information in step S3 and the method for distance is：
S33：Dangerous pedestrian is obtained according to effective pressure of the position of dangerous pedestrian target correspondence and the first model
The speed of target obtains dangerous pedestrian target according to the difference of effective pressure of different location on straight line and the second model
Distance, the speed of the danger pedestrian target is the relative velocity of dangerous pedestrian target and vehicle.
Specific embodiment 2
A kind of method of work of pedestrian's early warning system based on infrared imaging, includes the following steps：
S1：Obtain visual field scene Dynamic IR image, and according to the Dynamic IR image obtain pedestrian target and its
Time frame dynamic, the time frame dynamic are that change in location and area of the pedestrian target in the Dynamic IR image different frame become
Change；
S2：Dangerous pedestrian target is obtained according to pedestrian target and its time frame dynamic；
S3：The air flow information influenced by risk object is obtained, the air flow information includes more on the straight line in front
A effective pressure obtains the speed and distance of risk object according to the air flow information, according to the speed and distance of risk object
Collision time is obtained, and anti-collision warning signal is obtained according to the collision time.
The step S1 includes：
S11：The infrared picture comprising pedestrian target is obtained as positive sample collection；
S12：The infrared picture of barrier without pedestrian target is obtained as negative sample collection；
S13：According to positive and negative sample set, training obtains the Adaboost cascade classifiers with pedestrian's feature recognition；
S14：The infrared image obtained with Adaboost cascade classifiers according to infrared assembly, obtain pedestrian target and its
Variation on each frame infrared image.
The method of acquisition pedestrian target is specially in S14：
S141：By the infrared image equal area partition of current time frame in infrared image at least two subimage block；
S142：The subimage block is detected by Adaboost cascade classifiers, is rejected special without pedestrian target
The subimage block of sign retains the subimage block for including pedestrian target feature；
S143：By the further equal area partition of each subimage block of reservation at least two subimage block, pass through
Adaboost cascade classifiers are detected the subimage block finally divided, reject the subgraph without pedestrian target feature
As block, retain the subimage block for including pedestrian target feature；
S144：Judge whether the number divided is more than preset cycle-index, is to stop dividing, what fusion finally retained
Subimage block obtains pedestrian target, otherwise returns to S143.
The step S2 includes the following steps：
S21：Using the upper left corner of the infrared image as origin, pixel is unit, is positive direction of the x-axis towards the right side, downward for y just
Coordinate system is established in direction, in Δ t=3 frames, if | Δ xt|>1 judges the target to cross pedestrian target, if | Δ yt|>1 and |
kt|>|kt-1|, then judge that the target crosses pedestrian target to approach, obtain the first danger early warning signal；Otherwise enter S22, institute
State the time change that Δ t is t frames, Δ xtAbscissa for t frame one skilled in the art's targets changes, Δ ytVertical seat for t frame one skilled in the art's targets
Mark variation, the ktFor the acceleration slope of t frame one skilled in the art's targets, kt-1For the acceleration slope of (t-1) frame one skilled in the art's target, kt=Δ
yt/Δxt, kt-1=Δ yt-1/Δxt-1；
S22：Track of the pedestrian target in the time of Δ t=1 frames is crossed in judgement, if Δ xt>3, then judge the traverses rows
People's target crosses pedestrian target for the right side, obtains the second danger early warning signal；If Δ xt<- 3, then judge that pedestrian target is crossed for a left side
Pedestrian target obtains third danger early warning signal；Otherwise the pedestrian target is excluded.
Preliminary experiment is further included before the step S3：The distance for controlling pedestrian is constant, and the speed of multigroup pedestrian is set to obtain institute
Effective pressure of the position of pedestrian's correspondence is stated, the first relational model is obtained according to the speed of pedestrian and effective pressure；
The speed for controlling pedestrian is constant, and the distance of multigroup pedestrian is set to obtain the effective of different location on the straight line
The difference of pressure obtains the second relational model according to the distance of pedestrian and the difference of effective pressure；
The method that effective pressure is obtained in step S3 is：
S31：The original pressure of different location and benchmark wind pressure on the straight line of front are obtained, the benchmark wind pressure is certainly
The pressure that right wind generates；
S32：Benchmark wind pressure is subtracted from the original pressure, obtains effective pressure.
The dangerous speed of pedestrian target is obtained according to the air flow information in step S3 and the method for distance is：
S33：It is obtained according to effective pressure of the position of dangerous pedestrian target correspondence and the first relational model dangerous
The speed of pedestrian target obtains dangerous according to the difference of effective pressure of different location on straight line and the second relational model
The distance of pedestrian target, the speed of the danger pedestrian target is the relative velocity of dangerous pedestrian target and vehicle.
As it will be easily appreciated by one skilled in the art that the foregoing is merely illustrative of the preferred embodiments of the present invention, not to
The limitation present invention, all any modification, equivalent and improvement made all within the spirits and principles of the present invention etc., should all include
Within protection scope of the present invention.
Claims (10)
1. a kind of pedestrian's method for early warning based on infrared imaging, which is characterized in that including：
S1：The Dynamic IR image of visual field scene is obtained, and risk object is obtained according to the Dynamic IR image；
S2：The air flow information influenced by risk object is obtained, the air flow information includes multiple effective on the straight line of front
Pressure obtains the speed and distance of risk object according to the air flow information, according to the speed of risk object and apart from being touched
It hits the time, and anti-collision warning signal is obtained according to the collision time.
2. pedestrian's method for early warning based on infrared imaging as described in claim 1, which is characterized in that effective pressure is by original
Beginning pressure subtracts benchmark wind pressure and obtains.
3. pedestrian's method for early warning based on infrared imaging as claimed in claim 2, which is characterized in that according to Dynamic IR image
Obtain risk object method be：
S11：Pedestrian target and its time frame dynamic are obtained according to the Dynamic IR image；
S12：Risk object is obtained according to pedestrian target and its time frame dynamic.
4. pedestrian's method for early warning based on infrared imaging as claimed in claim 3, which is characterized in that root in the step S12
It is according to the method that pedestrian target and its time frame dynamic obtain risk object：Detection pedestrian's mesh when Δ t is equal to first threshold
The amplification number of accumulation is marked, if amplification number is more than second threshold, judges the pedestrian target for risk object, obtains the first danger
Otherwise dangerous pre-warning signal excludes the pedestrian target, time changes of the Δ t for t frames, t>1, the pedestrian target of being enlarged into
Area becomes larger.
5. pedestrian's method for early warning based on infrared imaging as claimed in claim 3, which is characterized in that root in the step S12
Include according to the method that pedestrian target and its time frame dynamic obtain risk object：
S121：When Δ t is equal to first threshold, if | Δ xt| the pedestrian target is then judged more than second threshold to cross target, if
|Δyt| more than third threshold value and | kt|>|kt-1|, then judge that the pedestrian target crosses target to approach, obtain the first danger early warning
Signal；Otherwise enter S122, time changes of the Δ t for t frames, t>1, Δ xtPosition for t frame one skilled in the art's target left and right directions
It moves, Δ ytFor the displacement in the front-back direction of t frame one skilled in the art's targets, the ktFor the displacement slope of t frame one skilled in the art's targets, kt-1For (t-
1) the displacement slope of frame one skilled in the art target, kt=Δ yt/Δxt, kt-1=Δ yt-1/Δxt-1, the Δ xt-1For (t-1) frame expert
The displacement of people's target left and right directions, Δ yt-1For the displacement in the front-back direction of (t-1) frame one skilled in the art's target；
S122：When Δ t is equal to first threshold, if crossing target Δ xtMore than the 4th threshold value, then judge that this crosses target as right horizontal stroke
Target is worn, obtains the second danger early warning signal；If cross target Δ xtFor negative and | Δ xt| more than the 4th threshold value, then judge to go
People's target crosses target for a left side, obtains third danger early warning signal；Otherwise the pedestrian target is excluded.
6. pedestrian's method for early warning based on infrared imaging as described in claim 4 or 5, which is characterized in that wrapped before step S11
It includes：There is the Adaboost cascade classifiers of pedestrian target feature recognition with Adaboost algorithm acquisition；In step s 11
It is described according to Dynamic IR image obtain pedestrian target method be：With the Adaboost cascade classifiers from the dynamic red
Pedestrian target is extracted in outer image.
A kind of 7. pedestrian's method for early warning based on infrared imaging as claimed in claim 6, which is characterized in that the extraction trip
People's mesh calibration method is specially：
S111：By the infrared image equal area partition of current time frame in infrared image at least two subimage block；
S112：The subimage block is detected with the Adaboost cascade classifiers, is rejected without pedestrian target feature
Subimage block, retain and include the subimage block of pedestrian target feature；
S113：By the further equal area partition of each subimage block of reservation at least two subimage block, cascaded with Adaboost
Grader is detected the subimage block finally divided, rejects the subimage block without pedestrian target feature, retains packet
The subimage block of the feature containing pedestrian target；
S114：Judge whether the number divided is more than preset cycle-index, is to stop dividing, and merges the subgraph finally retained
As block acquisition pedestrian target, S113 is otherwise returned.
8. a kind of pedestrian's early warning system based on infrared imaging, which is characterized in that including artificial side line unit, infrared assembly, master
Control system, prior-warning device, the first input end of the output terminal connection master control system of the artificial side line unit are described infrared
Second input terminal of the output terminal connection master control system of component, the output terminal of the master control system connect the defeated of prior-warning device
Enter end,
The infrared assembly is used to obtain the Dynamic IR image of visual field scene, and the master control system is used for according to visual field scene
Dynamic IR image obtain risk object, the artificial side line unit be used for according to multiple effective pressures on the straight line of front
The strong speed and distance for obtaining risk object, the master control system are further obtained according to the speed of the risk object with distance
Obtain anti-collision warning signal.
9. a kind of artificial side line unit for early warning system according to any one of claims 8, which is characterized in that including pressure sensor
Array, IC bus, STM32 controllers, the output terminal of the array of pressure sensors are connected by IC bus
The input terminal of the collection plate, the output terminal of the collection plate connect the input terminal of the STM32 controllers, the STM32 controls
Output terminal of the output terminal of device processed as the artificial side line unit,
The array of pressure sensors is used to obtain multiple having on the straight line of front according to the pressure information of different sensors
Imitate pressure, the STM32 controllers be used for according to multiple effective pressure on the front straight line obtain pedestrians speed,
Distance.
10. side line unit as claimed in claim 9 artificial, which is characterized in that the array of pressure sensors is along front air-flow
Pressure trace is set, including three pressure above sensors.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711251575.0A CN108256418B (en) | 2017-12-01 | 2017-12-01 | Pedestrian early warning method and system based on infrared imaging |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711251575.0A CN108256418B (en) | 2017-12-01 | 2017-12-01 | Pedestrian early warning method and system based on infrared imaging |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108256418A true CN108256418A (en) | 2018-07-06 |
| CN108256418B CN108256418B (en) | 2022-01-21 |
Family
ID=62722370
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711251575.0A Active CN108256418B (en) | 2017-12-01 | 2017-12-01 | Pedestrian early warning method and system based on infrared imaging |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108256418B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109145980A (en) * | 2018-08-16 | 2019-01-04 | 大连民族大学 | Magnetic force probabilistic algorithm and system for road pedestrian classification |
| CN111368883A (en) * | 2020-02-21 | 2020-07-03 | 浙江大华技术股份有限公司 | Obstacle avoidance method based on monocular camera, computing device and storage device |
| CN112632686A (en) * | 2020-12-30 | 2021-04-09 | 天津大学 | Early warning method for collision in offshore pile sinking construction process |
Citations (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060244632A1 (en) * | 2005-05-02 | 2006-11-02 | Mr. James Corcoran | Automatic Vehicle Collison Avoidance and Warning System |
| CN101497342A (en) * | 2009-01-13 | 2009-08-05 | 杭州创联电子技术有限公司 | System integrated supervision method and system thereof of self-movement special equipment safe operation |
| CN101826265A (en) * | 2010-04-29 | 2010-09-08 | 大连海事大学 | Green driving behavior training device |
| CN102381481A (en) * | 2011-08-01 | 2012-03-21 | 四川省科学城久利科技实业有限责任公司 | Unmanned aerial vehicle-mounted laser radar system |
| CN102627093A (en) * | 2011-02-04 | 2012-08-08 | 丰田自动车株式会社 | Vehicle side airbag apparatus |
| CN102673545A (en) * | 2011-03-16 | 2012-09-19 | 富士重工业株式会社 | Vehicle driving support apparatus |
| CN103761832A (en) * | 2014-01-22 | 2014-04-30 | 安徽科赫信息技术有限公司 | Wearable human body collision early warning and protection device and early warning and protection method thereof |
| CN203832189U (en) * | 2013-01-08 | 2014-09-17 | 福特全球技术公司 | Self-adaptation crash height adjusting system using active suspension |
| US20150066283A1 (en) * | 2013-09-04 | 2015-03-05 | Amazon Technologies, Inc. | System and method for transporting personnel within an active workspace |
| CN105291974A (en) * | 2015-11-11 | 2016-02-03 | 深圳市精能奥天导航技术有限公司 | Active warning system and method for preventing following vehicle collision |
| CN105333988A (en) * | 2015-11-25 | 2016-02-17 | 中国海洋大学 | Artificial lateral line pressure detection method |
| CN205209685U (en) * | 2015-11-25 | 2016-05-04 | 中国海洋大学 | Artifical side line pressure measurement |
| CN205428125U (en) * | 2016-03-15 | 2016-08-03 | 姚冰 | Bus safety monitoring system |
| CN106156725A (en) * | 2016-06-16 | 2016-11-23 | 江苏大学 | A kind of method of work of the identification early warning system of pedestrian based on vehicle front and cyclist |
| CN106203268A (en) * | 2016-06-28 | 2016-12-07 | 成都之达科技有限公司 | vehicle state monitoring method based on infrared identification |
| CN206003179U (en) * | 2016-08-31 | 2017-03-08 | 王文礼 | A kind of multifunctional vehicle accident early warning device |
| CN106683489A (en) * | 2017-03-01 | 2017-05-17 | 青岛海澄知识产权事务有限公司 | Early-warning device for avoiding secondary traffic accidents |
| CN106989758A (en) * | 2017-05-24 | 2017-07-28 | 北京大学 | Neighbouring underwater robot based on artificial lateral-line system closely cognitive method |
| CN107145105A (en) * | 2017-05-24 | 2017-09-08 | 北京大学 | A kind of artificial lateral-line system based on pressure transducer array |
| CN107195201A (en) * | 2017-07-18 | 2017-09-22 | 青岛厚科信息工程有限公司 | A chain of collision early warning system is prevented on highway |
| CN107368095A (en) * | 2017-08-29 | 2017-11-21 | 中国人民解放军国防科技大学 | Air anti-collision system and method for small-sized fixed-wing unmanned aerial vehicle |
| CN107364459A (en) * | 2017-06-28 | 2017-11-21 | 北京康拓红外技术股份有限公司 | A kind of rolling stock row inspection intelligent robot operating system and method |
-
2017
- 2017-12-01 CN CN201711251575.0A patent/CN108256418B/en active Active
Patent Citations (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060244632A1 (en) * | 2005-05-02 | 2006-11-02 | Mr. James Corcoran | Automatic Vehicle Collison Avoidance and Warning System |
| CN101497342A (en) * | 2009-01-13 | 2009-08-05 | 杭州创联电子技术有限公司 | System integrated supervision method and system thereof of self-movement special equipment safe operation |
| CN101826265A (en) * | 2010-04-29 | 2010-09-08 | 大连海事大学 | Green driving behavior training device |
| CN102627093A (en) * | 2011-02-04 | 2012-08-08 | 丰田自动车株式会社 | Vehicle side airbag apparatus |
| CN102673545A (en) * | 2011-03-16 | 2012-09-19 | 富士重工业株式会社 | Vehicle driving support apparatus |
| CN102381481A (en) * | 2011-08-01 | 2012-03-21 | 四川省科学城久利科技实业有限责任公司 | Unmanned aerial vehicle-mounted laser radar system |
| CN203832189U (en) * | 2013-01-08 | 2014-09-17 | 福特全球技术公司 | Self-adaptation crash height adjusting system using active suspension |
| US20150066283A1 (en) * | 2013-09-04 | 2015-03-05 | Amazon Technologies, Inc. | System and method for transporting personnel within an active workspace |
| CN103761832A (en) * | 2014-01-22 | 2014-04-30 | 安徽科赫信息技术有限公司 | Wearable human body collision early warning and protection device and early warning and protection method thereof |
| CN105291974A (en) * | 2015-11-11 | 2016-02-03 | 深圳市精能奥天导航技术有限公司 | Active warning system and method for preventing following vehicle collision |
| CN105333988A (en) * | 2015-11-25 | 2016-02-17 | 中国海洋大学 | Artificial lateral line pressure detection method |
| CN205209685U (en) * | 2015-11-25 | 2016-05-04 | 中国海洋大学 | Artifical side line pressure measurement |
| CN205428125U (en) * | 2016-03-15 | 2016-08-03 | 姚冰 | Bus safety monitoring system |
| CN106156725A (en) * | 2016-06-16 | 2016-11-23 | 江苏大学 | A kind of method of work of the identification early warning system of pedestrian based on vehicle front and cyclist |
| CN106203268A (en) * | 2016-06-28 | 2016-12-07 | 成都之达科技有限公司 | vehicle state monitoring method based on infrared identification |
| CN206003179U (en) * | 2016-08-31 | 2017-03-08 | 王文礼 | A kind of multifunctional vehicle accident early warning device |
| CN106683489A (en) * | 2017-03-01 | 2017-05-17 | 青岛海澄知识产权事务有限公司 | Early-warning device for avoiding secondary traffic accidents |
| CN106989758A (en) * | 2017-05-24 | 2017-07-28 | 北京大学 | Neighbouring underwater robot based on artificial lateral-line system closely cognitive method |
| CN107145105A (en) * | 2017-05-24 | 2017-09-08 | 北京大学 | A kind of artificial lateral-line system based on pressure transducer array |
| CN107364459A (en) * | 2017-06-28 | 2017-11-21 | 北京康拓红外技术股份有限公司 | A kind of rolling stock row inspection intelligent robot operating system and method |
| CN107195201A (en) * | 2017-07-18 | 2017-09-22 | 青岛厚科信息工程有限公司 | A chain of collision early warning system is prevented on highway |
| CN107368095A (en) * | 2017-08-29 | 2017-11-21 | 中国人民解放军国防科技大学 | Air anti-collision system and method for small-sized fixed-wing unmanned aerial vehicle |
Non-Patent Citations (2)
| Title |
|---|
| ZHENG, P等: "The effect of sensor errors on the performance of collision warning systems", 《INTELLIGENT TRANSPORTATION SYSTEMS IEEE》 * |
| 刘竞阳: "基于红外测距传感器的移动机器人路径规划系统设计", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109145980A (en) * | 2018-08-16 | 2019-01-04 | 大连民族大学 | Magnetic force probabilistic algorithm and system for road pedestrian classification |
| CN109145980B (en) * | 2018-08-16 | 2020-09-22 | 大连民族大学 | Magnetic force probability algorithm and system for road pedestrian classification |
| CN111368883A (en) * | 2020-02-21 | 2020-07-03 | 浙江大华技术股份有限公司 | Obstacle avoidance method based on monocular camera, computing device and storage device |
| CN111368883B (en) * | 2020-02-21 | 2024-01-19 | 浙江大华技术股份有限公司 | Obstacle avoidance method based on monocular camera, computing device and storage device |
| CN112632686A (en) * | 2020-12-30 | 2021-04-09 | 天津大学 | Early warning method for collision in offshore pile sinking construction process |
| CN112632686B (en) * | 2020-12-30 | 2022-07-22 | 天津大学 | Early warning method for collision in offshore pile sinking construction process |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108256418B (en) | 2022-01-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107031623B (en) | A kind of road method for early warning based on vehicle-mounted blind area camera | |
| CN110992683B (en) | Dynamic image perception-based intersection blind area early warning method and system | |
| CN105512623B (en) | Based on multisensor travelling in fog day vision enhancement and visibility early warning system and method | |
| CN107284355B (en) | A kind of Safety vehicle door unlatching processing method and system | |
| CN105922990B (en) | A kind of vehicle environmental based on high in the clouds machine learning perceives and control method | |
| CN105844222B (en) | The front vehicles collision warning systems and method of view-based access control model | |
| US10699567B2 (en) | Method of controlling a traffic surveillance system | |
| CN108596058A (en) | Running disorder object distance measuring method based on computer vision | |
| CN108263279A (en) | The pedestrian detection and pedestrian impact avoiding device and method integrated based on sensor | |
| KR101891460B1 (en) | Method and apparatus for detecting and assessing road reflections | |
| CN106240458A (en) | A kind of vehicular frontal impact method for early warning based on vehicle-mounted binocular camera | |
| CN104573646A (en) | Detection method and system, based on laser radar and binocular camera, for pedestrian in front of vehicle | |
| CN106485233A (en) | Drivable region detection method, device and electronic equipment | |
| CN106537180A (en) | Method for mitigating radar sensor limitations with video camera input for active braking for pedestrians | |
| CN110065494A (en) | A kind of vehicle collision avoidance method based on wheel detection | |
| CN105930787A (en) | Vehicle door opening early-warning method | |
| EP1671216A2 (en) | Moving object detection using low illumination depth capable computer vision | |
| CN109696676B (en) | Effective obstacle target determination method and device and vehicle | |
| CN107274678B (en) | A kind of night vehicle flowrate and model recognizing method based on Kinect | |
| CN204309672U (en) | A kind of anti-car rear-end prior-warning device based on image recognition | |
| CN108256418A (en) | A kind of pedestrian's method for early warning and system based on infrared imaging | |
| CN112101316B (en) | Target detection method and system | |
| CN108108680A (en) | A kind of front vehicle identification and distance measuring method based on binocular vision | |
| CN109389060A (en) | A kind of vehicle week anti-collision warning method of view-based access control model | |
| CN113706901B (en) | Intelligent accident prevention and control and early warning system for entrance section of expressway tunnel |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP03 | Change of name, title or address | ||
| CP03 | Change of name, title or address |
Address after: Building 4, No. 8 Deyi Road, Donghu New Technology Development Zone, Wuhan City, Hubei Province, 430205 Patentee after: Xuanyuan Idrive Technology (SHENZHEN) Co.,Ltd. Country or region after: China Address before: 518000 floor 12, building 1, yongxinhui, No. 4078, Dongbin Road, Nanshan street, Nanshan District, Shenzhen, Guangdong Patentee before: XUANYUAN IDRIVE TECHNOLOGY (SHENZHEN) CO.,LTD. Country or region before: China |