CN110312079A - Image collecting device and its application system - Google Patents
Image collecting device and its application system Download PDFInfo
- Publication number
- CN110312079A CN110312079A CN201810230565.7A CN201810230565A CN110312079A CN 110312079 A CN110312079 A CN 110312079A CN 201810230565 A CN201810230565 A CN 201810230565A CN 110312079 A CN110312079 A CN 110312079A
- Authority
- CN
- China
- Prior art keywords
- image
- unit
- collecting device
- image collecting
- exposure
- 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
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/64—Three-dimensional objects
-
- 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
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
- G06V40/166—Detection; Localisation; Normalisation using acquisition arrangements
-
- 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
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/70—Circuitry for compensating brightness variation in the scene
- H04N23/73—Circuitry for compensating brightness variation in the scene by influencing the exposure time
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Signal Processing (AREA)
- Studio Devices (AREA)
Abstract
The invention discloses a kind of multi-optical spectrum image collecting devices, and using object identification system, panorama camera system and the automobile navigation obstacle avoidance system of the device.Image collecting device according to the present invention, comprising: sensing unit, including imaging sensor and camera lens；Lighting unit, the driving circuit including luminescent device and driving luminescent device；Control unit is connect with sensing unit and lighting unit, and the exposure time series and exposure gain of the image sensor exposure for generating control sensing unit, and the illumination timing of control lighting unit；Wherein, within an exposure time series period, at least two exposed frame yield values are identical, and in the case where luminescent device is closed, at least two exposed frame yield values are different.The present invention can acquire multispectral image using monocular cam, and by improving the adaptability to illumination, and can export the depth map of 3D, being widely applied so as to provide to image exposure gain control and/or inter frame image processing.
Description
Technical field
The present invention relates to Image Acquisition field, in particular to a kind of image collecting device, object identification system, panorama camera
System and automobile navigation obstacle avoidance system.
Background technique
The monocular camera of image collecting device at present can only acquire the image or video of a spectrum in a period.
For example, general camera relies on ambient light illumination, visible light (RGB) image or video are acquired；The monitoring camera day and night switched, just
Under ordinary light shines, it is arranged to RGB mode, shoots the image or video of RGB spectral coverage, and in dark conditions, switch to near-infrared
(NIR needs luminescent device to illuminate) screening-mode.
In addition, the camera illuminated using near-infrared active luminescent device, such as three-dimensional imaging camera, it will receive environment light source
Such as the influence of near-infrared ingredient strong in sunlight, so that the radiation direction of near-infrared image and intensity receive influence, thus difficult
To reach expected camera effect.
In short, existing monocular camera can not acquire the image or video of multiple spectrum in the same period, and using close red
The camera of outer active light source, will receive the interference of environment light especially sunburst, to reduce output image quality, limit
Camera effective use scope.
Summary of the invention
The present invention provides a kind of images for being able to use monocular camera acquisition time multispectral image in a short time to adopt
Acquisition means, object identification system, panorama camera system and automobile navigation obstacle avoidance system.It can be under dark dim light and sunburst
Stablize output multispectral image.
To achieve the goals above, present invention employs following technical solutions:
According to one aspect of the present invention, a kind of image collecting device is provided, comprising: sensing unit, including image pass
Sensor and camera lens；Lighting unit, the driving circuit of luminescent device and driving luminescent device including at least one spectral band；Control
Unit processed, connect with sensing unit and lighting unit, and the exposure of the image sensor exposure for generating control sensing unit
Timing and exposure gain, and the illumination timing of control lighting unit；Wherein, within an exposure time series period, at least two frames
Exposure gain value is identical, and in the case where luminescent device is closed, at least two exposed frame yield values are different.
Preferably, image collecting device further includes the photosensitive unit for detecting environment illumination intensity, control unit according to
The intensity of illumination of photosensitive unit detection controls the exposure gain value of the framing inside an exposure time series period.
Preferably, sensing unit further includes filter unit, and filter unit is more bandpass filters or is arranged in the camera lens
Slide on more bandpass filter plated films；Filter unit allows the multi light spectrum hands light of luminescent device to pass through, and allows environment
The light of partial spectrum wave band passes through in illumination.
Preferably, image collecting device further includes for calculating the figure acquired when the luminescent device of sensing unit acquisition is opened
The calculation process for the difference image between image that picture acquires when closing with luminescent device.
Preferably, lighting unit further includes photo etched mask, for penetrating the photo etched mask under luminescent device irradiation,
Structure hot spot is formed on acquisition target, so that sensing unit obtains structure light spot image after carrying out shooting, collecting to acquisition target.
Preferably, image collecting device further includes first of the three dimensional depth of calculating acquisition target from structure light spot image
Three dimensional depth computing unit.
Preferably, image collecting device further includes carrying out the elimination of structure hot spot to structure light image, and rebuild complete image
Reconstruction unit.
Preferably, image collecting device further includes the second three dimensional depth meter that three dimensional depth is calculated using pulsed-beam time-of-flight methods
Calculate unit.
Preferably, a kind of object identification system, including image collecting device above-mentioned and object detection unit, the object
Detection unit detection object and obtains the image of institute's detection object from described image acquisition device acquired image.
Preferably, object identification system further includes the feature extraction unit that feature is extracted from the image of institute's detection object.
Preferably, object identification system further includes the classification that institute's detection object is judged from the feature that feature extraction unit is extracted
Object classification unit.
The present invention also provides a kind of panorama camera system, including the multiple images acquisition device that is shot to different perspectives and
The image mosaic system of image mosaic is carried out to the image of the different perspectives of multiple images acquisition device shooting；Multiple images acquisition
At least one of device is image collecting device above-mentioned.
The present invention also provides a kind of automobile navigation obstacle avoidance system, including image collecting device, Context awareness unit and aobvious
Show that unit, image collecting device include at least one image collecting device above-mentioned；Context awareness unit, to image collecting device
Ambient image collected is analyzed, and the distance of barrier in environment, the object in detection or environment-identification are measured；Display is single
Member is connect with image collecting device and Context awareness unit, and the image or environment for showing image acquisition device are known
The analysis result of other unit.
In conclusion image collecting device according to the present invention, by the way that the luminescent device of at least one spectral band is arranged,
And illumination timing is generated by control unit, to drive the impulse waveform of luminescent device alternate emission difference spectral band, close
Close or continuously transmit one or more impulse waveforms of same spectral band.Sensing unit is in exposure time series and exposure gain
Under driving, the image of timesharing shooting environmental light source image or the irradiation of different-waveband light source, to use a monocular cam
Multispectral image can be collected, and by the control of exposure gain, to improve products application performance, expand product and answer
Use range.
Detailed description of the invention
The attached drawing constituted part of this application is used to provide further understanding of the present invention, schematic reality of the invention
It applies example and its explanation is used to explain the present invention, do not constitute improper limitations of the present invention.In the accompanying drawings:
Fig. 1 is the schematic illustration of image collecting device according to the present invention；
Fig. 2 is the exposure time series of image collecting device according to the present invention and the schematic diagram of illumination timing；
Fig. 3 is schematic diagram of the image acquisition device according to the present invention with/without 2 frame images of active illumination；
Fig. 4 is the wave band schematic diagram of the multiband bandpass filter of image collecting device according to the present invention；
Fig. 5 is the schematic illustration of object identification system according to the present invention；
Fig. 6 is the schematic illustration of object identification system according to the present invention；
Fig. 7 is the schematic illustration of the feature extraction unit of object identification system according to the present invention；
Fig. 8 is the structural schematic diagram of panorama camera system according to the present invention；
Fig. 9 is the schematic illustration of the image mosaic of panorama camera system according to the present invention；
Figure 10 is the schematic illustration of vehicle according to the invention navigation obstacle avoidance system.
Specific embodiment
To make the object, technical solutions and advantages of the present invention clearer, below in conjunction with attached drawing to embodiment party of the present invention
Formula is described in further detail.
As shown in Figure 1, the first embodiment of the present invention provides a kind of image collecting device, comprising: sensing unit 11, one
As for, including imaging sensor 111 and corresponding camera lens 112；Lighting unit 12, shining including at least one spectral band
The driving circuit 122 of device 121 and driving luminescent device；Control unit 13 is connect with sensing unit 11 and lighting unit 12, and
The exposure time series and exposure gain that imaging sensor 111 for generating control sensing unit exposes, and control lighting unit
12 illumination timing；Wherein, within an exposure time series period, at least two exposed frame yield values are identical, and close in luminescent device
In the case where closing, at least two exposed frame yield values are different.That is, control unit 13 is according to setting program in image acquisition process
Or exposure time series, exposure gain and illumination timing are sent as needed, so that sensing unit 11 and photo unit 12 is matched work,
And the image of imaging sensor is read, to realize the mode of different-waveband timesharing to realize adopting for multiband spectrum image
Collection, and control the reading and output of image after the exposure of imaging sensor.
According to illumination Timing driver luminescent device (active light source), the alternate emission in different periods is different for illumination driving 122
The impulse waveform of spectral band, the one or more impulse waveforms for closing or continuously transmitting same spectral band.That is, illumination
The illumination driving 122 of unit 12 generates pulse current according to illumination timing, and timesharing drives the hair of some or multiple spectral bands
Optical device 121 issues the illumination of corresponding wave band, is irradiated on subject A, and illumination is returned from subject A surface reflection, thoroughly
Camera lens 112 is crossed, is received by imaging sensor 111, is reached on imaging sensor 111, is converted into picture signal, and export shooting figure
Picture.When lighting unit 12 be multispectral irradiation when, monocular cam can acquisition time to multiple spectrum pictures, to solve
The problem of needing two or more camera (images that each camera shoots a spectral band) in the prior art.
During multi-optical spectrum image collecting, it is seen that external environment light source photograph can also be used in the illumination of spectrum (RGB) wave band
Bright, control unit closes all luminescent devices at this time.Preferably, as shown in Figure 1, image collecting device further includes for detecting ring
The photosensitive unit 131 of border intensity of illumination, so that exposure when sensing unit is shot using ambient lighting to image is controlled,
The intensity of illumination that i.e. control unit can be detected according to photosensitive unit 131 controls at least partly frame within an exposure time series period
Exposure gain value.
Specifically, as shown in connection with fig. 2, exposure gain is mutually matched with exposure time series and illumination timing, to control image
111 photoreceptor signal amplification factor of sensor (or exposure shutter duration), and it is aided with image processing algorithm, so that imaging sensor
111 acquired images expose suitable (neither overexposure, also not under-exposure) as far as possible, and image pixel value is in normal range (NR).
For example, output is the range of 8bits pixel value 0~255, generally the pixel of the area-of-interest after exposure and image procossing
Value control is being preferred (not overexposure, not deficient exposure) between 50~200.
In addition, the exposure gain value that can control partial frame is identical, portion according to environment illumination intensity and image procossing needs
The exposure gain value of framing is not identical.For example, when luminescent device closes, only acquires the image of environment light source, in order to full respectively
Sufficient image is shown and the requirement of image recognition, can be different by control exposure gain value；When needing at a few frame image synthesis
Reason, such as comparison, difference, can set identical for the exposure gain value of the image of these frames, to have benchmark, favorably
In raising treatment effect and precision.
It is to be appreciated that subject A can be face, human body, vehicle, house, trees, direction board, move in Fig. 1
Object etc..
Preferably, luminescent device 121 is laser or LED light emitting device, or the combination of the two, to emit
The light of different spectral bands, including infrared light or the ultraviolet light etc. for being not limited to visible light, different wave length.
Preferably, sensing unit 11 further includes filter unit, and filter unit is independent more bandpass filters or setting exists
More bandpass filter plated films on the slide of camera lens 112；Filter unit only allows the light of interested spectral band to pass through, from
And make imaging sensor 111 that can only acquire the image of interested spectral band.
The wave band schematic diagram of multiband bandpass filter as shown in Figure 4, including RGB visible light wave range, central wavelength
For the ultraviolet band of 380nm, one or more groups of the near infrared band of central wavelength 808nm, 850nm, 905nm, 940nm
It closes.The multiband band logical property of multiband bandpass filter makes in the spectral band and/or environment light of luminescent device 121
Rgb light spectrum wave band passes through, and the light energy of other non-interest wave bands is inhibited to reach imaging sensor 111, provides desired characteristic
Image that is more consistent, being more advantageous to subsequent image analysis and object identification.
During image collecting device 10 acquires multispectral image, firstly, the control illumination of control unit 13 driving 122, root
Make illumination driving 122 in the defined period according to illumination timing, the electric current of timesharing output regulation spectral band simultaneously controls current strength,
The luminescent device 121 of corresponding wave band is lighted, the pulsed light of output regulation spectral band shines；Secondly, the control sensing of control unit 13
Unit 11 and pulsed light shine sync pulse jamming, or slightly lag shooting after waiting luminescent device 121 sufficiently to light.Wherein, it controls
According to the program of setting, the illumination pulses difference timesharing for controlling multiple and different spectral bands issues unit 13, or makes 2 or more
A illumination pulses be simultaneously emitted by (as at the same make two infraluminescences of different central wavelengths, i.e. nirA+nirB)；And it controls
The output current intensity of illumination driving achievees the effect that subsequent image processing application is required.The invention enables using a camera shooting
Head can acquire multispectral image, for example the following timesharing combination multispectral image sequence of 9 frames is acquired in one second:
RGB-nir-nir-RGB-nir-nir-RGB-nir-nir；
The following timesharing combination multispectral image sequence of 9 frames can also be acquired:
RGB-nirA+nirB-nirA-RGB-nirA+nirB-nirA-RGB-nirA+nirB-nirA。
As shown in connection with fig. 3, image collecting device of the invention further includes 2 or more for calculating the acquisition of sensing unit 11
The processing unit of difference image between a (frame) image.I.e. having the image acquired under the conditions of luminescent device 121 (for photophore
The reflected energy of spectral band and being superimposed for ambient lighting energy reflected energy that part 121 issues) subtract no photograph of luminescent device 121
The image (only reflecting uncontrolled ambient lighting reflected energy) acquired under the conditions of penetrating；To suppress or eliminate uncontrollable environment light
According to adverse effect, only provide light source it is controlled, be more advantageous to the image of subsequent image analysis and object identification.
Fig. 3 gives the further implementation method of 2 frame video captures acquisition, wherein the first exposed frame opens actively
Bright pulse, the second frame rely only on ambient lighting exposure；Previous frame image and a later frame are subtracted each other, can be obtained difference image.
It is highly preferred that one can also be subtracted according to the image acquired under irradiation condition simultaneously using two kinds or more of pulsed lights
Kind or a variety of pulsed lights shine under the conditions of the mode of image that acquires obtain difference image, namely improve controllable pulsed light and shine
The accounting of the relatively uncontrolled ambient lighting reflected energy of reflected energy, so as to more effectively suppress or eliminate uncontrollable ring
The adverse effect of border illumination.
Image recognition and vision are very big by the direction of illumination and intensity effect of environment light, even if there is luminescent device to shine from front
Bright (such as light bulb or flash lamp), it is also possible that cannot keeping under control the influence of environment light (such as sunburst).Of the invention passes through handle
Front active illumination image subtracts ambient lighting images, eliminates environment shadow and rings, and difference image mainly only includes frontlighting
The reflected ingredient on object, be able in direction of illumination and intensity it is controlled, be conducive to from image detect and identify object.
Generally, in order to obtain preferable picture quality in the sun, intensity of illumination Ying Jinke that lighting unit 12 issues
Can be big, it is preferred with reaching the 2 times or more of sunlight intensity；It is exported using pulse as narrow as possible, may make light illuminating unit moment
Illumination as strong as possible is issued, and keeps its mean power as low as possible, from the mean power of limit luminescent device, reaches extension hair
The effect in 121 service life of optical device.
According to the analysis to sunlight light intensity experimental result, when noon in summer, sunlight was most strong, through 808nm it is
The heart, the solar energy of the bandpass filter of 10nm bandwidth are equivalent to the illumination of 15W or so through this filter disc.Therefore, compared with
Goodly, the illumination of light illuminating unit output 30W or the above power, so that sunlight power and the ratio of luminescent device power are more than or equal to
2, the effect of desirable preferable inhibition sunlight.
If image is it is possible that random speckle using commonly relevant light laser is as luminescent device 121.So
For the present invention when selection is using laser device, selection uses laser array, is in a disguised form modulated to spatial light, desirable comparison
The effect of hot spot is eliminated well.
Preferably, the luminescent device 121 of the lighting unit 12 of described image acquisition device 1, further includes the one of proper focal length
A plano-concave lens can adjust lighting angle range so that the luminescent device is placed in the focus of the plano-concave lens.
Preferably, the shooting direction of image collecting device 1 can be adjusted in a certain range, or by adjusting photograph therein
Bright unit 12 and sensing unit 11 adjust shooting direction.
Preferably, the type selecting of imaging sensor 111 is examined according to the luminescent device and filter unit used by wave band
Consider making by indexs such as the response sensitivities, signal-to-noise ratio, dynamic range of wave band to filter unit, it is raw that Sony company can be used
The ICX633BKA global shutter ccd sensor of production or the AR0237 Rolling shutter CMOS of On Semiconductor company production
Sensor.
Preferably, the time for exposure of imaging sensor 111, between 50 microsecond~10 millisecond, the time for exposure is shorter, then it is outer
Interference brought by the uncontrolled near-infrared light source in portion is smaller, is conducive to controlled light emitting devices imaging；Analog gain is 1~4.5
Times, most preferably 1 times；Digital gain is 0.25~2 times, most preferably 1 times；Dynamic range is the bigger the better more preferably greater than 80dB.
Preferably, control unit 13 use embedding assembly chip, may be selected sea think Hi3516, Hi3519, Hi3556,
The Hi3559 or micro- RK3228 of auspicious core, etc..Control unit 13 not only controls light source and Image Acquisition, may further be used to be subsequent 3D
Depth image calculates, image object detects and identification.
Preferably, in image collecting device of the invention, lighting unit 12 further includes for forming knot on acquisition target
The photo etched mask of structure light, sensing unit 11 obtain structure light image to the acquisition target shooting, collecting in the case where structure light is irradiated.?
I.e. when opening luminescent device irradiation, structure light is formed on acquisition target, sensing unit 11 forms structure light figure to shooting
Picture can acquire the NIR image of hot spot such as when opening infrared light supply in the sun.Also, due to being clapped using multispectral timesharing
The mode taken the photograph, the image that can also obtain visible light is used to show, namely applies respectively to the image of the different spectrum of shooting, real
The function of now two or more cameras being needed just to be able to achieve in the prior art.
It is highly preferred that image collecting device further includes the first three dimensional depth meter for calculating three dimensional depth from structure light image
Unit is calculated, to calculate three dimensional depth from the structure light image of acquisition, obtains three-dimensional depth information.
Three dimensional depth is calculated from structure light image, LightCoding laser speckle method in the prior art can be used:
Laser beam after projecting collimation by laser generator, passes through diffractive-optical element DOE
(DiffractiveOpticalElements) diffraction is carried out, and then obtains the required image (speckle pattern comprising speckle pattern
Picture).These speckles have height randomness, and can with the difference of distance changing patterns.Pass through infrared photography again
Machine records each speckle in down space on each plane of reference, forms benchmark calibration.After obtaining initial data, IR sensor is captured
Laser speckle pattern after testee distorts (modulation).Calculated by chip, available known pattern with
Offset of the pattern on space (x, y, z) is received, the depth information of testee is solved.
Since difference image can eliminate the influence of uncontrolled environment light source (such as sunlight, light) to a certain extent, from
And the image under uncontrolled light source environment can be improved and obtain quality, and then lift structure light method calculates the essence of depth image
Degree.
It is highly preferred that the image reconstruction unit structure light that image collecting device further includes speckle point in elimination speckle pattern swashs
After light beam passes through diffractive-optical element DOE, what optical sensor received is speckle image, and wherein speckle pattern is by largely dissipating
Spot is constituted, and it is dark by the part that DOE pattern is blocked that these speckle points, which can regard a kind of as and block to true picture,
Point, non-shield portions are bright spot, thus speckle image can regard sparse dot pattern picture as.Image reconstruction unit is from speckle image
Complete (eliminating speckle) image is rebuild, is sparse image reconstruction (Image Reconstruction) problem, can adopt
It is realized with sparse numerical value interpolation and the method for extrapolation.
In another embodiment of image collecting device of the invention, image collecting device further includes using flight time side
Second three dimensional depth computing unit of method calculating three dimensional depth.Such as can under the irradiation of NIR, to the image sequence of acquisition into
Row processing, calculate three-dimensional depth information using prior art flight time (TOF) method: near infrared laser emits light to object
Pulse, picture receiver then light pulse, by calculating time difference from transmitter to object, i.e., light pulse emission connects to returning to
The runing time of device is received, to determine the distance of measured object.Due to difference image can eliminate to a certain extent it is uncontrollable
Light source (influence of such as sunlight), so as to use TOF method to calculate depth image under sunlight.
As shown in figure 5, the present invention also provides a kind of object identification system, including image collecting device above-mentioned.Namely
The image of image acquisition device is analyzed, is found out and is applied interested object in image and classified, it is such as dynamic
Object, face, human body, vehicle, license plate, etc..Wherein, feature extraction unit 22 proposes 10 acquired image 21 of image collecting device
Feature is taken, characteristic pattern is obtained；Still further, object identification system further includes judging institute's detection object based on the characteristic pattern
The object classification unit 23 of classification and position is classified and is positioned to the object in input picture.The target of system output
Classification and corresponding subgraph are, it can be achieved that corresponding apply, such as during automatic Pilot, all vehicles detected are divided into one
Class, the people detected is divided into one kind etc., to handle respectively, makes different judgements.
One more specific object identification system embodiment is as shown in Figure 6.In Fig. 6, include with collected one
For the road photo of pedestrian, automobile and bicycle, object identification system analyzes image, and feature extraction unit uses base
In the deep neural network of ResNet.As shown in connection with fig. 7, the deep neural network of the ResNet is by a series of N number of ResNet moulds
Block cascade is constituted；Wherein, a ResNet module is made of one shortcut of level 2 volume lamination (as dotted line is included portion in Fig. 7
Point).Object classification unit in Fig. 6, using Softmax neural network, to the various rulers on each position point of characteristics of image
The subgraph of degree is classified, and classifier exports the classification number of the subgraph as classification results, and such as in Fig. 6, automobile is returned
For one kind, pedestrian is classified as one kind, and bicycle is classified as one kind etc..
Further, in the feature base of extraction, recognition of face can be carried out: to two face subgraph Ii(this hair
Bright image collecting device face subgraph collected) and Ij(a face subgraph to be compared in database) extracts phase
Answer feature xi,xjAfterwards, the COS distance of its feature is calculated
That is similarity.Threshold tau is set according to FAR (misclassification rate), according to similarity size, as cos (xi,xj) >=τ then determines
Two face pictures belong to the same person, are not otherwise the same persons.
As shown in Figure 8 and Figure 9, the present invention also provides a kind of panorama camera systems, including to different more towards what is shot
A image collecting device and the image mosaic that image mosaic is carried out to the image of the different perspectives of multiple images acquisition device shooting
System；At least one of multiple images acquisition device is image collecting device 10 above-mentioned.For example, using 8 phases in Fig. 8
Machine is respectively facing 8 different directions acquisitions, is then spliced by the image mosaic system of Fig. 9 to image.If therein
Camera uses image collecting device 10 above-mentioned, can generate at least three benefit: (1) depth that the 3D depth map of its output is included
Spend information, can provide more applications may, for example real enhance；(2) the 3D depth map of its output helps to promote image spelling
Precision is connect, to promote the quality of panorama camera；(3) the 3D depth map of its output, facilitates lifting object recognition performance, etc..
Therefore, panorama camera system of the invention can expand the application function of panorama camera, promote value of the product.
As shown in Figure 10, the present invention also provides a kind of automobile navigation obstacle avoidance systems, including image collecting device, environment to know
Other unit and display unit, image collector are set to including at least one image collecting device above-mentioned；Context awareness unit, it is right
Image collecting device acquired image is analyzed, to the ranging of environment object, object detection and/or object identification, from
And make or prompt corresponding driving strategy；Display unit is connect, for showing with image collecting device and Context awareness unit
The image or the analysis of Context awareness unit, detection, recognition result for showing image acquisition device can also be shown corresponding
Prompt information, such as barrier prompt, drive early warning.
The above description is merely a specific embodiment, under above-mentioned introduction of the invention, those skilled in the art
Other improvement or deformation can be carried out on the basis of the above embodiments.It will be understood by those skilled in the art that above-mentioned tool
Body description only preferably explains that the purpose of the present invention, protection scope of the present invention should be subject to the protection scope in claims.
Claims (13)
1. a kind of image collecting device characterized by comprising
Sensing unit, including imaging sensor and camera lens；
Lighting unit, the driving circuit of luminescent device and the driving luminescent device including at least one spectral band；
Control unit is connect with the sensing unit and the lighting unit, and for generating the figure for controlling the sensing unit
As the exposure time series and exposure gain of exposure sensor, and the illumination timing of the control lighting unit；Wherein, it is exposed at one
In the light sequential period, at least two exposed frame yield values are identical, and in the case where the luminescent device is closed, at least two exposed frames
Yield value is different.
2. image collecting device according to claim 1, which is characterized in that
Described image acquisition device further includes the photosensitive unit for detecting environment illumination intensity, and described control unit is according to
Exposure gain value of the intensity of illumination control of photosensitive unit detection in an exposure time series period inner part exposure frame.
3. image collecting device according to claim 1, which is characterized in that
The sensing unit further includes filter unit, and the filter unit is more bandpass filters or the glass that the camera lens is arranged in
More bandpass filter plated films of on piece；
The filter unit allows the multi light spectrum hands light of the luminescent device to pass through, and allows partial spectrum in ambient lighting
The light of wave band passes through.
4. image collecting device according to claim 1, which is characterized in that
Described image acquisition device further includes acquiring for calculating when the luminescent device that the sensing unit obtains is opened
The calculation process for the difference image between image that image and luminescent device acquire when closing.
5. image collecting device according to any one of claim 1 to 4, which is characterized in that
The lighting unit further includes photo etched mask, for penetrating the photo etched mask under luminescent device irradiation, is being adopted
Collect and form structure hot spot on object, so that the sensing unit obtains structure light image after carrying out shooting, collecting to acquisition target.
6. image collecting device according to claim 5, which is characterized in that
Described image acquisition device further include calculated from the structure light spot image acquisition target surface three dimension depth away from
From the first three dimensional depth computing unit.
7. image collecting device according to claim 5, which is characterized in that
Described image acquisition device further includes carrying out the elimination of structure hot spot to the structure light image, and rebuild the weight of complete image
Build unit.
8. image collecting device according to any one of claim 1 to 4, which is characterized in that
Described image acquisition device further includes the second three dimensional depth computing unit that three dimensional depth is calculated using pulsed-beam time-of-flight methods.
9. a kind of object identification system, which is characterized in that including image collecting device described in any item of the claim 1 to 8
And object detection unit, the object detection unit detection object and are obtained from described image acquisition device acquired image
The image of institute's detection object.
10. object identification system according to claim 9, which is characterized in that
The object identification system further includes the feature extraction unit that feature is extracted from the image of institute's detection object.
11. object identification system according to claim 10, which is characterized in that
The object identification system further includes the classification that institute's detection object is judged from the feature that the feature extraction unit is extracted
Object classification unit.
12. a kind of panorama camera system, including the multiple images acquisition device shot to different perspectives and to described multiple images
The image of the different perspectives of acquisition device shooting carries out the image mosaic system of image mosaic；It is characterized in that,
At least one of described multiple images acquisition device is image collecting device described in any item of the claim 1 to 8.
13. a kind of automobile navigation obstacle avoidance system, including image collecting device, Context awareness unit and display unit, feature exist
In,
Described image acquisition device includes at least one image collecting device described in any item of the claim 1 to 8；
The Context awareness unit analyzes described image acquisition device ambient image collected, measures and hinders in environment
Hinder the distance of object, the object in detection or environment-identification；
The display unit is connect, for showing that described image is adopted with described image acquisition device and the Context awareness unit
The image of acquisition means acquisition or the analysis result of the Context awareness unit.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810230565.7A CN110312079A (en) | 2018-03-20 | 2018-03-20 | Image collecting device and its application system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810230565.7A CN110312079A (en) | 2018-03-20 | 2018-03-20 | Image collecting device and its application system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110312079A true CN110312079A (en) | 2019-10-08 |
Family
ID=68074137
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810230565.7A Pending CN110312079A (en) | 2018-03-20 | 2018-03-20 | Image collecting device and its application system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110312079A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111016094A (en) * | 2019-12-28 | 2020-04-17 | 吉丽雄 | Modular demolding equipment based on image recognition and working method thereof |
| CN111811432A (en) * | 2020-06-16 | 2020-10-23 | 中国民用航空飞行学院 | Three-dimensional imaging system and method |
| WO2020238805A1 (en) * | 2019-05-31 | 2020-12-03 | 杭州海康威视数字技术股份有限公司 | Facial recognition apparatus and door access control device |
| CN112825491A (en) * | 2019-11-21 | 2021-05-21 | 北京外号信息技术有限公司 | Method and system for enabling detection of light emitting devices |
| CN114584712A (en) * | 2022-03-21 | 2022-06-03 | 优尼科(青岛)微电子有限公司 | Method and equipment for rapidly acquiring tunable spectral image and storage medium |
| CN115802170A (en) * | 2022-11-03 | 2023-03-14 | 奥比中光科技集团股份有限公司 | Image output method, image sensor and image output system |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1910431A (en) * | 2004-01-23 | 2007-02-07 | 奥林巴斯株式会社 | Image processing system and camera |
| CN102428356A (en) * | 2009-05-29 | 2012-04-25 | 丰田自动车株式会社 | Spectrum measuring apparatus for mover |
| CN103575917A (en) * | 2013-10-15 | 2014-02-12 | 中国农业大学 | Multispectral image collecting method and device |
| CN104323858A (en) * | 2014-10-22 | 2015-02-04 | 中国科学院自动化研究所 | Handheld molecular imaging navigation system |
| CN107044886A (en) * | 2017-01-20 | 2017-08-15 | 武汉呵尔医疗科技发展有限公司 | A kind of multiband spectrum imaging source, imaging method and microscope |
-
2018
- 2018-03-20 CN CN201810230565.7A patent/CN110312079A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1910431A (en) * | 2004-01-23 | 2007-02-07 | 奥林巴斯株式会社 | Image processing system and camera |
| CN102428356A (en) * | 2009-05-29 | 2012-04-25 | 丰田自动车株式会社 | Spectrum measuring apparatus for mover |
| CN103575917A (en) * | 2013-10-15 | 2014-02-12 | 中国农业大学 | Multispectral image collecting method and device |
| CN104323858A (en) * | 2014-10-22 | 2015-02-04 | 中国科学院自动化研究所 | Handheld molecular imaging navigation system |
| CN107044886A (en) * | 2017-01-20 | 2017-08-15 | 武汉呵尔医疗科技发展有限公司 | A kind of multiband spectrum imaging source, imaging method and microscope |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020238805A1 (en) * | 2019-05-31 | 2020-12-03 | 杭州海康威视数字技术股份有限公司 | Facial recognition apparatus and door access control device |
| CN112825491A (en) * | 2019-11-21 | 2021-05-21 | 北京外号信息技术有限公司 | Method and system for enabling detection of light emitting devices |
| CN112825491B (en) * | 2019-11-21 | 2022-05-24 | 北京外号信息技术有限公司 | Method and system for enabling detection of light emitting devices |
| CN111016094A (en) * | 2019-12-28 | 2020-04-17 | 吉丽雄 | Modular demolding equipment based on image recognition and working method thereof |
| CN111811432A (en) * | 2020-06-16 | 2020-10-23 | 中国民用航空飞行学院 | Three-dimensional imaging system and method |
| CN114584712A (en) * | 2022-03-21 | 2022-06-03 | 优尼科(青岛)微电子有限公司 | Method and equipment for rapidly acquiring tunable spectral image and storage medium |
| CN115802170A (en) * | 2022-11-03 | 2023-03-14 | 奥比中光科技集团股份有限公司 | Image output method, image sensor and image output system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110312079A (en) | Image collecting device and its application system | |
| WO2021244414A1 (en) | Facial recognition monitoring system based on spectrum and multi-band fusion, and recognition method | |
| US9964643B2 (en) | Vehicle occupancy detection using time-of-flight sensor | |
| CN102592335B (en) | Determine the total number of persons in the infrared image obtained by infrared imaging system | |
| EP3161521B1 (en) | Gated sensor based imaging system with minimized delay time between sensor exposures | |
| CN105705962B (en) | Sensing system with active illumination | |
| JP5647118B2 (en) | Imaging system | |
| US6420704B1 (en) | Method and system for improving camera infrared sensitivity using digital zoom | |
| US10482347B2 (en) | Inspection of the contoured surface of the undercarriage of a motor vehicle | |
| US9996748B2 (en) | Emitter tracking system | |
| JP2019508717A5 (en) | Method and apparatus for illumination invariant image sensor of automatic subject detection and vision system | |
| CN1332221C (en) | Method and device for suppressing electromagnetic background radiation in image | |
| CN106772431A (en) | A kind of Depth Information Acquistion devices and methods therefor of combination TOF technologies and binocular vision | |
| CN108431630A (en) | Vehicle image acquiring device, control device, include vehicle image acquiring device or control device vehicle and vehicle image acquiring method | |
| EP3572975B1 (en) | A multispectral image processing system for face detection | |
| IL266025A (en) | System for characterizing surroundings of a vehicle | |
| JP2003307407A (en) | Method and apparatus for scannerless color sequential depth mapping system | |
| GB2539495A (en) | Improvements relating to time-of-flight cameras | |
| CN113557172A (en) | Gated camera, automobile, vehicle lighting, object recognition system, arithmetic processing device, object recognition method, image display system, inspection method, camera device, image processing device | |
| US11936973B2 (en) | Pulse chain-driven infrared imaging assembly | |
| WO2021054198A1 (en) | Imaging device, imaging system, and imaging method | |
| JP6942637B2 (en) | Vehicles equipped with a vehicle image acquisition device and a vehicle image acquisition device | |
| US11653832B2 (en) | Image acquisition with reflex reduction | |
| CN118962828A (en) | Device and method for detecting hidden lens based on combined camera | |
| CN113109787A (en) | Non-vision field imaging device and method based on thermal imaging camera |
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 |
Application publication date: 20191008 |
|
| RJ01 | Rejection of invention patent application after publication |