CN207397296U - A kind of object identification device based on figure viewed from behind imaging - Google Patents
A kind of object identification device based on figure viewed from behind imaging Download PDFInfo
- Publication number
- CN207397296U CN207397296U CN201721514568.0U CN201721514568U CN207397296U CN 207397296 U CN207397296 U CN 207397296U CN 201721514568 U CN201721514568 U CN 201721514568U CN 207397296 U CN207397296 U CN 207397296U
- Authority
- CN
- China
- Prior art keywords
- lens
- under test
- object under
- camera
- image
- 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.)
- Active
Links
Landscapes
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The utility model discloses a kind of object identification device based on figure viewed from behind imaging, including planar light source, lens, camera and image processing apparatus；The planar light source for being arranged in lens one side generates the principal direction less parallel light vertical with lens focal plane, after being blocked by object under test, incides into lens；Lens screen incident ray, will one group of parallel light for inciding into lens be converged in the certain point on lens opposite side focal plane, the camera being disposed in herein captures to form image, which includes the shade of object under test；The image that image processing apparatus captures camera is handled, and obtains the contour curve of object under test, it is matched with profile sample curve, to identify object under test.The utility model identifies that transmission positioning has the characteristics of discrimination is high, simple in structure, cheap especially suitable for subject images such as Packaging Bottles.
Description
Technical field
The utility model belongs to subject image identification technology field and discarded object intelligence recovery technology field, more particularly to
A kind of object identification device based on figure viewed from behind imaging in the small-sized objects removal process such as Packaging Bottle.
Background technology
The intelligent retracting device for developing and promoting the rubbish such as waste plastic bottle is considered as the recovery system of specification Packaging Bottle, is subtracted
Few " white garbage " pollution it is most effective by way of one of.In the intelligent retracting device of Packaging Bottle, the identification technology of object is
Core component.Efficiently, the identification prepared effectively preventing error can recycle and prevent recycling fraud, so as to greatly carry
The reliability of high machine.
Chinese patent application CN 202762657U, 200620092237.8 recognition methods are based on the item to Packaging Bottle
Code scanning.It is also easily damaged since bar code is easily copied.Therefore, the identification method based on bar code hinders intelligence recycling side
The popularization of formula.It is the recoverable development trend of intelligence using integrated object identification means, the image recognition technology of wherein object is
One of technology of most critical.Chinese patent application CN104408820A is not included using camera capture and the recycling comprising object
Cabin image is simultaneously as base map and comparing figure, so as to identify whether object is recyclable plastics bottle, but the accuracy of identification of the technology according to
Rely the quality in image.Chinese patent application CN101727703A discloses a kind of tank body pattern recognition device, passes through CCD camera
Intake image simultaneously compares itself and image in database, but does not provide the technical detail of realization.Application No.
200620092237.8 Chinese patent application places light emitting source and light-sensitive element respectively in beverage bottle both sides, it is done so that can only
The diameter of object is judged, it is difficult to which the beverage bottle of input is accurately detected and judged.In addition, in the outline drawing of plastic bottle
As the research in terms of identification also has some other progress, such as machine vision method, RGB methods, image comparison method, above-mentioned skill
Art is complicated, of high cost, and is not applied to really in recycling machine product also.
As described above, to solve the above-mentioned problems, and accelerate the popularization of intelligent recycling machine, it is accordingly required in particular to utility model is a kind of
In high precision, it is inexpensive, suitable for the pattern recognition device and method of the small-sized objects such as Packaging Bottle.
Utility model content
The technical issues of the utility model solves is in view of the deficiencies of the prior art, to provide a kind of based on figure viewed from behind imaging
Object identification device, especially suitable for the subject image identification in Packaging Bottle the like waste intelligence removal process, it is accurate to have
Reliably, the characteristics of simple in structure, cheap, can have as the substitute products of domestic and international similar image recognition sensor, future
It is applied to the potentiality in more identification fields.
The technical solution that the utility model is solved is：
A kind of object identification device based on figure viewed from behind imaging, fills including planar light source, lens, camera and image procossing
It puts；
The planar light source is arranged in lens one side, and planar light source surface is parallel with lens focal plane；Camera is arranged in
On the focal plane of lens opposite side (near focal point)；
Between detection zone is located at planar light source and lens respectively, for placing object under test；
The planar light source for being arranged in lens one side generates principal direction (principal direction is exactly the highest direction of brightness, planar light source
The light of outgoing is substantially the light source of diffusion, it can be understood as has infinite multigroup directional light, wherein perpendicular to planar light source surface
Parallel brightness highest) the less parallel light vertical with lens focal plane (parallel to lens primary optical axis or slightly exists inclined
Difference), after being blocked by object under test, lens are incided into, lens screen incident ray, by one group of parallel lens that incide into
Light converges in the certain point (certain point of the first near focal point) on lens opposite side focal plane, is disposed in camera shooting herein
Head capture forms image, which includes the shade of object under test；
The image that image processing apparatus captures camera is handled, and obtains the contour curve of object under test, and by its
It is matched with profile sample curve, to identify examined object.Existing skill can be used in the contours extract and contour curve extraction
Art is realized.
Further, the planar light source is disposed with ultra tiny blinds fenestrated membrane (conditioning film) between object under test, is used for
The light non-perpendicular to lens focal plane direction (parallel to lens key light direction of principal axis) being emitted from planar light source is filtered, so as to subtract
The light of few oblique incidence avoids some reflection lights by chance with the direction vertical with lens focal plane in object under test edge reflections
(the parallel direction of lens primary optical axis) entrance lens, are then refracted to camera, and reflection bright spot, shadow are formed at image outline edge
Ring the situation of contours extract.
Further, the camera is arranged in the focal position on the focal plane of lens, and one group of capture is perpendicular to lens
The light of lens is incided into focal plane direction (parallel to lens key light direction of principal axis), to obtain optimal picture quality.
Further, the lens rear (image space) is disposed with speculum, for by converged light line reflection to side,
Camera also is disposed on the convergent point of the convergence light after reflection at this time.This design will reduce device lateral size.
Further, the object identification device based on figure viewed from behind imaging, further includes object under test transport mechanism, determinand
Body transport mechanism is controlled by image processing apparatus；The brightness of detection zone on the image that image processing apparatus is captured according to camera
Information, the operation of control object under test transport mechanism or stopping.
The lens select Fresnel lens.
The operation principle of above device is：
1) planar light source for being arranged in object under test one side generates the principal direction less parallel light vertical with lens focal plane；
2) less parallel light after being blocked by object under test, is formed containing object under test shadow information by object under test
Less parallel light, incides into lens；
3) lens screen incident ray, and one group of parallel light for inciding into lens is converged in lens opposite side coke
Certain point in plane；
4) the light line formed image of camera capture convergence is mounted on, which contains the shade of object under test；Figure
It is uniform and light ground outside object under test shadow region as in, there is very strong contrast with object under test shade；
5) image processing apparatus carries out gray processing, binaryzation and contour extraction processing to the image that camera captures, and obtains
The contour curve of object under test；
6) contour curve of the object under test with profile sample curve is matched, identifies object under test.
Further, in object under test transmit process, detected on the image that image processing apparatus is captured according to camera
The luminance information in region, so as to quickly judge whether object under test reaches detection zone or leave detection zone；It is to be measured when determining
When whether object reaches detection zone, control object under test transport mechanism is out of service, and object under test is made to rest on detection zone,
So as to capture the image for including object under test shade in a static condition, so as to carry out more accurate contours extract, matching
Or other detections；After the completion of detection, object under test is sent from detection zone by the transport mechanism operation of control object under test；At image
When reason device determines that object under test leaves detection zone according to the image that camera captures, execution subsequent operation (such as after starting
The operations such as the detection of one object under test or sorting, the compression of current object under test).
Further, the Boundary Extraction is according to the image data after binaryzation, using Image Edge-Detection operator, is carried
Take data boundary.
Further, the contours extract, using certain point corresponding with shadow of object outer boundary in data boundary as starting point,
Using Contour searching algorithm, the outer profile curve of object under test is obtained.
Further, the outline is that the outer profile curve of object is carried out the conversion such as displacement and mirror image, and is calculated
Curve and the Accumulated deviation value of the profile sample curve in profile sample storehouse after the conversion such as displacement or mirror image every time；When accumulative inclined
When difference is less than some threshold value, then it is assumed that the object under test coincide with profile sample；Wherein the computational methods of Accumulated deviation value are such as
Under：
1) using method at equal intervals, N and M are chosen respectively on the object external outline curve after profile sample curve and conversion
(number of N and M can adjust a point according to precision and calculating speed, and the N and M of bigger correspond to higher precision, but calculating speed
It will reduce)；
2) to the N number of point chosen in profile sample curve, itself and M in the object external outline curve after conversion are calculated successively
The minimum value of the distance of point；
3) minimum value of N number of respective distances in profile sample curve is summed, obtains Accumulated deviation value.
In above-mentioned computational methods, profile sample curve and the object external outline curve after conversion can be exchanged.I.e.：
1) using method at equal intervals, distinguish and choose N and M in object external outline curve and profile sample curve after the conversion
A point；
2) to the N number of point chosen on the object external outline curve after conversion, itself and M in profile sample curve are calculated successively
The minimum value of the distance of point；
3) minimum value of N number of respective distances of object external outline curve after conversion is summed, obtains Accumulated deviation value.
Advantageous effect：
(1) solve currently marketed Packaging Bottle intelligence retracting device means of identification and depend on lacking for bar code scan
Point；By the contours extract to object and matching, accurately and rapidly Packaging Bottle can be identified.
(2) method based on figure viewed from behind imaging is innovatively utilized to solve the problems, such as contours extract and the matching of object.Identification dress
Put image recognition sensor simple in structure, cheap, that domestic and international similar functions can be substituted completely.
(3) common reflector is substituted using planar light source so that background is more uniformly distributed and becomes clear, image acquisition and wheel
Wide extraction rate and accuracy rate greatly improve.
(4) directly it is detected according to the position for getting image data measuring targets, so as to eliminate object under test
The needs of limit sensors are arranged in transmitting path.Once object under test reaches detection zone, image processing apparatus, which can control, to be treated
It is out of service to survey object transport mechanism, object under test is made to rest on detection zone, so as to carry out more accurate contours extract, matching
Leaving with other coherent detections or in object under test drives associated mechanisms to carry out subsequent sorting, compression etc. when detection zone is put
Operation.
(5) the utility model is also applied to the other field outside Packaging Bottle intelligently recycles, such as product sorting, rubbish
The fields such as classification, have a wide range of application.
Description of the drawings
Fig. 1 is the utility model device basic principle schematic；
Fig. 2 is 1 schematic diagram of the utility model embodiment；
Fig. 3 is 2 schematic diagram of the utility model embodiment；
Fig. 4 is the design sketch of the utility model embodiment 1；
Fig. 5 is the design sketch of the utility model embodiment 2；
Reference sign：In figure, 1 it is image processing apparatus, 2 is camera, 3 is focus, 4 is lens, 5 is determinand
Body, 6 be planar light source, 7 be speculum, 8 be ultra tiny blinds fenestrated membrane.
Specific embodiment
Fig. 1 is the utility model device basic principle schematic.It to be based on as shown in Figure 1, the utility model discloses one kind
The object identification device of figure viewed from behind imaging, including image processing apparatus 1, camera 2, lens 4 and planar light source 6；
The planar light source 6 for figure viewed from behind imaging for the lens 4 of imaging with being arranged in the two of detection zone
Side；Detection zone is used to place object under test 5；
Fresnel lens can be selected in the lens 4, for being screened to incident ray, wherein flat perpendicular to lens 4
The directional light of face incidence will be aggregated in the focus 3 of lens opposite side (these light include the shadow information of object)；
The camera 2 is arranged near the focus 3 on the focal plane of lens 4 (better closer to 3 effect of focus), uses
It is emitted in capture through planar light source 6, and after object under test 5 blocks, the parallel rays of lens 4 is incided into special angle, from
And obtain the image for including 5 shade of object under test.
The image processing apparatus 1 is used for the image gray processing for exporting camera 2, binaryzation, and row bound of going forward side by side carries
It takes, contours extract, outline etc..Described image gray processing, binaryzation, Boundary Extraction, contours extract, outline can adopt
It is realized with the prior art.
The Boundary Extraction is according to the image data after binaryzation, using Image Edge-Detection operator, extracts border
Data.
The contours extract, using starting point is in certain point corresponding with shadow of object outer boundary in data boundary, using profile
Searching algorithm obtains the outer profile curve of object under test.
The outline is to be shifted the outer profile curve of object and mirror transformation, and calculates displacement or mirror every time
As the curve after conversion and the Accumulated deviation value of the profile sample curve in profile sample storehouse；When Accumulated deviation value is less than some threshold
During value, then it is assumed that the object under test coincide with profile sample；The computational methods of wherein Accumulated deviation value are as follows：
1) using method at equal intervals, N and M are chosen respectively on the object external outline curve after profile sample curve and conversion
(number of N and M can adjust a point according to precision and calculating speed, and the N and M of bigger correspond to higher precision, but calculating speed
It will reduce)；
2) to the N number of point chosen in profile sample curve, itself and M in the object external outline curve after conversion are calculated successively
The minimum value of the distance of point；
3) minimum value of N number of respective distances in profile sample curve is summed, obtains Accumulated deviation value.
In above-mentioned computational methods, profile sample curve and the object external outline curve after conversion can be exchanged.
With reference to embodiment, the utility model is described in further detail.
Embodiment 1：
Using the contour of object being imaged based on the figure viewed from behind extraction, matching and localization method of the utility model, in an implementation
In example, speculum 7 is selected to be arranged in 4 rear end of lens, using beverage bottle as identification object.Specifically carry out by the following method：
1) light that the planar light source 6 of 5 one side of object under test projects is arranged in, forms less parallel light, less parallel light
Principal direction it is vertical with 4 plane of lens；
2) less parallel light forms the less parallel light containing 5 shade of object under test, entrance lens by object under test 5
4；
3) lens 4 screen incident ray, wherein the directional light perpendicular to lens 4 plane incidence will be aggregated saturating
Certain point on mirror opposite side focal plane；
4) speculum 7 changes the position of light convergent point, so as to reduce device lateral size；
5) extraction of camera 2 that the light of convergence is installed at this forms image, and the image of extraction contains object under test 5
Cross-sectional profiles shade.With being uniform and light ground outside 5 corresponding shadow region of object under test in image, with determinand
Body shade has very strong contrast；
6) by carrying out gray processing, binaryzation, Boundary Extraction, wheel to the image that camera captures to image processing apparatus 1
Wide extraction process obtains the contour curve of object under test；Fig. 4 is the recognition effect in the present embodiment, is ash successively from top to bottom
Degreeization, binaryzation, Boundary Extraction, contour extraction processing result；
7) contour curve and profile sample curve are subjected to outline, identify object under test.
In the present embodiment, select speculum be arranged in lens rear, for by converged light line reflection to side, meanwhile,
Camera need to also be disposed on the convergent point of the convergence light after reflecting according to mirror angle, so as to reduce device lateral ruler
It is very little.Preceding metallic-membrane plating reflector can be selected in the speculum, to reduce the ghost image of glass surface reflection.Fig. 4 shows beverage in the present embodiment
The Outside contour extraction effect of bottle fully meets identification requirement.
Embodiment 2：
The contour of object extraction being imaged using the utility model based on the figure viewed from behind and matching process, in one embodiment,
Select speculum 7 be arranged in 4 rear end of lens, while select ultra tiny blinds fenestrated membrane 8 be arranged in planar light source in object under test it
Between, using beverage bottle as identification object.Specifically carry out by the following method：
1) light that the planar light source 6 of 5 one side of object under test projects is arranged in, forms less parallel light, less parallel light
Principal direction it is vertical with 4 plane of lens；
2) ultra tiny blinds fenestrated membrane 8 is arranged in planar light source 6 between object under test 5, for filtering from planar light source 6
The light non-perpendicular to lens focal plane direction of outgoing so as to reduce the light of oblique incidence in 5 edge reflections of object under test, is kept away
Exempt from some reflection lights by chance with the direction entrance lens 4 vertical with 4 focal plane of lens, camera 2 is then refracted to, in image
Contour edge forms reflection bright spot, influences the situation of contours extract.
3) less parallel light forms the less parallel light containing 5 shade of object under test, entrance lens by object under test 5
4；
4) lens 4 screen incident ray, wherein the directional light perpendicular to the incidence of 4 focal plane of lens will be aggregated
In the focus of lens opposite side；
5) speculum 7 changes the position of light convergent point, so as to reduce device lateral size；
6) extraction of camera 2 that the light of convergence is installed at this forms image, and the image of extraction contains object under test 5
Cross-sectional profiles shade.With being uniform and light ground outside 5 corresponding shadow region of object under test in image, with determinand
Body shade has very strong contrast；
7) by carrying out gray processing, binaryzation, Boundary Extraction, wheel to the image that camera captures to image processing apparatus 1
Wide extraction process obtains the contour curve of object under test；Fig. 5 is the recognition effect in the present embodiment, is ash successively from top to bottom
Degreeization, binaryzation, Boundary Extraction, contour extraction processing result；
8) contour curve and profile sample curve are subjected to outline, identify object under test.
In the present embodiment, speculum is selected to be arranged in lens rear, for by converged light line reflection to side, reduction to fill
Put lateral dimension.Ultra tiny blinds fenestrated membrane is arranged in planar light source between object under test simultaneously, is avoided on image outline side
Edge forms reflection bright spot, so as to influence the situation of contours extract.Fig. 5 shows the Outside contour extraction effect of beverage bottle in the present embodiment
Fruit fully meets identification requirement.
Above example is merely to the thought and operation principle of the utility model described in more detail, it should be understood that
Embodiment mentioned above should not limit the utility model.Every all changes within the spirit and scope of the utility model and
It replaces, it should all be in the row of the protection of the utility model.
Claims (6)
1. a kind of object identification device based on figure viewed from behind imaging, it is characterised in that：Including planar light source, lens, camera and figure
As processing unit；
The planar light source is arranged in lens one side, and planar light source surface is parallel with lens focal plane；Camera is arranged in lens
On the focal plane of opposite side, for capturing the light line formed image of lens convergence；
Between detection zone is located at planar light source and lens respectively, for placing object under test.
2. the object identification device according to claim 1 based on figure viewed from behind imaging, it is characterised in that：The planar light source in
Ultra tiny blinds fenestrated membrane is disposed between object under test.
3. the object identification device according to claim 2 based on figure viewed from behind imaging, it is characterised in that：The camera arrangement
Focal position on the focal plane of lens.
4. the object identification device according to claim 1 based on figure viewed from behind imaging, it is characterised in that：The lens rear cloth
Speculum is equipped with, camera is arranged on the convergent point of convergence light after reflection.
5. the object identification device according to claim 1 based on figure viewed from behind imaging, it is characterised in that：Further include object under test
Transport mechanism, object under test transport mechanism are controlled by image processing apparatus.
6. according to the object identification device according to any one of claims 1 to 5 being imaged based on the figure viewed from behind, it is characterised in that：Institute
Lens are stated as Fresnel lens.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201721514568.0U CN207397296U (en) | 2017-11-14 | 2017-11-14 | A kind of object identification device based on figure viewed from behind imaging |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201721514568.0U CN207397296U (en) | 2017-11-14 | 2017-11-14 | A kind of object identification device based on figure viewed from behind imaging |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN207397296U true CN207397296U (en) | 2018-05-22 |
Family
ID=62326346
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201721514568.0U Active CN207397296U (en) | 2017-11-14 | 2017-11-14 | A kind of object identification device based on figure viewed from behind imaging |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN207397296U (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109460766A (en) * | 2018-09-29 | 2019-03-12 | 北京单多啦科技有限公司 | A kind of device and method for extracting bill images from bill paper |
| CN111524151A (en) * | 2020-04-29 | 2020-08-11 | 深圳市铭特科技有限公司 | Object detection method and system based on background recognition |
-
2017
- 2017-11-14 CN CN201721514568.0U patent/CN207397296U/en active Active
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109460766A (en) * | 2018-09-29 | 2019-03-12 | 北京单多啦科技有限公司 | A kind of device and method for extracting bill images from bill paper |
| CN109460766B (en) * | 2018-09-29 | 2022-03-15 | 济南企财通软件有限公司 | Device and method for extracting ticket image from ticket paper |
| CN111524151A (en) * | 2020-04-29 | 2020-08-11 | 深圳市铭特科技有限公司 | Object detection method and system based on background recognition |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110596129B (en) | Sheet glass edge flaw detection system based on image acquisition | |
| CN108896574B (en) | Bottled liquor impurity detection method and system based on machine vision | |
| CN110246122A (en) | Small size bearing quality determining method, apparatus and system based on machine vision | |
| CN109454006A (en) | Chemical fiber wire ingot is stumbled the device and its detection stage division of a defect on-line checking and classification | |
| CN107742119A (en) | A kind of contour of object extraction and coalignment and method based on figure viewed from behind imaging | |
| CN110108711A (en) | The vision detection system of annulus side wall defect | |
| JPH06100555B2 (en) | Defect inspection method and device for transparent object | |
| CN105865570A (en) | Machine vision-based glass bottle liquid level detection method | |
| CN113418933B (en) | Flying shooting visual imaging detection system and method for detecting large-size object | |
| CN114636706A (en) | Comprehensive method and device for image detection of solar cell after film coating | |
| CN106546600A (en) | A kind of PET bottle multi-orientation detection method and apparatus based on machine vision | |
| CN102374997A (en) | High-precision detection device of coin surface quality based on vision system | |
| CN207397296U (en) | A kind of object identification device based on figure viewed from behind imaging | |
| CN103712568B (en) | The reflective feature detection system of a kind of plastic pallet based on machine vision | |
| Liu et al. | Automatic detection technology of surface defects on plastic products based on machine vision | |
| CN204287068U (en) | For detecting the substrate imperfection pick-up unit of light-permeable base material | |
| CN105510339A (en) | Beverage bottle fullness detection method based on machine vision | |
| CN105973903B (en) | A kind of Oral liquid bottle lid detection method | |
| CN110044919A (en) | A kind of detection device and its detection method for mirror surfaces scratch | |
| CN207690117U (en) | A kind of device that article profile is obtained and identified | |
| CN204789392U (en) | Optic system of visual detection chamfer face | |
| CN107246841A (en) | A kind of multiple views grouping system device of auto parts and components | |
| CN115909359A (en) | Capacitance character detection method based on intelligent vision | |
| CN212180650U (en) | Paste surface air hole automatic detection equipment | |
| JP6937647B2 (en) | Optical display panel damage inspection method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| GR01 | Patent grant | ||
| GR01 | Patent grant |