CN107067431A - A kind of object volume computational methods based on Kinect - Google Patents
A kind of object volume computational methods based on Kinect Download PDFInfo
- Publication number
- CN107067431A CN107067431A CN201710027717.9A CN201710027717A CN107067431A CN 107067431 A CN107067431 A CN 107067431A CN 201710027717 A CN201710027717 A CN 201710027717A CN 107067431 A CN107067431 A CN 107067431A
- Authority
- CN
- China
- Prior art keywords
- image
- testee
- matrix
- kinect
- mtd
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20036—Morphological image processing
Landscapes
- Length Measuring Devices By Optical Means (AREA)
Abstract
The invention discloses a kind of object volume computational methods based on Kinect, including：（1）Utilize Kinect sampling depths image and coloured image；（2）Kinect colour imagery shot is demarcated；（3）The ROI region of depth image is set, image segmentation is carried out using the prospect coloured image comprising measured object of collection and the background color image of the test desk not comprising testee, obtains the bianry image of testee；（4）Background depth image ROI region is converted into background distance matrix, and it is pre-processed, foreground depth image ROI region is converted to prospect distance matrix by the element for being zero in filling background distance matrix；（5）Difference is subtracted according to preceding, background distance matrix and obtains height matrix；（6）Calculate object length, width and height size and object volume.The present invention efficiently solves the problem of Traditional Man measurement labor intensity is big, time of measuring is long, is a kind of non-contacting measurement means, measurement target is not injured, can meet the requirement of automation while improving measurement accuracy.
Description
Technical field
The present invention relates to a kind of object volume computational methods based on Kinect, belong to computer vision field.
Background technology
With the development of Digital Signal Processing and computer technology, video camera is obtained into external environment image and number is converted to
Word signal, is realized with computer and is referred to as computer vision to new technology formed by the processing overall process of visual information.It will calculate
Machine vision is applied to the accurate measurement and positioning of space geometry size, so as to produce vision measurement technology.Vision measurement as work as
One of modern new and high technology, obtains fast development, in production on the basis of image procossing and computer technology are constantly ripe and perfect
The fields such as product examine survey, reverse-engineering, robot navigation are obtained for extensive use.Computer vision measurement technology is with image sensing
Device detects the 3 d space coordinate of space object, and then size, shape and motion state of detection object etc. for means.
The Kinect somatosensory interactive device that Microsoft releases possesses a RGB camera, an infrared transmitter and one
Individual infrared camera, the coloured image and depth image of scene can be caught simultaneously, 3D rebuild, action capture, virtual reality,
Good application is obtained in terms of augmented reality, but is also seldom applied to technical field of visual measurement.Kinect phases
It is cheap for other depth extraction equipment, it is easy to operate.
At present in logistic industry, the express delivery point for being distributed in various regions is numerous.The Main Basiss of charging are the weight and body of goods
Product.In actual applications, mostly or by traditional based on the measurement of manually contact.External Cubiscan measurement series are set
Standby price is high, is unfavorable for being generalized to each scattered logistics post.Domestic box sizes Measuring light screen can only measurement rules object, nothing
Method meets quick, diversified require.
The content of the invention
Purpose：In view of the shortcomings of the prior art, the present invention provides a kind of object volume computational methods based on Kinect, leads to
Cross a kind of untouchable measurement means, measurement target do not injured, measured object nature is not disturbed, can automation condition
Lower application and convenient disassembly, it is convenient to install, solve traditional measurement method time of measuring length, the problem of cost is high.
Technical scheme is as follows：
A kind of object volume computational methods based on Kinect, specific steps include：
(1) foreground depth image and prospect coloured image comprising testee are gathered and not comprising quilt using Kinect
Survey the background depth image and background color image of the test desk of object；
(2) chessboard calibration is carried out to Kinect colour imagery shot, passes through the pass between image coordinate system and world coordinate system
System, calculates the inner parameter and external parameter of Kinect colour imagery shots,
(3) background color image that the prospect coloured image of collection in step (1) is subtracted into collection obtains subtracting the area after difference
Domain, carries out image segmentation to subtracting the region after difference, obtains the bianry image of testee, set according to the actual area of test desk
The ROI region of prospect ROI region and background depth image；
(4) image preprocessing is carried out to the background ROI region obtained in step (3), fills the cavity of ROI region, it is described
Image preprocessing includes mean filter, is filled, is iterated until all cavities with the average value of empty pixel neighbours thresholding
Pixel is filled and finished；
(5) foreground depth image and background depth image subtract obtaining testee height square after difference takes absolute value
Battle array, and be filtered pretreatment to the testee height matrix, the filter preprocessing include setting the height value upper limit with
Numerical value outside height value lower limit, bound is filtered into 0；
(6) the testee height matrix computations testee height value according to step (5) after processing, according to inverse projection
The principle of conversion, using the inner parameter and external parameter of step (2) the Kinect colour imagery shots, calculates testee pair
The world coordinates matrix answered, using the minimum enclosed rectangle for finding world coordinates matrix method calculate testee length and
It is wide；
(7) world coordinates matrix and testee height matrix according to step (6) is calculated using the method for integration
Testee volume.
Preferably, object under test is placed on measuring table in the step (1), by above measuring table
Kinect carries out IMAQ.
Preferably, step (2) is described carries out chessboard calibration to Kinect video camera, determines image coordinate system and world coordinates
The relation of system, calculates the inner parameter and external parameter of Kinect colour imagery shots, shown in all-purpose camera model such as formula (I)：
In formula (I), K is relevant with video camera internal structure, is intrinsic parameters of the camera；R, t are with video camera relative to the world
The orientation of coordinate system is relevant, is the external parameter of video camera, and R is spin matrix, and t is translation matrix, and [x y] is scene image
The coordinate of pixel, [X Y Z] is the coordinate of [x y] in world coordinate system.
Preferably, step (6) the object length, width and height size computing method is as follows:
The computational methods of (4a) world coordinates：According to the chessboard calibration data of step (2), using known parameters Z by general
Camera model formula (I) reverse goes out the corresponding world coordinate system coordinate (X, Y) of each point (x, y) on whole image coordinate system, and will
Result of calculation is saved in tables of data Xcoord_in_World and Ycoord_in_World, so that the quilt in step (3)
The object height matrix surveyed in the bianry image and step (5) of object obtains corresponding testee by way of tabling look-up
World coordinates；It is known that parameter Z is the element value height in the testee height matrix of acquisition in step (5), it is counted
Formula is calculated such as shown in (II)：
Height=Dground-Dobject (II)
Wherein, DgroundFor the element value in background depth image, the distance of Kinect each points to measuring table is represented,
DobjectFor the element value in foreground depth image, Kinect is represented to the distance of each point comprising testee；
(4b) object length and wide calculating：Specify world coordinates matrix maximum magnitude, newly-built world coordinates matrix
(XcoordRange, YcoordRange) and the element value for initializing world coordinates matrix is zero；Travel through the two-value of testee
Image, if the pixel value of bianry image is 1, tables look-up Xcoord_in_World and Ycoord_in_World according to step (4a)
Obtain correspondence world coordinates X and Y；And world coordinates matrix corresponding element value is set to 1；The world coordinates matrix obtained to calculating
Optimization is filtered, the world coordinates matrix minimum enclosed rectangle after filtering optimization is sought, if four summits of minimum enclosed rectangle
Coordinate be (X1, Y1), (X2, Y2), (X3, Y3), (X4, Y4), then the length of testee beThe width of testee is
Preferably, step (7) described object volume is as follows:
(7a) initialization volume volume is zero, travels through testee bianry image, if testee bianry image
Pixel value is 1, and world coordinates matrix TheWorld is mapped to using four adjacent pixels of the pixel as unit；
(7b) calculate this four pixels be mapped to world coordinates matrix long L and the width neighbor pixel of W, i.e., two away from
From corresponding realistic space actual distance, cellar area L × W is sought；
Cellar area is multiplied by the height value in the corresponding testee height matrix of pixel by (7c), and add up obtains one by one
Measured object volume.Calculation formula is such as shown in (III)：
Volume=volume+L × W × height (III)
Preferably, the filter preprocessing described in step (5) refer to set testee height matrix the height value upper limit and
Numerical value outside height value lower limit, bound is filtered into 0.
Preferably, the method for filtering optimization described in step (4b) utilizes for world coordinates matrix is narrowed down into 1/4 size
Morphology opening operation is removed amplifies 4 times, the world coordinates matrix after being optimized by image again after scattered noise.
Beneficial effects of the present invention：The present invention provides a kind of object volume computational methods based on Kinect, effectively solves
The problem of Traditional Man measurement labor intensity is big, time of measuring is long.With good stability and reliability, being that one kind is non-connects
Tactile measurement means, measurement target is not injured, can meet the requirement of automation while improving measurement accuracy.
Brief description of the drawings
Fig. 1 is the schematic diagram of measurement apparatus of the present invention.
Fig. 2 is the schematic diagram that the present invention is mapped to world coordinates by image coordinate.
Fig. 3 is the algorithm flow chart that the present invention calculates object length and width.
Fig. 4 is that the present invention calculates object volume algorithm flow chart.
Embodiment
In order that those skilled in the art more fully understand the technical scheme in the application, it is real below in conjunction with the application
The accompanying drawing in example is applied, the technical scheme in the embodiment of the present application is clearly and completely described, it is clear that described implementation
Example only some embodiments of the present application, rather than whole embodiments.Based on the embodiment in the application, this area is common
The every other embodiment that technical staff is obtained under the premise of creative work is not made, should all belong to the application protection
Scope.
As Figure 1-4, a kind of object volume computational methods based on Kinect, specific steps include：
(1) foreground depth image and prospect coloured image comprising testee are gathered and not comprising quilt using Kinect
Survey the background depth image and background color image of the test desk of object；
(2) chessboard calibration is carried out to Kinect colour imagery shot, passes through the pass between image coordinate system and world coordinate system
System, calculates the inner parameter and external parameter of Kinect colour imagery shots,；
(3) background color image that the prospect coloured image of collection in step (1) is subtracted into collection obtains subtracting the area after difference
Domain, carries out image segmentation to subtracting the region after difference, obtains the bianry image of testee, set according to the actual area of test desk
Prospect ROI region and background ROI region；
(4) image preprocessing is carried out to the background ROI region obtained in step (3), fills the cavity of ROI region, it is described
Image preprocessing includes mean filter, is filled, is iterated until all cavities with the average value of empty pixel neighbours thresholding
Pixel is filled and finished；
(5) foreground depth image and background depth image subtract obtaining testee height square after difference takes absolute value
Battle array, and pretreatment is filtered to the testee height matrix；
(6) the testee height matrix computations testee height value according to step (5) after processing, according to inverse projection
The principle of conversion, using the inner parameter and external parameter of step (2) the Kinect colour imagery shots, calculates testee pair
The world coordinates matrix answered, using the minimum enclosed rectangle for finding world coordinates matrix method calculate testee length and
It is wide；
(7) the testee height matrix described in world coordinates matrix according to step (6) and step (5) utilizes product
The method divided calculates testee volume.
Preferably, object under test is placed on measuring table in the step (1), by above measuring table
Kinect carries out IMAQ.
Preferably, step (2) is described carries out chessboard calibration to Kinect video camera, determines image coordinate system and world coordinates
The relation of system, calculates the inner parameter and external parameter of video camera, shown in all-purpose camera model such as formula (I)：
In formula (I), K is relevant with video camera internal structure, referred to as intrinsic parameters of the camera；R, t are with video camera relative to generation
The orientation of boundary's coordinate system is relevant, referred to as the external parameter of video camera, and R is spin matrix, and t is translation matrix, and [x y] is scene
The coordinate of image slices vegetarian refreshments, [X Y Z] is the coordinate of [x y] in world coordinate system.
Preferably, step (6) the object length Size calculation is as follows:
The computational methods of (4a) world coordinates：According to the chessboard calibration data of step (2), using known parameters Z by general
Camera model formula (I) reverse goes out the corresponding world coordinate system coordinate (X, Y) of each point (x, y) on whole image coordinate system, and will
Result of calculation is saved in tables of data Xcoord_in_World and Ycoord_in_World, so that the quilt in step (3)
The object height matrix surveyed in the bianry image and step (5) of object obtains corresponding testee by way of tabling look-up
World coordinates；It is known that parameter Z is the element value height in the testee height matrix of acquisition in step (5), it is counted
It is as shown in (II) to calculate formula：
Height=Dground-Dobject (II)
Wherein, DgroundFor the element value in background depth image, the distance of Kinect each points to measuring table is represented,
DobjectFor the element value in foreground depth image, Kinect is represented to the distance of each point comprising testee；
(4b) object length and wide calculating：Specify world coordinates matrix maximum magnitude, newly-built world coordinates matrix
(XcoordRange, YcoordRange) and the element value for initializing world coordinates matrix is zero；Travel through the two-value of testee
Image, if the pixel value of bianry image is 1, tables look-up Xcoord_in_World and Ycoord_in_World according to step (4a)
Obtain correspondence world coordinates X and Y；And world coordinates matrix corresponding element value is set to 1；The world coordinates matrix obtained to calculating
Optimization is filtered, the world coordinates matrix minimum enclosed rectangle after filtering optimization is sought, if four summits of minimum enclosed rectangle
Coordinate be (X1, Y1), (X2, Y2), (X3, Y3), (X4, Y4), then the length of testee beThe width of testee is
Preferably, step (7) described object volume is calculated as follows：
(5a) initialization volume volume is zero, travels through testee bianry image, if bianry image pixel value is 1,
World coordinates matrix is mapped to using four adjacent pixels of the pixel as unit；
(5b) calculate this four pixels be mapped to world coordinates matrix long L and the width neighbor pixel of W, i.e., two away from
From corresponding realistic space actual distance, cellar area L × W is sought；
Cellar area is multiplied by the height value in the corresponding testee height matrix of pixel by (5c), and add up obtains one by one
Measured object volume, its calculation formula is such as shown in (III)：
Volume=volume+L × W × height (III)
Preferably, the filter preprocessing described in step (5) refer to set testee height matrix the height value upper limit and
Numerical value outside height value lower limit, bound is filtered into 0.
Preferably, the method for filtering optimization described in step (4b) utilizes for world coordinates matrix is narrowed down into 1/4 size
Morphology opening operation is removed amplifies 4 times, the world coordinates matrix after being optimized by image again after scattered noise.
Embodiment 1：
(1a) gathers foreground depth image and prospect coloured image comprising testee using Kinect and not included
The background depth image and background color image of the test desk of testee；
(1b) carries out chessboard calibration to Kinect colour imagery shot, passes through the pass between image coordinate system and world coordinate system
System, calculates the inner parameter and external parameter of video camera, and all-purpose camera model is represented by as shown in formula (I)：
In formula (I), K is only relevant with video camera internal structure, referred to as intrinsic parameters of the camera；R, t are only relative with video camera
Relevant in the orientation of world coordinate system, the referred to as external parameter of video camera, R is spin matrix, and t is translation matrix, and [x y] is figure
As the coordinate of pixel, [X Y Z] is the coordinate of [x y] in world coordinate system；The inside ginseng of colour imagery shot in the present invention
Number and external parameter are calculated according to chessboard calibration, and the method for demarcation uses Zhang Zhengyou scaling methods, and R, t is referred to as shooting
The external parameter of machine, calculates according to the calibration tool case in MATLAB, belongs to conventional technical means.
The background color image that the prospect coloured image of collection in step (1a) subtracts collection is obtained subtracting after difference by (1c)
Region, carries out image segmentation to subtracting the region after difference, obtains the bianry image of testee, set according to the actual area of test desk
Put prospect ROI region and background ROI region；
(1d) carries out image preprocessing to the background ROI region obtained in step (1c), fills the sky of background ROI region
Hole, described image pretreatment includes mean filter, is filled, is iterated until all with the average value of empty pixel neighbours thresholding
Empty pixel be filled and finish；
Foreground depth image and background depth image subtract obtaining testee height square after difference takes absolute value by (1e)
Battle array, and be filtered pretreatment to the testee height matrix, the filter preprocessing include setting the height value upper limit with
Numerical value outside height value lower limit, bound is filtered into 0；
The testee height matrix computations testee height value of (1f) according to step (1e) after processing, according to inverse throwing
The principle of shadow conversion, testee is calculated using the inner parameter and external parameter of step (1b) the Kinect colour imagery shots
Corresponding world coordinates matrix, testee length and width are calculated using the method for the minimum enclosed rectangle for finding world coordinates matrix,
Calculate obtain after object height matrix just can using height matrix element value as world coordinate system Z parameter.
As shown in Fig. 2 a pixel can be mapped back in turn by reality three-dimensional by Inverse projection, but not
The coordinate of corresponding reality three-dimensional point can be obtained by the coordinate of image pixel, its essential reason is due to three-dimensional objective scene
It is conversion of the multiple spot to any to be mapped to image plane.When a world of the three dimensions point of known generation image slices vegetarian refreshments is sat
Mark, i.e., Z coordinate by camera model formula (I), it is known that just can obtain two other world coordinates X, Y.Because R is 3 × 3 squares
Battle array, t is 1 × 3 matrix, and K is 3 × 3 matrixes, soFor 4 × 3 matrixes.
IfFor 4 × 3 matrixes,
It can be obtained after then calculatingWherein, matrix T and Z is, it is known that then may be used
In the hope of going out world coordinates X, Y.
As shown in figure 3, calculating object length and width flow chart, in the following way：World coordinates matrix maximum magnitude is specified, i.e.,
The actually active region (XcoordRange, YcoordRange) of measuring table, newly-built world coordinates matrix TheWorld
(XcoordRange, YcoordRange) and the element value for initializing world coordinates matrix is zero；Travel through the two-value of testee
Image, if pixel value is 1, obtains correspondence world coordinates X and Y；And world coordinates matrix corresponding element value is put 1 i.e.
TheWorld (X, Y)=1；Calculate obtained world coordinates matrix TheWorld and there is scattered noise, the calculating on length and width influences
It is larger.The world coordinates matrix TheWorld obtained to remove noise to mapping is filtered optimization, by world coordinates matrix
1/4 size is narrowed down to, is removed using morphology opening operation and image is amplified 4 times again after scattered noise, seek the generation after filtering optimization
Boundary's coordinates matrix minimum enclosed rectangle, if four apex coordinates of minimum enclosed rectangle be (X1, Y1), (X2, Y2), (X3, Y3),
(X4, Y4), then length isAs the length of testee, width is
It is used as the width of testee.
As shown in figure 4, calculating volume flow chart, by the way of cumulative, step is as follows：Initializing volume volume is
Zero, testee bianry image is traveled through, if pixel value is 1, is mapped according to four adjacent pixels of the pixel for unit
To world coordinates matrix TheWorld；
Calculate long L and the width neighbor pixel of W, i.e., two that this four pixels are mapped to world coordinates matrix TheWorld
The corresponding realistic space actual distance of distance, cellar area is multiplied by pixel corresponding tested by computing unit area L × W
Height value in object height matrix, adds up obtain measured object volume one by one.Calculation formula is such as shown in (II)：
Volume=volume+L × W × height (II)
The foregoing description of the disclosed embodiments, enables professional and technical personnel in the field to realize or using the present invention.
A variety of modifications to these embodiments will be apparent for those skilled in the art, as defined herein
General Principle can be realized in other embodiments without departing from the spirit or scope of the present invention.Therefore, it is of the invention
The embodiments shown herein is not intended to be limited to, and is to fit to and principles disclosed herein and features of novelty phase one
The most wide scope caused.
Claims (7)
1. a kind of object volume computational methods based on Kinect, it is characterised in that specific steps include：
(1) foreground depth image and prospect coloured image comprising testee are gathered and not comprising measured object using Kinect
The background depth image and background color image of the test desk of body；
(2) chessboard calibration is carried out to Kinect colour imagery shot, passes through the relation between image coordinate system and world coordinate system, meter
Calculate the inner parameter and external parameter of Kinect colour imagery shots；
(3) background color image that the prospect coloured image of collection in step (1) is subtracted into collection obtains subtracting the region after difference, right
Subtract the region after difference and carry out image segmentation, obtain the bianry image of testee, prospect is set according to the actual area of test desk
ROI region and background ROI region；
(4) image preprocessing is carried out to the background ROI region obtained in step (3), fills the cavity of background ROI region, it is described
Image preprocessing includes mean filter, is filled, is iterated until all cavities with the average value of empty pixel neighbours thresholding
Pixel is filled and finished；
(5) foreground depth image and background depth image subtract obtaining testee height matrix after difference takes absolute value, and
Pretreatment is filtered to the testee height matrix；
(6) the testee height matrix computations testee height value according to step (5) after processing, according to Inverse projection
Principle, using the inner parameter and external parameter of step (2) the Kinect colour imagery shots, calculate testee corresponding
World coordinates matrix, the length of testee is calculated using the method for the minimum enclosed rectangle for finding world coordinates matrix；
(7) the testee height matrix described in world coordinates matrix according to step (6) and step (5) utilizes integration
Method calculates testee volume.
2. a kind of object volume computational methods based on Kinect according to claim 1, it is characterised in that the step
(1) object under test is placed on measuring table in, and IMAQ is carried out by the Kinect above measuring table.
3. a kind of object volume computational methods based on Kinect according to claim 1, it is characterised in that step (2)
It is described that chessboard calibration is carried out to Kinect video camera, the relation of image coordinate system and world coordinate system is determined, video camera is calculated
Inner parameter and external parameter, shown in all-purpose camera model such as formula (I)：
<mrow>
<mi>s</mi>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<mi>x</mi>
</mtd>
<mtd>
<mi>y</mi>
</mtd>
<mtd>
<mn>1</mn>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>=</mo>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<mi>X</mi>
</mtd>
<mtd>
<mi>Y</mi>
</mtd>
<mtd>
<mi>Z</mi>
</mtd>
<mtd>
<mn>1</mn>
</mtd>
</mtr>
</mtable>
</mfenced>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<mi>R</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mi>t</mi>
</mtd>
</mtr>
</mtable>
</mfenced>
<mi>K</mi>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mi>I</mi>
<mo>)</mo>
</mrow>
</mrow>
In formula (I), K is relevant with video camera internal structure, referred to as intrinsic parameters of the camera；R, t are sat with video camera relative to the world
The orientation for marking system is relevant, referred to as the external parameter of video camera, and R is spin matrix, and t is translation matrix, and [x y] is scene image
The coordinate of pixel, [X Y Z] is the coordinate of [x y] in world coordinate system.
4. a kind of object volume computational methods based on Kinect according to claim 3, it is characterised in that step (6)
The object length Size calculation is as follows:
The computational methods of (4a) world coordinates：According to the chessboard calibration data of step (2), using known parameters Z by universal camera shooting
Machine model formula (I) reverse goes out the corresponding world coordinate system coordinate (X, Y) of each point (x, y) on whole image coordinate system, and will calculate
As a result it is saved in tables of data Xcoord_in_World and Ycoord_in_World, so that the measured object in step (3)
Object height matrix in the bianry image and step (5) of body obtains the world of corresponding testee by way of tabling look-up
Coordinate；It is known that parameter Z is the element value height in the testee height matrix of acquisition in step (5), it calculates public
Formula is such as shown in (II)：
Height=Dground-Dobject (II)
Wherein, DgroundFor the element value in background depth image, the distance of Kinect each points to measuring table, D are representedobject
For the element value in foreground depth image, Kinect is represented to the distance of each point comprising testee；
(4b) object length and wide calculating：Specify world coordinates matrix maximum magnitude, newly-built world coordinates matrix
(XcoordRange, YcoordRange) and the element value for initializing world coordinates matrix is zero；Travel through the two-value of testee
Image, if the pixel value of bianry image is 1, tables look-up Xcoord_in_World and Ycoord_in_World according to step (4a)
Obtain correspondence world coordinates X and Y；And world coordinates matrix corresponding element value is set to 1；The world coordinates matrix obtained to calculating
Optimization is filtered, the world coordinates matrix minimum enclosed rectangle after filtering optimization is sought, if four summits of minimum enclosed rectangle
Coordinate be (X1, Y1), (X2, Y2), (X3, Y3), (X4, Y4), then the length of testee beThe width of testee is
5. a kind of object volume computational methods based on Kinect according to claim 1, it is characterised in that step (7)
The object volume is calculated as follows：
(5a) initialization volume volume is zero, travels through testee bianry image, if bianry image pixel value is 1, with this
Four adjacent pixels of pixel are that unit is mapped to world coordinates matrix；
(5b) calculates this four pixels and is mapped to the long L of world coordinates matrix and the distance pair of the width neighbor pixel of W, i.e., two
The realistic space actual distance answered, seeks cellar area L × W；
Cellar area is multiplied by the height value in the corresponding testee height matrix of the pixel by (5c), one by one add up obtain by
Object product is surveyed, its calculation formula is such as shown in (III)：
Volume=volume+L × W × height (III)
6. a kind of object volume computational methods based on Kinect according to claim 1, it is characterised in that step (5)
Described in filter preprocessing refer to the height value upper limit and height value lower limit that testee height matrix is set, outside bound
Numerical value is filtered into 0.
7. a kind of object volume computational methods based on Kinect according to claim 4, it is characterised in that step (4b)
Described in filtering optimization method for world coordinates matrix is narrowed down into 1/4 size, remove scattered make an uproar using morphology opening operation
Image is amplified into 4 times, the world coordinates matrix after being optimized again after point.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710027717.9A CN107067431B (en) | 2017-01-16 | 2017-01-16 | Kinect-based object volume calculation method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710027717.9A CN107067431B (en) | 2017-01-16 | 2017-01-16 | Kinect-based object volume calculation method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107067431A true CN107067431A (en) | 2017-08-18 |
| CN107067431B CN107067431B (en) | 2020-07-03 |
Family
ID=59597965
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710027717.9A Active CN107067431B (en) | 2017-01-16 | 2017-01-16 | Kinect-based object volume calculation method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107067431B (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108332662A (en) * | 2018-02-06 | 2018-07-27 | 新石器龙码（北京）科技有限公司 | A kind of object measuring method and device |
| CN108537329A (en) * | 2018-04-18 | 2018-09-14 | 中国科学院计算技术研究所 | A kind of method and apparatus carrying out operation using Volume R-CNN neural networks |
| CN108629756A (en) * | 2018-04-28 | 2018-10-09 | 东北大学 | A kind of Kinect v2 depth images Null Spot restorative procedure |
| CN109060830A (en) * | 2018-06-21 | 2018-12-21 | 珠海格力智能装备有限公司 | Method and device for detecting impurities of display screen |
| CN109186461A (en) * | 2018-07-27 | 2019-01-11 | 南京阿凡达机器人科技有限公司 | A kind of measurement method and measuring device of cabinet size |
| CN109357637A (en) * | 2018-12-11 | 2019-02-19 | 长治学院 | A kind of veneer reeling machine roll bending radius of curvature and method for measuring thickness based on depth camera |
| CN109448045A (en) * | 2018-10-23 | 2019-03-08 | 南京华捷艾米软件科技有限公司 | Plane polygon object measuring method and machine readable storage medium based on SLAM |
| CN109472822A (en) * | 2018-12-29 | 2019-03-15 | 上海大学 | Dimension of object measurement method based on depth image processing |
| CN109631764A (en) * | 2018-11-22 | 2019-04-16 | 南京理工大学 | Size Measuring System and method based on RealSense camera |
| CN110017773A (en) * | 2019-05-09 | 2019-07-16 | 福建（泉州）哈工大工程技术研究院 | A kind of package volume measuring method based on machine vision |
| CN110136193A (en) * | 2019-05-08 | 2019-08-16 | 广东嘉腾机器人自动化有限公司 | Cubold cabinet three-dimensional dimension measurement method and storage medium based on depth image |
| CN110148166A (en) * | 2019-01-25 | 2019-08-20 | 华晨宝马汽车有限公司 | Measure system, method, apparatus and the storage medium of dimension of object |
| CN110378952A (en) * | 2019-07-10 | 2019-10-25 | 深圳前海微众银行股份有限公司 | A kind of image processing method and device |
| CN110838142A (en) * | 2019-11-05 | 2020-02-25 | 沈阳民航东北凯亚有限公司 | Luggage size identification method and device based on depth image |
| CN111524193A (en) * | 2020-04-17 | 2020-08-11 | 西安交通大学 | Method and device for measuring two-dimensional size of object |
| CN112070736A (en) * | 2020-09-01 | 2020-12-11 | 上海电机学院 | Object volume vision measurement method combining target detection and depth calculation |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2008112766A1 (en) * | 2007-03-12 | 2008-09-18 | Conversion Works, Inc. | Systems and methods for allowing a user to dynamically manipulate stereoscopic parameters |
| US20090189889A1 (en) * | 2008-01-10 | 2009-07-30 | Siemens Aktiengesellschaft | Method and apparatus for visualizing 3D image data from tomographic imaging modalities |
| CN103207674A (en) * | 2013-03-21 | 2013-07-17 | 苏州展科光电科技有限公司 | Electronic demonstration system based on motion sensing technology |
| CN104408724A (en) * | 2014-11-27 | 2015-03-11 | 中南大学 | Depth information method and system for monitoring liquid level and recognizing working condition of foam flotation |
| CN104899883A (en) * | 2015-05-29 | 2015-09-09 | 北京航空航天大学 | Indoor object cube detection method for depth image scene |
| CN104951073A (en) * | 2015-06-19 | 2015-09-30 | 济南大学 | Gesture interaction method based on virtual interface |
| CN105243658A (en) * | 2015-09-08 | 2016-01-13 | 江苏大学 | Plant depth feature recognition method based on Kinect sensor |
-
2017
- 2017-01-16 CN CN201710027717.9A patent/CN107067431B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2008112766A1 (en) * | 2007-03-12 | 2008-09-18 | Conversion Works, Inc. | Systems and methods for allowing a user to dynamically manipulate stereoscopic parameters |
| US20090189889A1 (en) * | 2008-01-10 | 2009-07-30 | Siemens Aktiengesellschaft | Method and apparatus for visualizing 3D image data from tomographic imaging modalities |
| CN103207674A (en) * | 2013-03-21 | 2013-07-17 | 苏州展科光电科技有限公司 | Electronic demonstration system based on motion sensing technology |
| CN104408724A (en) * | 2014-11-27 | 2015-03-11 | 中南大学 | Depth information method and system for monitoring liquid level and recognizing working condition of foam flotation |
| CN104899883A (en) * | 2015-05-29 | 2015-09-09 | 北京航空航天大学 | Indoor object cube detection method for depth image scene |
| CN104951073A (en) * | 2015-06-19 | 2015-09-30 | 济南大学 | Gesture interaction method based on virtual interface |
| CN105243658A (en) * | 2015-09-08 | 2016-01-13 | 江苏大学 | Plant depth feature recognition method based on Kinect sensor |
Cited By (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108332662B (en) * | 2018-02-06 | 2020-02-11 | 新石器龙码（北京）科技有限公司 | Object measuring method and device |
| CN108332662A (en) * | 2018-02-06 | 2018-07-27 | 新石器龙码（北京）科技有限公司 | A kind of object measuring method and device |
| CN108537329A (en) * | 2018-04-18 | 2018-09-14 | 中国科学院计算技术研究所 | A kind of method and apparatus carrying out operation using Volume R-CNN neural networks |
| CN108629756A (en) * | 2018-04-28 | 2018-10-09 | 东北大学 | A kind of Kinect v2 depth images Null Spot restorative procedure |
| CN108629756B (en) * | 2018-04-28 | 2021-06-25 | 东北大学 | Kinectv2 depth image invalid point repairing method |
| CN109060830A (en) * | 2018-06-21 | 2018-12-21 | 珠海格力智能装备有限公司 | Method and device for detecting impurities of display screen |
| CN109060830B (en) * | 2018-06-21 | 2021-03-26 | 珠海格力智能装备有限公司 | Method and device for detecting impurities of display screen |
| CN109186461A (en) * | 2018-07-27 | 2019-01-11 | 南京阿凡达机器人科技有限公司 | A kind of measurement method and measuring device of cabinet size |
| CN109448045A (en) * | 2018-10-23 | 2019-03-08 | 南京华捷艾米软件科技有限公司 | Plane polygon object measuring method and machine readable storage medium based on SLAM |
| CN109631764A (en) * | 2018-11-22 | 2019-04-16 | 南京理工大学 | Size Measuring System and method based on RealSense camera |
| CN109631764B (en) * | 2018-11-22 | 2020-12-04 | 南京理工大学 | Dimension measuring system and method based on RealSense camera |
| CN109357637B (en) * | 2018-12-11 | 2021-12-10 | 长治学院 | Method for measuring curvature radius and thickness of plate rolling machine plate rolling based on depth camera |
| CN109357637A (en) * | 2018-12-11 | 2019-02-19 | 长治学院 | A kind of veneer reeling machine roll bending radius of curvature and method for measuring thickness based on depth camera |
| CN109472822A (en) * | 2018-12-29 | 2019-03-15 | 上海大学 | Dimension of object measurement method based on depth image processing |
| CN110148166A (en) * | 2019-01-25 | 2019-08-20 | 华晨宝马汽车有限公司 | Measure system, method, apparatus and the storage medium of dimension of object |
| CN110136193A (en) * | 2019-05-08 | 2019-08-16 | 广东嘉腾机器人自动化有限公司 | Cubold cabinet three-dimensional dimension measurement method and storage medium based on depth image |
| CN110017773A (en) * | 2019-05-09 | 2019-07-16 | 福建（泉州）哈工大工程技术研究院 | A kind of package volume measuring method based on machine vision |
| CN110378952A (en) * | 2019-07-10 | 2019-10-25 | 深圳前海微众银行股份有限公司 | A kind of image processing method and device |
| CN110838142A (en) * | 2019-11-05 | 2020-02-25 | 沈阳民航东北凯亚有限公司 | Luggage size identification method and device based on depth image |
| CN110838142B (en) * | 2019-11-05 | 2023-09-08 | 沈阳民航东北凯亚有限公司 | Luggage size recognition method and device based on depth image |
| CN111524193A (en) * | 2020-04-17 | 2020-08-11 | 西安交通大学 | Method and device for measuring two-dimensional size of object |
| CN111524193B (en) * | 2020-04-17 | 2022-05-03 | 西安交通大学 | Method and device for measuring two-dimensional size of object |
| CN112070736A (en) * | 2020-09-01 | 2020-12-11 | 上海电机学院 | Object volume vision measurement method combining target detection and depth calculation |
| CN112070736B (en) * | 2020-09-01 | 2023-02-24 | 上海电机学院 | Object volume vision measurement method combining target detection and depth calculation |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107067431B (en) | 2020-07-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107067431A (en) | A kind of object volume computational methods based on Kinect | |
| CN104501720B (en) | Non-contact object size and range image measuring instrument | |
| CN103530881B (en) | Be applicable to the Outdoor Augmented Reality no marks point Tracing Registration method of mobile terminal | |
| Li et al. | An improved building boundary extraction algorithm based on fusion of optical imagery and LIDAR data | |
| CN107945268A (en) | A kind of high-precision three-dimensional method for reconstructing and system based on binary area-structure light | |
| CN110174056A (en) | A kind of object volume measurement method, device and mobile terminal | |
| CN103605978A (en) | Urban illegal building identification system and method based on three-dimensional live-action data | |
| CN104665836A (en) | length measuring method and length measuring device | |
| CN106981091A (en) | Human body three-dimensional modeling data processing method and processing device | |
| CN106530337A (en) | Non local stereopair dense matching method based on image gray scale guiding | |
| CN105938619A (en) | Visual odometer realization method based on fusion of RGB and depth information | |
| CN110648362B (en) | Binocular stereo vision badminton positioning identification and posture calculation method | |
| CN104700385A (en) | Binocular vision positioning device based on FPGA | |
| CN105277144A (en) | Land area rapid detection method based on binocular vision and detection device thereof | |
| CN112489099A (en) | Point cloud registration method and device, storage medium and electronic equipment | |
| CN108805980A (en) | A kind of horse body measurement system and data correcting method based on 3D | |
| CN110207702B (en) | Target positioning method and device | |
| Abdul-Rahman et al. | Innovations in 3D geo information systems | |
| CN116071424A (en) | Fruit space coordinate positioning method based on monocular vision | |
| CN107705335A (en) | Demarcate the non-method that ken line sweeps laser range finder and measures camera orientation altogether | |
| CN106264537A (en) | The measurement system and method for human body attitude height in image | |
| CN106846324A (en) | A kind of irregular object height measurement method based on Kinect | |
| CN114569114A (en) | Height measuring method and device | |
| CN103399652A (en) | 3D (three-dimensional) input method on basis of OpenCV (open source computer vision library) camera calibration | |
| Han | Tree height measurement based on image processing with 3-points correction |
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 |