CN110220456A - A kind of hand-held box sizes measuring device and measuring method - Google Patents
A kind of hand-held box sizes measuring device and measuring method Download PDFInfo
- Publication number
- CN110220456A CN110220456A CN201910560263.0A CN201910560263A CN110220456A CN 110220456 A CN110220456 A CN 110220456A CN 201910560263 A CN201910560263 A CN 201910560263A CN 110220456 A CN110220456 A CN 110220456A
- Authority
- CN
- China
- Prior art keywords
- tested
- image
- treatment device
- depth camera
- intelligent treatment
- 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
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/02—Measuring arrangements characterised by the use of optical techniques for measuring length, width or thickness
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/02—Measuring arrangements characterised by the use of optical techniques for measuring length, width or thickness
- G01B11/06—Measuring arrangements characterised by the use of optical techniques for measuring length, width or thickness for measuring thickness ; e.g. of sheet material
- G01B11/0608—Height gauges
Abstract
The invention discloses a kind of hand-held box sizes measuring devices, comprising: depth camera, for the RGB image and depth image of tested cabinet to be shot and exported to tested cabinet；Intelligent treatment device is electrically connected with depth camera, is carried out image data acquiring for controlling depth camera, is handled and be calculated and be shown calculated result to the image data of depth camera acquisition；Hand-held support, for constant depth camera and intelligent treatment device, hand-held support includes rack body, the upper grip block being movably set in above rack body and the side grip block for being movably set in rack body left or right side, depth camera is fixed on rack body by upper grip block, and intelligent treatment device is fixed on rack body by side grip block.The measuring device miniature portable, it is convenient hand-held, the measurement process of box sizes volume is automatically completed, solves the problems, such as that one line addressee personnel of logistic industry generally require time-consuming and laborious manual measurement box sizes volume.
Description
Technical field
The present invention relates to size measurement technique fields, and in particular to a kind of hand-held box sizes measuring device and measurement side
Method.
Background technique
With the fast development of e-commerce, logistics box body of every number of days in terms of necessarily, which needs to receive, to be posted, transports.Logistic industry
It needs the dimension volume of rapid survey cabinet to calculate transportation cost, optimizes storage space.Current automatic contactless survey
Amount device usually utilizes the sensors such as linear laser, infrared light curtain, measures size volume by uniform scanning logistics box body, because
To be limited to detection mode, such device is limited only to the fixed platforms such as conveyer belt, transport rack, lacks portability.For a line
Logistics pickup personnel can only carry out the measurement of human contact's property, inefficiency to package using tools such as tape measures.Therefore, object at present
Popular industry extremely lacks a kind of logistics box body dimension volume self-operated measuring unit of miniature portable.
It is different from traditional RGB/ gray scale camera, depth camera can not only export RGB/ gray level image, moreover it is possible in utilization
The optical sensor acquisition range information output depth image set.The each pixel and RGB/ gray level image one of depth image are a pair of
Answer, have recorded pixel correspond to object to camera plane distance.I.e. be based on depth camera provides a kind of hand-held cabinet to the present invention
Dimension measuring device realizes size and the volume letter of the quick obtaining logistics box body under everyday scenes in conjunction with visual detection algorithm
Breath.
Summary of the invention
To overcome the above deficiencies, the invention provides a kind of hand-held box sizes measuring device and measurement sides
Method.
The present invention overcomes the technical solution used by its technical problem to be:
A kind of hand-held box sizes measuring device, comprising:
Depth camera, for the RGB image and depth image of tested cabinet to be shot and exported to tested cabinet；
Intelligent treatment device is electrically connected with depth camera, carries out image data acquiring, to depth for controlling depth camera
The image data of camera acquisition is handled and is calculated and be shown calculated result；And
Hand-held support is used for constant depth camera and intelligent treatment device, and the hand-held support includes rack body, activity
The upper grip block that is set to above rack body and the side grip block for being movably set in rack body left or right side, wherein
The depth camera is fixed on rack body by upper grip block, and the intelligent treatment device is fixed on branch by side grip block
On frame ontology.
Further, the upper plate that the upper grip block includes overhead gage and is fixedly connected with overhead gage, the side clamping
Plate includes side baffle and the side plate one and side plate two that are fixedly connected with side baffle, and the upper rear portion of the rack body is set
There are upper slot, back side middle part to be equipped with side slot one and side slot two, is set respectively in the upper slot, side slot one and side slot two
There are upper spring, lateral spring one and lateral spring two, one end of the slot bottom of slot, lateral spring one is fixed in one end of the upper spring
The slot bottom of side slot two is fixed in one end of the slot bottom, lateral spring two that are fixed on side slot one, and the free end of the upper plate is inserted
Enter in upper slot and be fixedly connected with the other end of upper spring, the free end of side plate one is inserted into side slot one and and lateral spring
One other end is fixedly connected, fixes company in the free end of side plate two insertion side slot two and with the other end of lateral spring two
It connects.
Further, the hand-held support further includes gripping handle, and the gripping handle is set on rack body and clamping
The opposite side of plate two.Further, the gripping handle is hollow structure, for accommodating at connection depth camera and intelligence
Manage the USB line beam of device.
Further, the intelligent treatment device is PDA or smart phone comprising interactive interface.
The present invention also provides a kind of measurement methods using hand-held box sizes measuring device described above, including
Steps are as follows:
S1, Image Acquisition: shooting tested cabinet by depth camera and guarantees to take three sides, depth phase
The RBG image and depth image of tested cabinet are transferred to intelligent treatment device by machine；
S2, target detection: intelligent treatment device carries out target detection to the RGB image of tested cabinet, obtains multiple candidates
Object matching region；
S3, image segmentation: intelligent treatment device reads the depth image of tested cabinet, rejects the erroneous matching in step S2
Region splits tested housing area from background image to get to the two dimensional image of tested housing area；
S4, three-dimensional point cloud processing: two dimensional image combination depth of the intelligent treatment device to the tested housing area being partitioned into
Image generates three-dimensional point cloud, then rejects outlier, then cloud is carried out it is down-sampled, obtain comprising tested three sides of cabinet and
The three-dimensional point cloud of bottom surface；
S5, dimension volume calculate: the distance that intelligent treatment device calculates tested box top to ground obtains tested cabinet
Height, the average distance for calculating data point to two sides on tested box top obtains the length and width of tested cabinet, logical
It crosses length, width and height and then obtains the volume of tested cabinet.
Further, in the step S2, using the Adaboost cascade classifier based on Haar feature to tested cabinet
RGB image carry out target detection, specifically include as follows:
S21, the Haar characteristic value construction Weak Classifier for calculating positive sample and negative sample, then pass through Adaboost algorithm and connect
Multiple Weak Classifiers form strong classifier, formula are as follows:
Wherein, h (x；am) it is Weak Classifier, βmFor Weak Classifier in strong classifier proportion, pass through each weak typing
Device linear, additive forms strong classifier FM(x；P)；
After S22, intelligent treatment device read the RGB image of depth camera shooting, the sliding window point of different scale is created
Not Ji Suan Haar characteristic value, by Adaboost cascade classifier Select Error be less than threshold value match window as candidate target
Matching area.
Further, in the step S3, the specific method for rejecting the erroneous matching region in step S2 includes the following:
S31, using prior information, i.e., three sides of tested cabinet are intersected in a wedge angle and the wedge angle is located at matching window
The middle part of mouth, wherein the wedge angle shows as local extremum in depth image；
S32, window Image (x is obtained by step S2start, ystart, xend, yend)；
S33, depth image is read in, and traversal finds minimum in the window described in step s 32, then judges minimum
It is worth region, if minimum is located in the middle part of window, matching is correct, if minimum is located at window edge, matching error is simultaneously
The erroneous matching is rejected.
Further, the step S4 is specifically included as follows:
S41, intelligent treatment device generate the two dimensional image combination depth image for the tested housing area being partitioned into three-dimensional
Point cloud, formula are as follows:
Wherein, Z (i, j) is depth information, can be directly read from depth map, fx、fyFor lens focus；
S42, the statistical zero-knowledge based on k nearest neighbor algorithm are filtered rejecting outlier, specifically include: calculating each point and arrive
The average distance of all K neighborhood points calculates the average and standard deviation of point set；Successively more each point rejects and is higher than threshold value
Point；
S43, cloud is carried out using voxel filter it is down-sampled.
Further, it in the step S5, first passes through RANSAC algorithm and obtains the flat of tested three sides of cabinet and ground
Face equation, then the dimension volume for being tested cabinet is calculated based on plane equation, the detailed process for obtaining plane equation includes:
S51, the target to be estimated of setting, i.e. plane equation: then ax+by+cz=d randomly chooses K point from cloud
As interior point set S and initialize the plane equation；
It is not belonging to remaining point of this K point in S52, traversal point cloud, the error between plane equation is calculated, if being less than threshold
Value is then added to interior point set S, expands interior point set S；
If S53, interior point set S quantity are greater than to being set the goal, reevaluate plane equation with interior point set S, otherwise, again with
Machine selects K point estimation plane equation；
Iteration above step S51-S53, until error is stablized.
The beneficial effects of the present invention are:
1, measuring device of the invention realizes the measurement of box sizes volume using small-sized depth camera, the measuring device
Miniature portable, it is convenient hand-held, the measurement process of box sizes volume is automatically completed, solves one line addressee of logistic industry
The problem of member generally requires time-consuming and laborious manual measurement box sizes volume, brings great convenience to it.
2, measurement method runing time of the present invention ensure that real-time, measurement accuracy error is 5% within 1s
Within, it ensure that accuracy, greatly improve work efficiency.
Detailed description of the invention
Fig. 1 is the decomposition texture schematic diagram of hand-held box sizes measuring device described in the embodiment of the present invention.
Fig. 2 is the overall structure diagram of hand-held box sizes measuring device described in the embodiment of the present invention.
Fig. 3 is the structural schematic diagram of hand-held support described in the embodiment of the present invention.
Fig. 4 is the use operation chart of hand-held box sizes measuring device described in the embodiment of the present invention.
In figure, 1, depth camera, 2, intelligent treatment device, 3, hand-held support, 31, rack body, 311, upper slot, 312,
Side slot one, 313, side slot two, 32, upper grip block, 321, overhead gage, 322, upper plate, 33, side grip block, 331, side block
Plate, 332, side plate one, 333, side plate two, 34, upper spring, 35, lateral spring one, 36, lateral spring two, 37, gripping handle.
Specific embodiment
For a better understanding of the skilled in the art, being done in the following with reference to the drawings and specific embodiments to the present invention
It is further described, it is following to be merely exemplary that the scope of protection of the present invention is not limited.
As shown in Figure 1, 2, a kind of hand-held box sizes measuring device described in the present embodiment, including depth camera 1, intelligence
It can processing unit 2 and hand-held support 3.
The depth camera 1 is used to shoot tested cabinet and export the RGB image and depth image of tested cabinet.
The intelligent treatment device 2 is electrically connected with depth camera 1 by USB line beam, carries out figure for controlling depth camera 1
As data acquisition, the image data acquired to depth camera 1 are handled and are calculated and be shown calculated result；The Intelligent treatment
Device 2 is PDA or smart phone comprising interactive interface and internal operation has dimension volume Measurement Algorithm software, the interaction
Interface carries out control shooting to depth camera 1 for operator to acquire image data.
The hand-held support 3 is used for constant depth camera 1 and intelligent treatment device 2, as shown in figure 3, the hand-held support 3
Including rack body 31, the upper grip block 32 that is movably set in the top of rack body 31, be movably set on the left of rack body or
The side grip block 33 on right side and on rack body 31 and the gripping handle 37 of the side opposite with side grip block 33；Its
In, the depth camera 1 is fixed on rack body 31 by upper grip block 32, and the intelligent treatment device 2 is clamped by side
Plate 33 is fixed on rack body 31.In the present embodiment, the upper grip block 32 includes overhead gage 321 and fixes company with overhead gage
The upper plate 322 connect, the side grip block 33 include side baffle 331 and the side plate 1 that is fixedly connected with side baffle and
Side plate 2 333, the upper rear portion of the rack body 31 is equipped with upper slot 311, back side middle part is equipped with side slot 1 and side
Slot 2 313 is respectively equipped with upper spring 34, lateral spring 1 in the upper slot 311, side slot 1 and side slot 2 313
With lateral spring 2 36, the slot bottom of slot 311 is fixed in one end of the upper spring 34, side is fixed in one end of lateral spring 1
The slot bottom of slot 1, one end of lateral spring 2 36 are fixed on the slot bottom of side slot 2 313, the free end of the upper plate 322
It is fixedly connected in insertion in slot 311 and with the other end of upper spring 34, the free end of side plate 1 insertion side slot one
Be fixedly connected in 312 and with the other end of lateral spring 1, in the free end of side plate 2 333 insertion side slot 2 313 and with
The other end of lateral spring 2 36 is fixedly connected.Further, the gripping handle 37 is hollow structure, for accommodating connection depth
The USB line beam of camera 1 and intelligent treatment device 2, specifically, one end of USB line beam connect with depth camera 1, the other end and intelligence
Energy processing unit 2 connects, and the USB line beam is used to the shooting instruction of intelligent treatment device 2 being transferred to depth camera 1 and will be deep
The photo that degree camera 1 is shot is transferred to intelligent treatment device 2 by USB line beam.
The present embodiment uses hand-held box sizes measuring device described above to measure with depth tested cabinet
For the RGB image and depth image data that camera 1 acquires as input, Output Size volume measurements are shown in Intelligent treatment dress
It sets on 2 interactive interface, the measurement method comprises the following steps that
Step S1, Image Acquisition
Tested cabinet is shot by depth camera 1 and guarantees to take three sides, as shown in figure 4, depth phase
The RBG image and depth image of tested cabinet are transferred to intelligent treatment device 2 by machine 1.
Step S2, target detection
Intelligent treatment device 2 carries out target detection to the RGB image of tested cabinet, obtains multiple candidate target Matching bands
Domain.In the present embodiment, target is carried out using RGB image of the Adaboost cascade classifier based on Haar feature to tested cabinet
Detection, specifically includes as follows:
S21, the Haar characteristic value construction Weak Classifier for calculating positive sample and negative sample, then pass through Adaboost algorithm and connect
Multiple Weak Classifiers form strong classifier, formula are as follows:
Wherein, h (x；am) it is Weak Classifier, βmFor Weak Classifier in strong classifier proportion, pass through each weak typing
Device linear, additive forms strong classifier FM(x；P)；In the present embodiment, using the picture of 1000 cabinets as positive sample, 10000
The cabinet picture of mixed and disorderly background is opened as negative sample, and training is used to based on Haar feature by positive sample and negative sample
Adaboost cascade classifier；
When S22, actual motion, the algorithm software of intelligent treatment device 2 is transported after reading the RGB image that depth camera 1 is shot
Row algorithm of target detection, the sliding window for creating different scale calculate separately Haar characteristic value, pass through Adaboost cascade sort
Device Select Error is less than the match window of threshold value as candidate target matching area；It is limited to algorithm performance, which can generate
Multiple match windows need to be further processed wherein including some erroneous matchings.
Step S3, image segmentation
Intelligent treatment device 2 reads the depth image of tested cabinet, rejects the erroneous matching region in step S2, will be tested
Housing area is split from background image to get to the two dimensional image of tested housing area.Wherein, it rejects in step S2
The specific method in erroneous matching region includes the following:
S31, using prior information, i.e., three sides of tested cabinet are intersected in a wedge angle and the wedge angle is located at matching window
The middle part of mouth, wherein the wedge angle shows as local extremum in depth image；
S32, window Image (x is obtained by step S2start, ystart, xend, yend)；
S33, depth image is read in, and traversal finds minimum in the window described in step s 32, then judges minimum
It is worth region, if minimum is located in the middle part of window, matching is correct, if minimum is located at window edge, matching error is simultaneously
The erroneous matching is rejected.
When actual motion, the algorithm software of intelligent treatment device 2 reads the depth image that depth camera 1 is shot, and checks step
The match window that rapid S32 is obtained rejects error hiding, then expands match window boundary according to the position of minimum in the window, makes
It can completely include logistics box body, complete image segmentation.
Step S4, three-dimensional point cloud is handled
Intelligent treatment device 2 generates three-dimensional point to the two dimensional image combination depth image for the tested housing area being partitioned into
Then cloud rejects outlier, then down-sampled to cloud progress, obtains the three-dimensional point comprising tested three sides of cabinet and bottom surface
Cloud specifically includes as follows:
S41, intelligent treatment device 2 generate the two dimensional image combination depth image for the tested housing area being partitioned into three-dimensional
Point cloud, formula are as follows:
Wherein, Z (i, j) is depth information, can be directly read from depth map, fx、fyFor lens focus；
S42, by Environmental Noise Influence, the point cloud directly generated there are outlier, need by filtering eliminate；Algorithm uses
Statistical zero-knowledge based on k nearest neighbor algorithm is filtered rejecting outlier, the working principle of the filter are as follows: (1) calculates each
Point arrives the average distance of all K neighborhood points, and (2) calculate the average and standard deviation of point set, and (3) successively more each point is rejected
Higher than the point of threshold value；
S43, to guarantee real-time, need to carry out cloud down-sampled, algorithm uses voxel filter, whole in holding point cloud
Point cloud scale, the working principle of the filter are as follows: be divided into a cloud small one by one are cut down on the basis of body geometry
Then space three-dimensional cube substitutes other points with the center of gravity of all the points in three-dimensional cubic body.
Step S5, dimension volume calculates
The distance that intelligent treatment device 2 calculates tested box top to ground obtains the height of tested cabinet, calculates tested
The average distance of data point to two sides obtains the length and width of tested cabinet on box top, passes through length, width and height
It spends and then obtains the volume of tested cabinet.Specifically, it first passes through RANSAC algorithm and obtains tested three sides of cabinet and ground
Plane equation, then the dimension volume for being tested cabinet is calculated based on plane equation, the detailed process for obtaining plane equation includes:
S51, the target to be estimated of setting, i.e. plane equation: then ax+by+cz=d randomly chooses K point from cloud
As interior point set S and initialize the plane equation；
It is not belonging to remaining point of this K point in S52, traversal point cloud, the error between plane equation is calculated, if being less than threshold
Value is then added to interior point set S, expands interior point set S；
If S53, interior point set S quantity are greater than to being set the goal, reevaluate plane equation with interior point set S, otherwise, again with
Machine selects K point estimation plane equation；
Iteration above step S51-S53, until error is stablized.
When actual motion, the plane equation on tested three sides of cabinet and ground is successively obtained using RANSAC algorithm, so
The distance for calculating tested box top to ground afterwards obtains the height of tested cabinet, calculates and is tested on box top data point to two
The average distance of side and then the length and width for obtaining tested cabinet, after the three-dimensional dimension measurement for completing tested cabinet, in turn
The volume of available tested cabinet.
Within 1s, measurement accuracy error ensure that in real time algorithm total run time described in the present embodiment within 5%
Property and accuracy.
Above only describes basic principle of the invention and preferred embodiment, those skilled in the art can be according to foregoing description
Many changes and improvements are made, these changes and improvements should be within the scope of protection of the invention.
Claims (10)
1. a kind of hand-held box sizes measuring device characterized by comprising
Depth camera, for the RGB image and depth image of tested cabinet to be shot and exported to tested cabinet；
Intelligent treatment device is electrically connected with depth camera, carries out image data acquiring, to depth camera for controlling depth camera
The image data of acquisition is handled and is calculated and be shown calculated result；And
Hand-held support, is used for constant depth camera and intelligent treatment device, and the hand-held support includes rack body, activity setting
Upper grip block above rack body and the side grip block for being movably set in rack body left or right side, wherein described
Depth camera is fixed on rack body by upper grip block, and the intelligent treatment device is fixed on bracket sheet by side grip block
On body.
2. hand-held box sizes measuring device according to claim 1, which is characterized in that the upper grip block includes upper
Baffle and the upper plate being fixedly connected with overhead gage, the side grip block include side baffle and the side that is fixedly connected with side baffle
Plate one and side plate two, the upper rear portion of the rack body is equipped with upper slot, back side middle part is equipped with side slot one and side is inserted
Slot two, is respectively equipped with upper spring, lateral spring one and lateral spring two in the upper slot, side slot one and side slot two, it is described on
The slot bottom of slot is fixed in one end of spring, the slot bottom of side slot one is fixed in one end of lateral spring one, lateral spring two one
The slot bottom of side slot two is fixed at end, and the free end of the upper plate, which is inserted into upper slot and is fixed with the other end of upper spring, to be connected
It connects, be fixedly connected in the free end of side plate one insertion side slot one and with the other end of lateral spring one, the freedom of side plate two
It is fixedly connected in end insertion side slot two and with the other end of lateral spring two.
3. hand-held box sizes measuring device according to claim 1 or 2, which is characterized in that the hand-held support is also
Including gripping handle, the gripping handle is set to side opposite with grip block two on rack body.
4. hand-held box sizes measuring device according to claim 3, which is characterized in that the gripping handle is hollow
Structure, for accommodating the USB line beam of connection depth camera and intelligent treatment device.
5. hand-held box sizes measuring device according to claim 1, which is characterized in that the intelligent treatment device is
PDA or smart phone comprising interactive interface.
6. a kind of measurement method using the described in any item hand-held box sizes measuring devices of claim 1-5, feature
It is, comprises the following steps that
S1, Image Acquisition: shooting tested cabinet by depth camera and guarantees to take three sides, and depth camera will
The RBG image and depth image of tested cabinet are transferred to intelligent treatment device；
S2, target detection: intelligent treatment device carries out target detection to the RGB image of tested cabinet, obtains multiple candidate targets
Matching area；
S3, image segmentation: intelligent treatment device reads the depth image of tested cabinet, rejects the erroneous matching area in step S2
Domain splits tested housing area from background image to get to the two dimensional image of tested housing area；
S4, three-dimensional point cloud processing: two dimensional image combination depth image of the intelligent treatment device to the tested housing area being partitioned into
Three-dimensional point cloud is generated, outlier is then rejected, then is down-sampled to cloud progress, is obtained comprising tested three sides of cabinet and bottom surface
Three-dimensional point cloud；
S5, dimension volume calculate: the distance that intelligent treatment device calculates tested box top to ground obtains the height of tested cabinet
Degree, the average distance for calculating data point to two sides on tested box top obtain the length and width of tested cabinet, pass through length
Degree, width and height and then the volume for obtaining tested cabinet.
7. measurement method according to claim 6, which is characterized in that in the step S2, using based on Haar feature
Adaboost cascade classifier carries out target detection to the RGB image of tested cabinet, specifically includes as follows:
S21, the Haar characteristic value for calculating positive sample and negative sample construct Weak Classifier, then are connected by Adaboost algorithm multiple
Weak Classifier forms strong classifier, formula are as follows:
Wherein, h (x；am) it is Weak Classifier, βmFor Weak Classifier in strong classifier proportion, pass through each Weak Classifier line
Property be added composition strong classifier FM(x；P)；
After S22, intelligent treatment device read the RGB image of depth camera shooting, the sliding window for creating different scale is counted respectively
Haar characteristic value is calculated, is matched by the match window that Adaboost cascade classifier Select Error is less than threshold value as candidate target
Region.
8. measurement method according to claim 7, which is characterized in that in the step S3, reject the mistake in step S2
The specific method of matching area includes the following:
S31, using prior information, i.e., three sides of tested cabinet are intersected in a wedge angle and the wedge angle is located at match window
Middle part, wherein the wedge angle shows as local extremum in depth image；
S32, window Image (x is obtained by step S2start, ystart, xend, yend)；
S33, depth image is read in, and traversal finds minimum in the window described in step s 32, then judges minimum institute
In region, if minimum is located in the middle part of window, matching is correct, if minimum is located at window edge, matching error simultaneously should
Erroneous matching is rejected.
9. measurement method according to claim 8, which is characterized in that the step S4 specifically includes as follows:
S41, intelligent treatment device generate three-dimensional point cloud to the two dimensional image combination depth image for the tested housing area being partitioned into,
Formula are as follows:
Wherein, Z (i, j) is depth information, can be directly read from depth map, fx、fyFor lens focus；
S42, the statistical zero-knowledge based on k nearest neighbor algorithm are filtered rejecting outlier, specifically include: calculating each point to all
The average distance of K neighborhood point calculates the average and standard deviation of point set；Successively more each point rejects the point for being higher than threshold value；
S43, cloud is carried out using voxel filter it is down-sampled.
10. measurement method according to claim 9, which is characterized in that in the step S5, first pass through RANSAC algorithm and obtain
It must be tested the plane equation on three sides of cabinet and ground, then calculate the dimension volume of tested cabinet based on plane equation, it is described
Obtain plane equation detailed process include:
S51, the target to be estimated of setting, i.e. plane equation: then ax+by+cz=d randomly chooses K point conduct from cloud
Interior point set S simultaneously initializes the plane equation；
It is not belonging to remaining point of this K point in S52, traversal point cloud, calculates the error between plane equation, if being less than threshold value,
It is added to interior point set S, expands interior point set S；
If S53, interior point set S quantity are greater than to setting the goal, plane equation is reevaluated with interior point set S, otherwise, random choosing again
Select K point estimation plane equation；
Iteration above step S51-S53, until error is stablized.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910560263.0A CN110220456A (en) | 2019-06-26 | 2019-06-26 | A kind of hand-held box sizes measuring device and measuring method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910560263.0A CN110220456A (en) | 2019-06-26 | 2019-06-26 | A kind of hand-held box sizes measuring device and measuring method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110220456A true CN110220456A (en) | 2019-09-10 |
Family
ID=67814746
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910560263.0A Pending CN110220456A (en) | 2019-06-26 | 2019-06-26 | A kind of hand-held box sizes measuring device and measuring method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110220456A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112683167A (en) * | 2020-12-17 | 2021-04-20 | 上海嗨酷强供应链信息技术有限公司 | Based on AR volume side equipment |
| CN112697042A (en) * | 2020-12-07 | 2021-04-23 | 深圳市繁维科技有限公司 | Handheld TOF camera and strong-adaptability wrapping volume measuring method thereof |
| CN113362385A (en) * | 2021-06-21 | 2021-09-07 | 山东产研信息与人工智能融合研究院有限公司 | Cargo volume measuring method and device based on depth image |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2014033614A1 (en) * | 2012-08-27 | 2014-03-06 | Koninklijke Philips N.V. | Patient-specific and automatic x-ray system adjustment based on optical 3d scene detection and interpretation |
| CN205610770U (en) * | 2016-04-29 | 2016-09-28 | 深圳市大疆创新科技有限公司 | Shoot device, shooting equipment and portable electronic equipment |
| CN106251353A (en) * | 2016-08-01 | 2016-12-21 | 上海交通大学 | Weak texture workpiece and the recognition detection method and system of three-dimensional pose thereof |
| CN107767456A (en) * | 2017-09-22 | 2018-03-06 | 福州大学 | A kind of object dimensional method for reconstructing based on RGB D cameras |
| CN108734087A (en) * | 2018-03-29 | 2018-11-02 | 京东方科技集团股份有限公司 | Object automatic identifying method and system, shopping apparatus and storage medium |
| CN108981570A (en) * | 2018-07-12 | 2018-12-11 | 浙江大学 | A kind of portable type physical distribution package volume measurement device |
| CN109584297A (en) * | 2018-10-24 | 2019-04-05 | 北京升哲科技有限公司 | Object detection method and device |
| CN109801333A (en) * | 2019-03-19 | 2019-05-24 | 北京华捷艾米科技有限公司 | Volume measuring method, device, system and calculating equipment |
| CN109886961A (en) * | 2019-03-27 | 2019-06-14 | 重庆交通大学 | Medium-and-large-sized measurement of cargo measurement method based on depth image |
-
2019
- 2019-06-26 CN CN201910560263.0A patent/CN110220456A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2014033614A1 (en) * | 2012-08-27 | 2014-03-06 | Koninklijke Philips N.V. | Patient-specific and automatic x-ray system adjustment based on optical 3d scene detection and interpretation |
| CN205610770U (en) * | 2016-04-29 | 2016-09-28 | 深圳市大疆创新科技有限公司 | Shoot device, shooting equipment and portable electronic equipment |
| CN106251353A (en) * | 2016-08-01 | 2016-12-21 | 上海交通大学 | Weak texture workpiece and the recognition detection method and system of three-dimensional pose thereof |
| CN107767456A (en) * | 2017-09-22 | 2018-03-06 | 福州大学 | A kind of object dimensional method for reconstructing based on RGB D cameras |
| CN108734087A (en) * | 2018-03-29 | 2018-11-02 | 京东方科技集团股份有限公司 | Object automatic identifying method and system, shopping apparatus and storage medium |
| CN108981570A (en) * | 2018-07-12 | 2018-12-11 | 浙江大学 | A kind of portable type physical distribution package volume measurement device |
| CN109584297A (en) * | 2018-10-24 | 2019-04-05 | 北京升哲科技有限公司 | Object detection method and device |
| CN109801333A (en) * | 2019-03-19 | 2019-05-24 | 北京华捷艾米科技有限公司 | Volume measuring method, device, system and calculating equipment |
| CN109886961A (en) * | 2019-03-27 | 2019-06-14 | 重庆交通大学 | Medium-and-large-sized measurement of cargo measurement method based on depth image |
Non-Patent Citations (1)
| Title |
|---|
| 赵兴东等: "《矿用三维激光数字测量原理及其工程应用》", 31 January 2016, 冶金工业出版社 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112697042A (en) * | 2020-12-07 | 2021-04-23 | 深圳市繁维科技有限公司 | Handheld TOF camera and strong-adaptability wrapping volume measuring method thereof |
| CN112697042B (en) * | 2020-12-07 | 2023-12-05 | 深圳市繁维科技有限公司 | Handheld TOF camera and method for measuring volume of package by using same |
| CN112683167A (en) * | 2020-12-17 | 2021-04-20 | 上海嗨酷强供应链信息技术有限公司 | Based on AR volume side equipment |
| CN113362385A (en) * | 2021-06-21 | 2021-09-07 | 山东产研信息与人工智能融合研究院有限公司 | Cargo volume measuring method and device based on depth image |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109801337B (en) | 6D pose estimation method based on instance segmentation network and iterative optimization | |
| US11720766B2 (en) | Systems and methods for text and barcode reading under perspective distortion | |
| CN110220456A (en) | A kind of hand-held box sizes measuring device and measuring method | |
| Lowe | Distinctive image features from scale-invariant keypoints | |
| CN109102547A (en) | Robot based on object identification deep learning model grabs position and orientation estimation method | |
| CN108381549B (en) | Binocular vision guide robot rapid grabbing method and device and storage medium | |
| CN109213202B (en) | Goods placement method, device, equipment and storage medium based on optical servo | |
| CN113850865A (en) | Human body posture positioning method and system based on binocular vision and storage medium | |
| CN107704867A (en) | Based on the image characteristic point error hiding elimination method for weighing the factor in a kind of vision positioning | |
| Ansari | A review on SIFT and SURF for underwater image feature detection and matching | |
| CN105979248B (en) | Image processing system and its operating method with interacting depth estimation | |
| Deepak et al. | Identification of fake notes and denomination recognition | |
| Netz et al. | Recognition using specular highlights | |
| CN113021355B (en) | Agricultural robot operation method for predicting sheltered crop picking point | |
| Peng et al. | Robot visual guide with Fourier-Mellin based visual tracking | |
| Thammasorn et al. | Real-time method for counting unseen stacked objects in mobile | |
| Madessa et al. | Leveraging an instance segmentation method for detection of transparent materials | |
| Zia et al. | Exploring chromatic aberration and defocus blur for relative depth estimation from monocular hyperspectral image | |
| Slossberg et al. | Freehand Laser Scanning Using Mobile Phone. | |
| CN110728222B (en) | Pose estimation method for target object in mechanical arm grabbing system | |
| CN111507177A (en) | Identification method and device for metering turnover cabinet | |
| CN115457494A (en) | Object identification method and system based on infrared image and depth information fusion | |
| Kim et al. | The coil recognition system for an unmanned crane using stereo vision | |
| Nakashima et al. | Constructing illumination image basis from object motion | |
| Bleier et al. | Towards robust self-calibration for handheld 3D line laser scanning |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190910 |
|
| WD01 | Invention patent application deemed withdrawn after publication |