CN106127727B - A kind of domestic animal body surface three-dimensional data acquisition methods - Google Patents
A kind of domestic animal body surface three-dimensional data acquisition methods Download PDFInfo
- Publication number
- CN106127727B CN106127727B CN201610395906.7A CN201610395906A CN106127727B CN 106127727 B CN106127727 B CN 106127727B CN 201610395906 A CN201610395906 A CN 201610395906A CN 106127727 B CN106127727 B CN 106127727B
- Authority
- CN
- China
- Prior art keywords
- domestic animal
- point cloud
- body surface
- animal body
- dimensional
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- 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
- G06T2207/10028—Range image; Depth image; 3D point clouds
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30181—Earth observation
- G06T2207/30188—Vegetation; Agriculture
Abstract
The present invention relates to three-dimensional (3D) computer vision field, especially a kind of domestic animal body surface three-dimensional data acquisition methods based on inexpensive single depth camera.The step of this method includes plane monitoring-network module, principal component local coordinate system detection module, coordinate rectification module and mirror module.The present invention is to provide a kind of domestic animal body surface three-dimensional data fast acquiring methods, utilize unilateral domestic animal body surface point cloud data, points cloud processing technology and computer graphics techniques, acquire complete domestic animal body surface point cloud data, equipment cost and human cost are reduced, versatility, flexibility and the degree of automation are improved.
Description
Technical field
The present invention relates to three-dimensional (3D) computer vision fields more particularly to a kind of based on inexpensive single depth camera
Domestic animal body surface three-dimensional data fast acquiring method.
Background technique
Animal body measurement is the important component of domestic animal appearance evaluation, and the figure and features of the domestic animals such as cattle,pig and sheep, which is evaluated, at present needs
The body measurement index to be carried out is more, and it is huge to be obtained and measured workload by the way of artificial, to contactless dynamic
Object ruler auto acquisition system, which carries out research, has certain realistic meaning, and the acquisition of contactless animal information is that research is non-
The basis of contact animal body ruler auto acquisition system.
The research of current domestic animal acquisition of information focuses mostly in two-dimentional field, domestic to study two-dimensional domestic animal acquisition of information
It is more, it focuses mostly in pig, the medium-and-large-sized livestock such as ox, if publication number is CN200710119509.8 respectively,
The patent of CN201410096476.X.But with the development of science and technology, three-dimensional (3D) computer vision has been popularized, for dynamic
The acquisition demand of object table three-dimensional data increases, and especially not only becomes parsing family dirty swine to the accurate measurement of domestic animal character and measurement
One of key of life rule, exploitation disease control methods, while being also beneficial to solve phenotype and genotype detection and being associated with something lost
The problem in science in mechanism analysis is passed, it is therefore necessary to propose a kind of domestic animal body surface three-dimensional data fast acquiring method, thus real
The quick obtaining of existing domestic animal body surface three-dimensional data.
Summary of the invention
The present invention is to provide a kind of simple, quick domestic animal body surface three-dimensional data acquisition methods.
Method includes the following steps:
A) point cloud data inputs, and the domestic animal body surface three-dimensional point cloud number of the most sides comprising ground is obtained using depth camera
According to, and input point cloud data;
B) plane monitoring-network carries out the cloud detection of floor point to the point cloud data inputted in the step a), deletes ground
Point cloud obtains the ground normal vector for being directed toward domestic animal body surface three dimensional point cloud by eliminating normal vector ambiguity;
C) local coordinate system is established, and is carried out feature extraction to the domestic animal body surface three dimensional point cloud of most sides, is obtained domestic animal
Center of gravity and orthogonal three reference axis feature vectors, are denoted as X-axis, Y-axis, Z axis respectively, so that establishing with domestic animal center of gravity is original
The local coordinate system of point;
D) point cloud coordinate conversion, calculates rotational translation matrix, and according to rotational translation matrix to domestic animal body surface three-dimensional point cloud
The coordinate of data carries out coordinate conversion, while snapping in global coordinate system;
E) rotational correction is corrected in the step d) in conjunction with the ground normal vector obtained in the step b) through sitting
The coordinate of domestic animal body surface three dimensional point cloud after mark conversion;
F) root of the tail midpoint detection, by judging that the plane perpendicular to X-axis intersects acquisition with domestic animal body surface three dimensional point cloud
Point cloud quantity, determine the X axis coordinate at root of the tail midpoint;
G) translationai correction translates rotation in the step e) and rectifys according to the X axis coordinate at the root of the tail midpoint in the step f)
The coordinate of domestic animal body surface three dimensional point cloud after just;
H) mirror image processing merely enters the domestic animal body surface three dimensional point cloud that Z axis is greater than zero, is with X-axis-Y axis coordinate plane
The plane of symmetry does mirror image processing, to obtain complete domestic animal body surface three dimensional point cloud.
Preferably, the domestic animal body surface three dimensional point cloud of the most sides in the step a) comprising ground does not include shooting back
Scape.
Preferably, in the step b) plane monitoring-network, it is as follows to eliminate normal vector ambiguity method: it is flat to first pass through the ground
Face detects two normal vectors, then along the direction of described two normal vectors, apart from the floor point cloud distance
It for the position of Nr, does two parallel planes and intersects respectively with domestic animal body surface three dimensional point cloud, when parallel along a normal vector direction
When the plane of the floor intersects the point cloud quantity obtained with domestic animal point cloud and is greater than given threshold n, it is determined that the method
Vector is the ground normal vector for being directed toward domestic animal body surface three dimensional point cloud.
Preferably, during the step c) local coordinate system is established, feature extraction uses Principal Component Analysis, i.e. PCA method.
Preferably, the rotational translation matrix in step d) the point cloud coordinate conversion, passes through the step c) local coordinate
The three reference axis feature vectors and domestic animal center of gravity obtained in system's foundation are constructed.
Preferably, in the step e) rotational correction, by calculating the ground normal vector obtained in the step b)
With the angle between existing coordinate Z axis, construct spin matrix, then according to spin matrix in the step d) through a cloud coordinate
The coordinate of domestic animal body surface three dimensional point cloud after conversion is rotated.
Preferably, in step d) the root of the tail midpoint detection, the judgment method of the X-axis position at the root of the tail midpoint is edge
Domestic animal point cloud X-direction minimum value is at a certain distance interval, makees the plane vertical with X-axis respectively, until domestic animal body surface is three-dimensional
The maximum value of point cloud data X-direction, when the point cloud quantity for intersecting acquisition with domestic animal body surface three dimensional point cloud by plane is small
When some threshold value n, recording x value locating for the plane is Xn, and it is maximum to obtain the Y-axis that the plane at Xn intersects with domestic animal point cloud
Point ymaxWith Y-axis smallest point ymin, the y value for calculating root of the tail midpoint is ymedium=(ymax+ymin)/2。
Preferably, in the step g) translationai correction, the Y direction distance of the translation is the y at the root of the tail midpoint
Value.
Preferably, the mirror image processing of the step h) includes that domestic animal body surface three dimensional point cloud is cut and domestic animal body surface three-dimensional
Point cloud data symmetrical treatment.
Above-mentioned technical proposal of the invention has the advantages that provided by the invention based on inexpensive single depth camera
Domestic animal body surface three-dimensional data fast acquiring method simply and efficiently solves the accuracy that previous domestic animal body surface three-dimensional data obtains
Low, the problems such as equipment cost is high, data volume is big.Data volume needed for the present invention halves, it is only necessary to which the half side point cloud data of target domestic animal is
Can, this greatly reduces the difficulty and equipment cost of data acquisition.And it can make to calculate more by modified hydrothermal process in the present invention
Accurately, all data of more objective surveyed domestic animal is obtained.
Detailed description of the invention
Fig. 1 is that the domestic animal body surface three-dimensional data provided in an embodiment of the present invention based on inexpensive single depth camera quickly obtains
Take the schematic diagram of method;
Fig. 2 is that the domestic animal body surface three-dimensional data provided in an embodiment of the present invention based on inexpensive single depth camera quickly obtains
Take method flow diagram;
Fig. 3 is the specific workflow figure of Fig. 2;
Fig. 4 a-4b is the tested domestic animal body surface three dimensional point cloud figure of plane monitoring-network segmentation front and back;
Fig. 5 a-5b is the location drawing that PCA conversion front and back is tested domestic animal body surface three dimensional point cloud;
Fig. 6 a-6b is the location drawing of tested domestic animal body surface three dimensional point cloud before and after rotation transformation;
Fig. 7 a-7b is the location drawing of tested domestic animal body surface three dimensional point cloud before and after translation transformation;
Fig. 8 a-8b is the location drawing of tested domestic animal body surface three dimensional point cloud before and after mirror image.
Specific embodiment
Specific embodiments of the present invention will be described in further detail with reference to the accompanying drawings and examples.Following embodiment
For illustrating the present invention, but it is not intended to limit the scope of the invention.
As shown in Figure 1, the embodiment of the invention provides a kind of domestic animal body surface based on inexpensive single depth camera is three-dimensional
Data fast acquiring method, this method comprises:
First step S001, point cloud data input, the domestic animal body surface of the most sides comprising ground is obtained using depth camera
Three dimensional point cloud, and these point cloud datas are inputted.
Second step S002, plane monitoring-network detect the domestic animal body surface three dimensional point cloud of the above-mentioned most sides comprising ground (no
Including shooting background), floor point cloud is detected, ground point cloud is deleted, by eliminating ambiguity, obtains and is directed toward domestic animal point cloud
Normal vector.
Third step S003, local coordinate system are established, and are carried out feature extraction to the half side body surface point cloud data of target domestic animal, are obtained
Domestic animal center of gravity and three reference axis feature vectors perpendicular to each other, are denoted as X-axis, Y-axis, Z axis respectively, are established whereby with domestic animal center of gravity
For the space coordinates of origin;
4th step S004, point cloud coordinate conversion, according to rotational translation matrix, converts domestic animal point cloud coordinate, by it
Snap to global coordinate system.
5th step S005, rotational correction, the domestic animal in conjunction with the ground normal vector obtained, after correcting the coordinate conversion
Point cloud coordinate.
6th step S006, root of the tail midpoint detection intersect the point cloud number obtained with domestic animal point cloud by judging the plane along X-axis
Amount, determines the X-axis position at root of the tail midpoint.
7th step S007, translationai correction, the family according to the X axis coordinate at the root of the tail midpoint, after translating the rotational correction
Poultry point cloud.
8th step S008, mirror module, i.e., complete domestic animal point cloud obtain, and the domestic animal data that Z axis is greater than zero are merely entered, with X
Axis-Y axis coordinate plane is that the plane of symmetry does mirror image processing, obtains complete domestic animal point cloud data.
As Figure 2-3, domestic animal body surface three dimension provided in an embodiment of the present invention based on inexpensive single depth camera
It is described as follows according to fast acquiring method specific workflow figure:
Step S0, plane monitoring-network module is first carried out.Step S0 includes step S01, step S02 and step S03:
Step S01 is executed, ground point cloud detection simultaneously deletes ground point cloud.In this example, input is more comprising ground
Half side domestic animal body surface three dimensional point cloud, is denoted as Pall, as shown in fig. 4 a.The present embodiment is by taking pig as an example, CtBe input include
The global coordinate system of the domestic animal body surface three dimensional point cloud of the most sides on ground, is denoted as Pall={ pi| i=0,1 ..., Ns, it utilizes
Stochastical sampling consistency algorithm RANSAC, the planar point cloud detected are set as pplaneBy to PallAnd pplaneCarry out difference fortune
It calculates, obtains the domestic animal point cloud data not comprising ground dataAs shown in Figure 4 b.
Step S02 is executed, ground normal vector extracts.It is mentioned using the plane formula Ax+By+Cz+D=0 obtained in S01
Follow the example of vector n 1 and n2.
Step S03 is executed, normal vector ambiguity is eliminated.Two extracted along the extraction of the ground S02 normal vector respectively
The direction that normal vector n1 and n2 are directed toward, apart from the plan range p detectedplaneFor Nr, do two parallel planes
pplane1And pplane2Respectively at domestic animal point cloud data PallIntersection.If along a certain normal direction apart from the detection plan range
pplaneFor NrPlane and domestic animal point cloud PallWhen the point cloud quantity that intersection obtains is greater than threshold value num, we determine that the direction
Normal vector is the ground normal vector n for being directed toward domestic animal body surface three dimensional point cloud, in the present embodiment Nr=0.1 meter, num is
1000。
Then step S1, principal component local coordinate system detection module are executed.Step S1 includes step S11 and step S12:
Step S11 is executed, focus point calculates.Utilize formulaInstitute invocation point pmAs center of gravity, wherein n
For the point set quantity of the obtained domestic animal body surface three dimensional point cloud not comprising ground data.
Step S12 is executed, principal component characteristic value and feature vector (principal component local coordinate reference axis) obtain.Building is not wrapped
The covariance matrix of domestic animal point cloud data P containing ground data:
The matrix is symmetrical real matrix, seeks its non-negative eigenvalue λ1、λ2、λ3。
Utilize formula Cp ei=λl ei
Acquire e1、e2、e3, that is, obtain principal component local coordinate system.In the present embodiment, e1For not comprising ground data
Domestic animal point cloud data P spatially orders the most intensive direction of cloud, is set as X-axis;e2To cross focus point and and e1Vertical Plane-point
The most intensive direction of cloud, is set as Y-axis;e3It is determined using the right-hand rule, is set as Z axis.
Then step S2, coordinate rectification module are executed.Step S2 includes step S21, step S22, step S23 and step
S24:
Execute step S21, coordinate conversion.Utilize the characteristic value e of acquisition1、e2、e3, construct spin matrix R=(e1、e2、
e3), then to each of the obtained domestic animal point cloud data P not comprising ground data vertex pi, utilize following formula:
pi=R (pi-pm)+pm, (i=1,2,3 ..., n)
P is obtained after conversiontransform, as shown in Figure 5 b, the tested domestic animal body surface three dimensional point cloud location drawing before conversion is such as
Shown in Fig. 5 a.
Execute step S22: rotational correction.In conjunction with the normal vector perpendicular to ground point cloud of acquisition, formula is utilized:
Acquire the angle a of rotation.
Then the angle a, structural matrix R of rotation are utilizedtransform,
For the point set P after conversiontransformEach point (x, y, z), construct vector (x, y, z, 1), according to what is acquired
Structural matrix Rtransform, converted using following formula.
(x ', y ', z ', 1)=(x, y, z, 1) * Rtransform
Point set R using (x ', y ', z ', 1) vector after conversion, after obtaining rotational correction conversionrote, as shown in Figure 6 b,
The tested domestic animal body surface three dimensional point cloud location drawing before rotational correction conversion is as shown in Figure 6 a.
Execute step S23: root of the tail midpoint detection.The judgment method of the X-axis position at root of the tail midpoint is, along domestic animal point cloud X-axis
Direction minimum value, at a certain distance XcorFor interval, make the plane vertical with X-axis respectively, until domestic animal point cloud X-direction most
Big value records locating for the plane when intersecting the point cloud quantity obtained with domestic animal point cloud by plane less than some threshold value n
X value is Xn, obtain XnThe Z axis maximum point z that the plane at place intersects with pig body point cloudmaxWith Z axis smallest point zmin, calculate root of the tail midpoint
Z value be zmedium=(zmax+zmin)/2.In the present embodiment, xcorIt is set as 1cm.
Execute step S24: translationai correction.Utilize the z obtainedmediumWith the point set R after rotational correction conversionrote, such as
Shown in Fig. 7 a, for RroteThe three-dimensional coordinate (x, y, z) of each of set point, is calculated as follows:
X '=x
Y '=y
Z '=- z
Point set P after obtaining translationai correctiontranslation, as shown in Figure 7b.
Finally execute step S3, mirror module.Step S3 includes step S31 and step S32:
Step S31 is executed, domestic animal point cloud is cut.To the point set P after translationai correctiontranslationEach point sentenced
It is disconnected, if z value is greater than 0, deleted, cloud P is put after being cutdelete, as shown in Figure 8 a.
Step S32 is executed, domestic animal point cloud is symmetrical.To point cloud P after being cutdeleteEach of point z value take it is negative,
Obtain a cloud PsymThen cloud P is put after being integrated into cuttingdeleteIn, obtain P to the endcomplete, Pcomplete=Psym+Pdelete, such as
Shown in Fig. 8 b.
Domestic animal point cloud data in the present embodiment is not limited thereto by taking pig as an example, other animals such as ox, horse, sheep
Equal animals can be used the domestic animal body surface three-dimensional data acquisition methods provided by the invention based on inexpensive single depth camera quick
Obtain complete point cloud data.
In conclusion the present invention is to provide a kind of domestic animal body surface three-dimensional data based on inexpensive single depth camera is fast
Fast acquisition methods, using unilateral domestic animal body surface point cloud data, points cloud processing technology and computer graphics techniques are acquired complete
Domestic animal body surface point cloud data.Complete family is sought using most side domestic animal body surface point cloud datas the innovation of the invention consists in that proposing
The method for raiseeing point cloud data.The present invention is capable of handling using single depth camera, is remembered using most side domestic animal body surface point cloud datas
Record and restore complete family's carcass footage according to the problem of.
Finally, it should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations;Although
Present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that: it still may be used
To modify the technical solutions described in the foregoing embodiments or equivalent replacement of some of the technical features;
And these are modified or replaceed, technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution spirit and
Range.
Claims (7)
1. a kind of domestic animal body surface three-dimensional data acquisition methods, which comprises the following steps:
A) point cloud data inputs, and the domestic animal body surface three dimensional point cloud of the most sides comprising ground is obtained using depth camera,
And input point cloud data;
B) plane monitoring-network carries out the cloud detection of floor point to the point cloud data inputted in the step a), deletes floor
Point cloud obtains the ground normal vector for being directed toward domestic animal body surface three dimensional point cloud by eliminating normal vector ambiguity;
C) local coordinate system is established, and carries out feature extraction to the half side body surface three dimensional point cloud of domestic animal, obtain domestic animal center of gravity and
Orthogonal three reference axis feature vectors, are denoted as X-axis, Y-axis, Z axis respectively, to establish using domestic animal center of gravity as the office of origin
Portion's coordinate system;
D) point cloud coordinate conversion, calculates rotational translation matrix, and according to rotational translation matrix to domestic animal body surface three dimensional point cloud
Coordinate carry out coordinate conversion, while snapping in global coordinate system;
E) rotational correction is corrected in the step d) and is sat through cloud in conjunction with the ground normal vector obtained in the step b)
The coordinate of domestic animal body surface three dimensional point cloud after mark conversion;
F) root of the tail midpoint detection intersects the point obtained with domestic animal body surface three dimensional point cloud by judging the plane perpendicular to X-axis
Cloud quantity determines the X axis coordinate at root of the tail midpoint;
G) translationai correction translates in the step e) after rotational correction according to the X axis coordinate at the root of the tail midpoint in the step f)
Domestic animal body surface three dimensional point cloud coordinate;
H) mirror image processing, input Z axis is greater than zero domestic animal body surface three dimensional point cloud, using X-axis-Y axis coordinate plane as the plane of symmetry
Mirror image processing is done, to obtain complete domestic animal body surface three dimensional point cloud.
2. domestic animal body surface three-dimensional data acquisition methods according to claim 1, which is characterized in that include in the step a)
The domestic animal body surface three dimensional point cloud of the most sides on ground does not include shooting background.
3. domestic animal body surface three-dimensional data acquisition methods according to claim 1, which is characterized in that step b) the plane inspection
In survey, it is as follows to eliminate normal vector ambiguity method: first passing through the floor point cloud detection and detects two normal vectors, then
Along the direction of described two normal vectors, with the position apart from floor point cloud distance for Nr, do two parallel planes respectively with
Domestic animal body surface three dimensional point cloud intersection, when the plane and domestic animal body surface three for being oriented parallel to the floor along a normal vector
When tieing up the point cloud quantity of point cloud data intersection acquisition greater than preset threshold, it is determined that the normal vector is three-dimensional to be directed toward domestic animal body surface
The ground normal vector of point cloud data.
4. domestic animal body surface three-dimensional data acquisition methods according to claim 1, which is characterized in that the step c) is locally sat
During mark system establishes, feature extraction uses Principal Component Analysis.
5. domestic animal body surface three-dimensional data acquisition methods according to claim 1, which is characterized in that step d) the point cloud is sat
Rotational translation matrix in mark conversion passes through the three reference axis feature vectors obtained in step c) local coordinate system foundation
It is constructed with domestic animal center of gravity.
6. domestic animal body surface three-dimensional data acquisition methods according to claim 1, which is characterized in that the step e) rotation is rectified
Center, by calculating the angle between the ground normal vector and existing coordinate Z axis that have obtained in the step b), construction rotation
Matrix, then according to spin matrix to the seat of the domestic animal body surface three dimensional point cloud in the step d) after cloud coordinate conversion
Mark is rotated.
7. domestic animal body surface three-dimensional data acquisition methods according to claim 1, which is characterized in that the mirror image of the step h)
Processing includes that domestic animal body surface three dimensional point cloud is cut and domestic animal body surface three dimensional point cloud symmetrical treatment.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610395906.7A CN106127727B (en) | 2016-06-07 | 2016-06-07 | A kind of domestic animal body surface three-dimensional data acquisition methods |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610395906.7A CN106127727B (en) | 2016-06-07 | 2016-06-07 | A kind of domestic animal body surface three-dimensional data acquisition methods |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106127727A CN106127727A (en) | 2016-11-16 |
CN106127727B true CN106127727B (en) | 2019-01-04 |
Family
ID=57270046
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610395906.7A Active CN106127727B (en) | 2016-06-07 | 2016-06-07 | A kind of domestic animal body surface three-dimensional data acquisition methods |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106127727B (en) |
Families Citing this family (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109238264B (en) * | 2018-07-06 | 2020-09-01 | 中国农业大学 | Livestock position and posture normalization method and device |
CN111915725B (en) * | 2020-08-21 | 2024-04-12 | 青岛联合创智科技有限公司 | Human body measurement method based on motion reconstruction |
CN115222799B (en) * | 2021-08-12 | 2023-04-11 | 达闼机器人股份有限公司 | Method and device for acquiring image gravity direction, electronic equipment and storage medium |
CN114694186B (en) * | 2022-06-01 | 2022-08-26 | 南京优牧大数据服务有限公司 | Method and device for processing cattle face identification data |
CN116071417B (en) * | 2023-01-31 | 2024-01-12 | 河北农业大学 | Sheep body ruler weight acquisition system and method based on Azure Kinect |
CN115994905B (en) * | 2023-03-22 | 2023-08-01 | 淄博特凯泽自动化工程有限公司 | Positioning detection method for opening of livestock |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104504682A (en) * | 2014-11-25 | 2015-04-08 | 中国农业大学 | Animal shape phenotype automatic measuring method |
CN104517236A (en) * | 2014-11-25 | 2015-04-15 | 中国农业大学 | Automatic animal shape phenotype measuring system |
-
2016
- 2016-06-07 CN CN201610395906.7A patent/CN106127727B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104504682A (en) * | 2014-11-25 | 2015-04-08 | 中国农业大学 | Animal shape phenotype automatic measuring method |
CN104517236A (en) * | 2014-11-25 | 2015-04-15 | 中国农业大学 | Automatic animal shape phenotype measuring system |
Non-Patent Citations (3)
Title |
---|
Using visual image analysis to describe pig growth in terms of size and shape;AB Doeschiwilson 等;《Animal Science An International Journal of Fundamental & Applied Research》;20041231;第79卷(第6期);第415-427页 |
基于深度图像的猪体尺检测系统;李卓 等;《农业机械学报》;20160331;第47卷(第3期);第311-317页 |
基于点云数据的猪体曲面三维重建与应用;刘同海 等;《农业机械学报》;20140630;第45卷(第6期);第291-295页 |
Also Published As
Publication number | Publication date |
---|---|
CN106127727A (en) | 2016-11-16 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106127727B (en) | A kind of domestic animal body surface three-dimensional data acquisition methods | |
CN111199564B (en) | Indoor positioning method and device of intelligent mobile terminal and electronic equipment | |
US9400941B2 (en) | Method of matching image features with reference features | |
EP2249311B1 (en) | Systems and methods for extracting planar features, matching the planar features, and estimating motion from the planar features | |
CN109724603A (en) | A kind of Indoor Robot air navigation aid based on environmental characteristic detection | |
CN109425348B (en) | Method and device for simultaneously positioning and establishing image | |
Zhou et al. | T-loam: truncated least squares lidar-only odometry and mapping in real time | |
WO2015096508A1 (en) | Attitude estimation method and system for on-orbit three-dimensional space object under model constraint | |
WO2016199605A1 (en) | Image processing device, method, and program | |
Jessup et al. | Robust and efficient multirobot 3-d mapping merging with octree-based occupancy grids | |
CN103988226A (en) | Method for estimating camera motion and for determining three-dimensional model of real environment | |
US11315264B2 (en) | Laser sensor-based map generation | |
Urbach et al. | DPDist: Comparing point clouds using deep point cloud distance | |
CN109443207A (en) | A kind of light pen robot in-situ measurement system and method | |
JP2012128744A (en) | Object recognition device, object recognition method, learning device, learning method, program and information processing system | |
CN109872366A (en) | Object dimensional method for detecting position and device based on depth fitting degree assessment network | |
Tomažič et al. | Fusion of visual odometry and inertial navigation system on a smartphone | |
CN110148177A (en) | For determining the method, apparatus of the attitude angle of camera, calculating equipment, computer readable storage medium and acquisition entity | |
Ruiz-Ruiz et al. | A multisensor LBS using SIFT-based 3D models | |
US20230342972A1 (en) | Depth sensor activation for localization based on data from monocular camera | |
CN109238264A (en) | A kind of domestic animal posture method for normalizing and device | |
Barros et al. | Real-time human pose estimation from body-scanned point clouds | |
Huang et al. | Circle detection and fitting using laser range finder for positioning system | |
Pan et al. | A robust RGB-D image-based SLAM system | |
Friedel | Event-based visual-inertial odometry using smart features |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |