CN107607053A - A kind of standing tree tree breast diameter survey method based on machine vision and three-dimensional reconstruction - Google Patents
A kind of standing tree tree breast diameter survey method based on machine vision and three-dimensional reconstruction Download PDFInfo
- Publication number
- CN107607053A CN107607053A CN201710849961.3A CN201710849961A CN107607053A CN 107607053 A CN107607053 A CN 107607053A CN 201710849961 A CN201710849961 A CN 201710849961A CN 107607053 A CN107607053 A CN 107607053A
- Authority
- CN
- China
- Prior art keywords
- tree
- standing tree
- algorithm
- height
- standing
- 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
Abstract
The present invention discloses a kind of standing tree tree breast diameter survey method based on machine vision and three-dimensional reconstruction, by algorithm in detail below and model realization：(1) intelligent mobile end equipment camera calibration model and algorithm, lens distortion correction model and algorithm；(2) two dimensional image coordinate system three-dimensional world coordinate system reconstruction model；The model solves the problems, such as that unit pixel corresponds to real world physical size difference under different distance, so as to carry out standing tree tree breast diameter survey under different distance.(3) standing tree trunk contour detecting and extraction algorithm under natural environment；(4) the minimum external square type detection algorithm of standing tree trunk profile；(5) standing tree trunk height and tree breast diameter survey algorithm；The model solves the tree breast diameter survey of higher in single image.The present invention is solved in environment complexity, labor intensity is big, depends on problem present in the forest inventory investigation work of manual measurement unduly, and reduces labor intensity and human cost, is raised labour efficiency.
Description
Technical field
The present invention relates to Investigation Forestry Resources field, it is based on the measuring method, in particular to one kind of the standing tree diameter of a cross-section of a tree trunk 1.3 meters above the ground
The standing tree tree breast diameter survey method of machine vision and three-dimensional reconstruction.
Background technology
Forest resource data collection is the basis of forest resource monitoring, and it is the forest reserves to improve data acquisition efficiency and quality
Information-based important content.In all kinds of forest inventory investigations, the measurement of the standing tree diameter of a cross-section of a tree trunk 1.3 meters above the ground factor is of crucial importance, and existing main
Investigation method relies on artificial field investigation and carries out hand dipping by tape measure or diameter of a cross-section of a tree trunk 1.3 meters above the ground chi, labor intensity is big, human cost is high,
Efficiency is low.
In tree breast diameter survey, traditional measuring method is time-consuming, effort and efficiency are low, although using close range photogrammetry method
Can solve these problems well, but calibration algorithm is complicated, and the two dimensional image coordinate in picture is recovered to three-dimensional world to sit
Timestamp precision is not high, and error is larger and needs to operate completion at PC ends.Existing standing tree tree breast diameter survey soft and hardware system, model
Majority is based on computer or digital camera, and measurement can not be completed directly on more pervasive smart mobile phone.Such as king's allusion quotation et al.
《Standing tree tree breast diameter survey method based on optical similarity triangulation method》, this method by camera and a laser and need to give
Optics similar triangles method measures；Feng Zhong sections et al.《It is a kind of to utilize the method for pinpointing the angular distance method measurement standing tree diameter of a cross-section of a tree trunk 1.3 meters above the ground》,
This method need using it is preceding from standing tree predetermined distance to be measured dispose measuring instrument, it is necessary to take multiple measurements, operation side
Method is complex, can not realize and measure at an arbitrary position；And existing most mobile terminal standing tree tree breast diameter survey methods are by
The mark passing ratio of principal dimensions calculates the diameter of a cross-section of a tree trunk 1.3 meters above the ground, and needs to carry out post-processing by PC ends.
In recent years smart mobile phone manufacturing industry develop rapidly, increasing sensor, as gravity sensor, GPS module,
GPRS module, camera etc., be all integrated with smart mobile phone so that its function is stronger and stronger, and smart mobile phone popularity rate
Extensively, easy to carry, it will be forest workers and forestry section to design a kind of standing tree diameter of a cross-section of a tree trunk 1.3 meters above the ground method for fast measuring based on smart mobile phone
The personnel of grinding bring great convenience.The present invention is exactly caused in this context.
The content of the invention
In view of this, it is an object of the present invention to provide one kind in order to solve environment is complicated, labor intensity is big, depends on people unduly
Problem present in the forest inventory investigation work of work measurement, and labor intensity and human cost are reduced, raise labour efficiency.Cause
This provides a kind of easy to operate, and cost is low, precision meet forestry survey requirement based on machine vision and three-dimensional reconstruction
Standing tree tree breast diameter survey method.
In order to solve the above-mentioned technical problem, the technical scheme is that：One kind is based on machine vision and three-dimensional reconstruction skill
The standing tree tree breast diameter survey method of art, by algorithm in detail below and model realization：
(1) intelligent mobile end equipment camera calibration model and algorithm, distortion correction model and algorithm；
(2) two dimensional image coordinate system-three-dimensional world coordinate system reconstruction model；The model solves unit picture under different distance
Element corresponds to the problem of real world physical size difference, so as to carry out standing tree tree breast diameter survey under different distance.
(3) standing tree contour detecting and extraction algorithm under natural environment；
(4) the minimum external square type detection algorithm of standing tree profile；
(5) standing tree trunk height and tree breast diameter survey algorithm；The model solves the tree breast diameter survey of higher in single image.
Further, intelligent mobile end equipment camera calibration model and algorithm, distortion correction model and algorithm (1), due to
The computing capability of intelligent sliding moved end is limited and camera lens have different degrees of distortion, including radial distortion and tangential distortion
Deng, radial distortion and tangential distortion correction model are introduced to reduce influence of the distortion to camera calibration precision, it is higher so as to obtain
The inside and outside parameter of camera of precision；In parameter calculation procedure parameter iteration efficiency is improved using L-M algorithms.
Further, two dimensional image coordinate system-three-dimensional world coordinate system reconstruction model (2) solves unit under different distance
Pixel corresponds to the problem of real world physical size difference, so as under different distance by image two dimensional image sit
Mark carries out three-dimensional world coordinate and rebuild so as to realize the measurement of physical size.
Further, standing tree contour detecting and extraction algorithm (3) pass through the target standing tree to be measured to acquisition under natural environment
Picture carries out the extraction of color space component, chooses the tone H components in the deal of Lab color spaces three and hsv color space respectively,
Using entire image in Lab space the arithmetic mean of instantaneous value of each component and each pixel after 3 × 3 operator convolution algorithms
The difference between value is obtained afterwards, obtains vision significance expression of the original image based on frequency tuning, at the same it is balanced to tone H components
Change is handled, and fusion figure is obtained by Fusion Features, and image segmentation is carried out on the basis of figure is merged；Utilize morphological dilations and corruption
Lose combinatorial operation, with reference to standing tree main outline and interference profile girth gap it is big the characteristics of, complete denoising and smoothing processing；
Further, asked first by the estimation of target major axes orientation in the minimum external square type detection algorithm (4) of standing tree profile
Go out standing tree profile minimum enclosed rectangle, the initial rectangular of target is determined according to main shaft, using the geometric center of initial rectangular as rotation
Center rotates to it, so as to find optimal rectangle posture, and translates the postrotational standing tree profile minimum enclosed rectangle of optimization.
After standing tree minimum enclosed rectangle is obtained, regard rectangle height of tree direction as trunk height, obtain the pixel value of the rectangle length of side
Heightpixels。
Further, standing tree trunk height and tree breast diameter survey algorithm (5) are to obtain trunk height Pixel Information
HeightpixelsOn the basis of, with reference to the unit obtained in two dimensional image coordinate system-three-dimensional world coordinate system reconstruction model (2)
Pixel corresponds to real world physical size L3D-realTo calculate trunk height value：Height=Heightpixels*L3D-real.By list
Position pixel physical size information and the pixel-parameters information in trunk height direction, obtain the diameter of a cross-section of a tree trunk 1.3 meters above the ground pixel of 1.3 meters of opening positions of the height of tree
Parameter DBHpixels；Finally, diameter of a cross-section of a tree trunk 1.3 meters above the ground calculation formula is：DBH=DBHpixels*L3D-real。
Another technical problem to be solved by the invention comprises the following steps to provide a kind of standing tree image processing flow：
(1) smart mobile phone shoots the picture of standing tree to be measured, is demarcated using suspension scaling board to camera parameter；
(2) camera calibration process, the inside and outside parameter of camera is obtained；
(3) by inside and outside parameter, three-dimensional reconstruction is carried out, is obtained under the shooting distance, the corresponding reality of unit pixel in picture
The physical size in the world；
(4) standing tree picture processing to be measured, visual saliency map is built by Lab color model and 3*3 operators, and passes through HSV
Or the H components in HSI color model strengthen standing tree trunk profile information；
(5) image dividing processing binaryzation, noise reduction is then carried out by the combination operation for expanding and corroding, removes tiny make an uproar
Sound, fill the hole in trunk profile；
(6) girth of each object in the picture after processing is extracted, the girth according to standing tree profile is most long spy
Point, extract the profile of standing tree trunk；
(7) minimum enclosed rectangle of standing tree trunk profile is extracted, obtains the pixel value in standing tree height of tree direction, and according to (3)
The physical size that the unit pixel of middle acquisition corresponds to real world is multiplied with the pixel value in height of tree direction, can obtain image neutrality ebon
Dry elevation information, and intercept the pixel value in the diameter of a cross-section of a tree trunk 1.3 meters above the ground direction of 1.3 meters of height and positions, the physics of real world corresponding with unit pixel
Size is multiplied, you can calculates diameter of a cross-section of a tree trunk 1.3 meters above the ground value.
The present invention has advantages below：(1) measurement accuracy of the invention is satisfied by what the State Administration of Forestry promulgated for 2014《State
Family's continuous forest inventory technical stipulation》, Article 7：Investigation allowable error, the 4th section：Tree breast diameter survey：The diameter of a cross-section of a tree trunk 1.3 meters above the ground is less than 20cm's
Trees, measurement error are less than 0.3cm；The diameter of a cross-section of a tree trunk 1.3 meters above the ground is more than or equal to 20cm trees, and measurement error is less than 1.5%；The 5th section of height of tree is surveyed
Amount：When the height of tree is less than 10m, measurement error is less than 3%；When the height of tree is more than or equal to 10m, measurement error is less than 5%.(2)
Simple and convenient is operated, only being needed when measuring the diameter of a cross-section of a tree trunk 1.3 meters above the ground can perpendicular to ground photographic subjects standing tree by the video camera of intelligent sliding moved end
The diameter of a cross-section of a tree trunk 1.3 meters above the ground of automatic measurement standing tree, and automatically save measurement data and image information.(3) cost is low, the pervasive rate of equipment is wide.According to
The modern electronic equipments such as solution, tree breast diameter survey equipment such as total powerstation are expensive, and complex operation.(4) it can be widely applied to it
His industry field.It can solve the problem that in environment is complicated, labor intensity is big, depends on the forest inventory investigation work of manual measurement unduly
The problem of existing, and labor intensity and human cost are reduced, raise labour efficiency.Therefore offer is a kind of easy to operate, and cost is low,
Precision meets forestry survey requirement.
Brief description of the drawings
The present invention is further described with example below in conjunction with the accompanying drawings.
Fig. 1 is technical scheme and schematic diagram；
Fig. 2 is the frequency tuning vision significance expression figure in the standing tree image processing process of the present invention；
Fig. 3 is that the H subscales in the standing tree image processing process of the present invention reach figure；
Fig. 4 is the amalgamation and expression figure in the standing tree image processing process of the present invention；
Fig. 5 is the schematic diagram of the input picture of the present invention；
Fig. 6 is the schematic diagram of the binary image of the present invention；
Fig. 7 is the schematic diagram of the main outline extraction of the present invention；
Fig. 8 is the schematic diagram of the trunk extraction of the present invention；
Fig. 9 is the schematic diagram of the initial rectangular of the present invention；
Figure 10 is the schematic diagram of the minimum enclosed rectangle of the optimized rotation of the present invention；
Figure 11 is that the diameter of a cross-section of a tree trunk 1.3 meters above the ground position positioning of the present invention and the diameter of a cross-section of a tree trunk 1.3 meters above the ground calculate schematic diagram.
Embodiment
Below in conjunction with accompanying drawing 1-11, the embodiment of the present invention is described in further detail, so that the technology of the present invention side
Case is more readily understood and grasped.
A kind of standing tree tree breast diameter survey method based on machine vision and three-dimensional reconstruction, by algorithm in detail below and model
Realize：(1) intelligent mobile end equipment camera calibration model and algorithm, distortion correction model and algorithm；(2) two dimensional image coordinate
System-three-dimensional world coordinate system reconstruction model；The model solves unit pixel under different distance and corresponds to real world physical size
The problem of different, so as to carry out standing tree tree breast diameter survey under different distance.(3) standing tree contour detecting and extraction under natural environment
Algorithm；(4) the minimum external square type detection algorithm of standing tree profile；(5) standing tree trunk height and tree breast diameter survey algorithm；The model solves
The tree breast diameter survey of higher in single image.
As shown in figure 1, intelligent mobile end equipment camera calibration model and algorithm, distortion correction model and algorithm (1), due to
The computing capability of intelligent sliding moved end is limited and camera lens have the distortion for not allowing degree, including radial distortion and tangential distortion,
Radial distortion and tangential distortion correction model are introduced to reduce influence of the distortion to camera calibration precision, so as to obtain degree of precision
The inside and outside parameter of camera；In parameter calculation procedure parameter iteration efficiency is improved using L-M algorithms.Two dimensional image coordinate system-
Three-dimensional world coordinate system reconstruction model (2) solves unit pixel under different distance and corresponds to that real world physical size is different to ask
Topic, so as to be rebuild under different distance by carrying out three-dimensional world coordinate to the two dimensional image coordinate in image so as to realize thing
Manage the measurement of size.Standing tree contour detecting and extraction algorithm (3) pass through the target standing tree picture to be measured to acquisition under natural environment
The extraction of color space component is carried out, chooses the tone H components in the deal of Lab color spaces three and hsv color space respectively, is utilized
The entire image arithmetic mean of instantaneous value of each component and each pixel rear institute after 3 × 3 operator convolution algorithms in Lab space
Obtain the difference between value, obtain original image based on frequency tuning vision significance expression, while to tone H component equalization at
Reason, fusion figure is obtained by Fusion Features, and image segmentation is carried out on the basis of figure is merged；Utilize morphological dilations and corrosion group
Close computing, with reference to standing tree main outline and interference profile girth gap it is big the characteristics of, complete denoising and smoothing processing；Standing tree profile is most
Standing tree profile minimum enclosed rectangle is obtained by the estimation of target major axes orientation first in small external square type detection algorithm (4), according to
Main shaft determines the initial rectangular of target, it is rotated using the geometric center of initial rectangular as pivot, so as to find most
Excellent rectangle posture, and translate the postrotational standing tree profile minimum enclosed rectangle of optimization., will after standing tree minimum enclosed rectangle is obtained
Trunk height is regarded in rectangle height of tree direction as, obtains the pixel value Height of the rectangle length of sidepixels.Standing tree trunk height and the diameter of a cross-section of a tree trunk 1.3 meters above the ground are surveyed
Quantity algorithm (5) is to obtain trunk height Pixel Information HeightpixelsOn the basis of, with reference to two dimensional image coordinate system-three-dimensional generation
Unit pixel obtained in boundary's coordinate system reconstruction model (2) corresponds to real world physical size L3D-realTo calculate trunk height
Value：Height=Heightpixels*L3D-real.Believed by the pixel-parameters in unit pixel physical size information and trunk height direction
Breath, obtain the diameter of a cross-section of a tree trunk 1.3 meters above the ground pixel-parameters DBH of 1.3 meters of opening positions of the height of treepixels；Finally, diameter of a cross-section of a tree trunk 1.3 meters above the ground calculation formula is：DBH=DBHpixels*
L3D-real。
A kind of standing tree image processing flow, comprises the following steps：
(1) smart mobile phone shoots the picture of standing tree to be measured, is demarcated using suspension scaling board to camera parameter；
(2) camera calibration process, the inside and outside parameter of camera is obtained；
(3) by inside and outside parameter, three-dimensional reconstruction is carried out, is obtained under the shooting distance, the corresponding reality of unit pixel in picture
The physical size in the world；
(4) standing tree picture processing to be measured, visual saliency map is built by Lab color model and 3*3 operators, and passes through HSV
Or the H components in HSI color model strengthen standing tree trunk profile information；
(5) image dividing processing binaryzation, noise reduction is then carried out by the combination operation for expanding and corroding, removes tiny make an uproar
Sound, fill the hole in trunk profile；
(6) girth of each object in the picture after processing is extracted, the girth according to standing tree profile is most long spy
Point, extract the profile of standing tree trunk；
(7) minimum enclosed rectangle of standing tree trunk profile is extracted, obtains the pixel value in standing tree height of tree direction, and according to (3)
The physical size that the unit pixel of middle acquisition corresponds to real world is multiplied with the pixel value in height of tree direction, can obtain image neutrality ebon
Dry elevation information, and intercept the pixel value in the diameter of a cross-section of a tree trunk 1.3 meters above the ground direction of 1.3 meters of height and positions, the physics of real world corresponding with unit pixel
Size is multiplied, you can calculates diameter of a cross-section of a tree trunk 1.3 meters above the ground value.
In the present embodiment, standing tree tree breast diameter survey method uses the improved camera mark suitable for intelligent sliding moved end camera
Cover half type and distortion correction model；The standing tree profile salient region that machine vision technique is included under natural environment detects, standing tree
Trunk area-of-interest detects and image segmentation；After the image segmentation for completing standing tree profile and trunk part, combining camera demarcation
The inside and outside parameter of camera of acquisition carries out the reconstruction that two-dimensional image information carries out three-dimensional world coordinate, then carries out the measurement of the diameter of a cross-section of a tree trunk 1.3 meters above the ground
Calculate.
Described camera calibration model and distortion correction model：Corrected including pin-hole imaging model (formula 1), camera distortion
Model formation (formula 2, formula 5), the radial distortion pattern function (formula 3) for ignoring higher order term, ignore the tangential abnormal of higher order term
Varying model function (formula 4)；Image ideal coordinates point (xu, yu) due to the distortion of camera lens, its actual coordinate point for (x,
Y), by introducing nonlinear distortion variate δx, δyTo describe existing distortion.Formula (4) is obtained by formula (1), (2), (3)
Distortion correction pattern function, wherein containing k1、k2、p1、p24 kilrrfactors.
Meanwhile according to the relation during camera calibration between image coordinate system and pixel coordinate system, in image coordinate system
The pixel coordinate point of point (x, y) be (u0, v0), due to manufacturing process problem, camera imaging unit is not regular square,
Physical size of each pixel in x-axis, y-axis direction is d in image planesx, dy.Understand that any one pixel is at two in image
Meet following relation in coordinate system：
Convolution (5), linear camera is modeled as into homogeneous coordinates is with matrix form：
The improved camera calibration model and distortion correction model suitable for mobile terminal camera more than, with reference to opening just
The parameter algorithm of friendly standardization, you can calculate the inside and outside parameter of camera.
, can be from uncalibrated image by camera calibration when carrying out the reconstruction of three-dimensional world coordinate system by two dimensional image coordinate system
In, calculate the actual physical distance between two pixels under three-dimensional world coordinate system on image.Rectified with reference to acquisition through distorting
Mobile phone camera parameter after just, its algorithmic procedure are：
The first step：The physical size of each pixel can be released by size sensor：Lpixels=LCCD/pixels；
Second step：Focal length of camera fc physical size is equal to Lfc=fc*Lpixels；
3rd step：On pixel planes, the distance between 2 points are Distance=sqrt ((u1-u2)^2+(v1-v2))*
Lfc；
Therefore, the actual physical size in image in the unit pixel corresponding three-dimensional world is：L3D-real=Distance*
Depth(3)/Lpixels.Through above-mentioned algorithm can unit of account pixel corresponding three-dimensional world coordinates physical size.
As in Figure 2-4, when standing tree profile is extracted under natural environment：
A) extraction of color space component is carried out to collection image, extracts L, a in Lab color model, b3 component, Lab
Three-component extracts：
B) using Lab color spaces as characteristics of image, for each Color Channel L, a, b, calculate each pixel (x,
Y) with the average color difference of entire image and squared, the value of these three passages is then added the significance value as the pixel：
C) using entire image in Lab space the arithmetic mean of instantaneous value of each component and each pixel (x, y) through too high
The difference between value is obtained after this filtering, obtains vision significance expression of the original image shown in b based on frequency tuning；
D) H components in hsv color space are introduced into be fused in target standing tree trunk feature representation, notable figure is contrasted
The limited adaptive histogram equalization adjustment of degree, strengthen the color contrast of image local, capture the standing tree trunk of brown colour system
Detail differences between the background of green system.The equalization H components of standing tree trunk image and frequency tuning visual saliency map two
Individual feature is merged by formula (10), is expressed as：
Wherein Ifusion(x, y) represents the fusion feature of each pixel (x, y), and H (x, y) represents the equal of pixel (x, y)
Weighing apparatusization tone, S (x, y) represent the frequency tuning vision significance value of the pixel (x, y).Modulated by feature, reduce illumination
Influence of the Strength Changes for vision significance, so as to strengthen target standing tree trunk profile.
Fusion figure is subjected to image segmentation, now target image still suffers from that outline portion is imperfect and some residual noises,
Independent pictorial element is partitioned into by opening operation, eliminates small objects, is reached in very thin place's separating objects and smooth larger thing
The effect on body border；And by element adjacent in closed operation connection figure picture, the minuscule hole in filler body, it is adjacent to reach connection
The effect of nearly object and smooth boundary.On this basis, the characteristics of big using the girth gap of standing tree main outline and interference profile,
The less distracter of girth is rejected, retains main outline, completes the output of standing tree profile.
When calculating the standing tree profile external square type of minimum, the minimum external square of standing tree profile is obtained by the estimation of target major axes orientation
Shape, the initial rectangular of target is determined according to main shaft, it is rotated using the geometric center of initial rectangular as pivot, so as to
Optimal rectangle posture is found, and translates the postrotational rectangle of optimization, as shown in Figure 5., will after standing tree minimum enclosed rectangle is obtained
Trunk height is regarded in rectangle height of tree direction as, by the pixel value Height for obtaining the rectangle length of sidepixelsTrunk height calculating is carried out,
Calculation formula is：Height=Heightpixels*L3D-real。
When detecting profile at the standing tree diameter of a cross-section of a tree trunk 1.3 meters above the ground, according to HeightpixelsAnd L3D-realTrunk height is calculated, and the height of tree can be calculated
The diameter of a cross-section of a tree trunk 1.3 meters above the ground pixel-parameters DBH of 1.3 meters of opening positionspixels, with reference to diameter of a cross-section of a tree trunk 1.3 meters above the ground direction Pixel Information DBH at 1.3 meters of the height of treepixelsAnd unit
Pixel physical size information L3D-real, diameter of a cross-section of a tree trunk 1.3 meters above the ground calculation formula is：DBH=DBHpixels*L3D-real。
Diameter of a cross-section of a tree trunk 1.3 meters above the ground DBH Measurement Algorithm is as follows：
Trunk height Height=Heightpixels*L3D-real；
Wherein unit pixel corresponds to real world actual physical size information L3D-realBy being measured after 2.3 section three-dimensional reconstructions.
By trunk height and unit pixel physical size L3D-realThe Pixel Information P of 1.3 meters of height can be calculated1.3m：
P1.3m=1.3*1000/L3D-real；
Intercept diameter of a cross-section of a tree trunk 1.3 meters above the ground direction pixel value DBH at P1.3mpixels, then the diameter of a cross-section of a tree trunk 1.3 meters above the ground be：DBH=DBHpixels*L3D-real。
The present invention has advantages below：(1) measurement accuracy of the invention is satisfied by what the State Administration of Forestry promulgated for 2014《State
Family's continuous forest inventory technical stipulation》, Article 7：Investigation allowable error, the 4th section：Tree breast diameter survey：The diameter of a cross-section of a tree trunk 1.3 meters above the ground is less than 20cm's
Trees, measurement error are less than 0.3cm；The diameter of a cross-section of a tree trunk 1.3 meters above the ground is more than or equal to 20cm trees, and measurement error is less than 1.5%；The 5th section of height of tree is surveyed
Amount：When the height of tree is less than 10m, measurement error is less than 3%；When the height of tree is more than or equal to 10m, measurement error is less than 5%.(2)
Simple and convenient is operated, only being needed when measuring the diameter of a cross-section of a tree trunk 1.3 meters above the ground can perpendicular to ground photographic subjects standing tree by the video camera of intelligent sliding moved end
The diameter of a cross-section of a tree trunk 1.3 meters above the ground of automatic measurement standing tree, and automatically save measurement data and image information.(3) cost is low, the pervasive rate of equipment is wide.According to
The modern electronic equipments such as solution, tree breast diameter survey equipment such as total powerstation are expensive, and complex operation.(4) it can be widely applied to it
His industry field.It can solve the problem that in environment is complicated, labor intensity is big, depends on the forest inventory investigation work of manual measurement unduly
The problem of existing, and labor intensity and human cost are reduced, raise labour efficiency.Therefore offer is a kind of easy to operate, and cost is low,
Precision meets forestry survey requirement.
Certainly, the above is the representative instance of the present invention, and in addition, the present invention can also have other a variety of specific implementations
Mode, all technical schemes formed using equivalent substitution or equivalent transformation, is all fallen within the scope of protection of present invention.
Claims (6)
- A kind of 1. standing tree tree breast diameter survey method based on machine vision and three-dimensional reconstruction, it is characterised in that：This method be by Algorithm and model realization in detail below：(1) intelligent mobile end equipment camera calibration model and algorithm, distortion correction model and algorithm；(2) two dimensional image coordinate system-three-dimensional world coordinate system reconstruction model；The model solves unit pixel pair under different distance The problem of answering real world physical size difference, so as to realize progress standing tree tree breast diameter survey under different distance.(3) standing tree contour detecting and extraction algorithm under natural environment；(4) the minimum external square type detection algorithm of standing tree profile；(5) standing tree trunk height and tree breast diameter survey algorithm；The model solves the tree breast diameter survey of higher in single image.
- 2. a kind of standing tree tree breast diameter survey method based on machine vision and three-dimensional reconstruction according to claim 1, its It is characterised by：Intelligent mobile end equipment camera calibration model and algorithm, distortion correction model and algorithm (1), due to intelligent mobile The computing capability at end is limited and camera lens have the distortion for not allowing degree, including radial distortion and tangential distortion, introduces radially Distort with tangential distortion correction model to reduce influence of the distortion to camera calibration precision, so as to obtain the camera of degree of precision Inside and outside parameter；In parameter calculation procedure parameter iteration efficiency is improved using L-M algorithms.
- 3. a kind of standing tree tree breast diameter survey method based on machine vision and three-dimensional reconstruction according to claim 1, its It is characterised by：It is corresponding that two dimensional image coordinate system-three-dimensional world coordinate system reconstruction model (2) solves unit pixel under different distance Real world physical size difference the problem of, so as under different distance by image two dimensional image coordinate carry out Three-dimensional world coordinate is rebuild so as to realize the measurement of physical size.
- 4. a kind of standing tree tree breast diameter survey method based on machine vision and three-dimensional reconstruction according to claim 1, its It is characterised by：Standing tree contour detecting and extraction algorithm (3) to the target standing tree picture to be measured of acquisition by carrying out under natural environment The extraction of color space component, the tone H components in the deal of Lab color spaces three and hsv color space are chosen respectively, utilize view picture Image arithmetic mean of instantaneous value of each component in Lab space obtains value afterwards with each pixel after 3 × 3 operator convolution algorithms Between difference, obtain vision significance expression of the original image based on frequency tuning, while to tone H component equalization processings, Fusion figure is obtained by Fusion Features, image segmentation is carried out on the basis of figure is merged；Combined using morphological dilations and corrosion Computing, with reference to standing tree main outline and interference profile girth gap it is big the characteristics of, complete denoising and smoothing processing.
- 5. a kind of standing tree tree breast diameter survey method based on machine vision and three-dimensional reconstruction according to claim 1, its It is characterised by：Standing tree wheel is obtained by the estimation of target major axes orientation first in the minimum external square type detection algorithm (4) of standing tree profile Wide minimum enclosed rectangle, the initial rectangular of target is determined according to main shaft, be pivot to it using the geometric center of initial rectangular Rotated, so as to find optimal rectangle posture, and translate the postrotational standing tree profile minimum enclosed rectangle of optimization.It is vertical obtaining After wooden minimum enclosed rectangle, regard rectangle height of tree direction as trunk height, obtain the pixel value Height of the rectangle length of sidepixels。
- 6. a kind of standing tree tree breast diameter survey method based on machine vision and three-dimensional reconstruction according to claim 1, its It is characterised by：Standing tree trunk height and tree breast diameter survey algorithm (5) are to obtain trunk height Pixel Information HeightpixelsBase On plinth, real generation is corresponded to reference to the unit pixel obtained in two dimensional image coordinate system-three-dimensional world coordinate system reconstruction model (2) Boundary physical size L3D-realTo calculate trunk height value：Height=Heightpixels*L3D-real.By unit pixel physical size Information and the pixel-parameters information in trunk height direction, obtain the diameter of a cross-section of a tree trunk 1.3 meters above the ground pixel-parameters DBH of 1.3 meters of opening positions of the height of treepixels；Most Afterwards, diameter of a cross-section of a tree trunk 1.3 meters above the ground calculation formula is：DBH=DBHpixels*L3D-real。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710849961.3A CN107607053B (en) | 2017-09-20 | 2017-09-20 | A kind of standing tree tree breast diameter survey method based on machine vision and three-dimensional reconstruction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710849961.3A CN107607053B (en) | 2017-09-20 | 2017-09-20 | A kind of standing tree tree breast diameter survey method based on machine vision and three-dimensional reconstruction |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107607053A true CN107607053A (en) | 2018-01-19 |
| CN107607053B CN107607053B (en) | 2019-09-20 |
Family
ID=61061345
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710849961.3A Active CN107607053B (en) | 2017-09-20 | 2017-09-20 | A kind of standing tree tree breast diameter survey method based on machine vision and three-dimensional reconstruction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107607053B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109269430A (en) * | 2018-08-12 | 2019-01-25 | 浙江农林大学 | The more plants of standing tree diameter of a cross-section of a tree trunk 1.3 meters above the ground passive measurement methods based on depth extraction model |
| CN110749283A (en) * | 2019-11-29 | 2020-02-04 | 杭州柏晗装饰设计工程有限公司 | Method for measuring house interior full decoration |
| CN111667879A (en) * | 2020-07-17 | 2020-09-15 | 浙江农林大学 | Moso bamboo forest stand structure optimization method based on minimum entropy |
| CN111798444A (en) * | 2020-07-17 | 2020-10-20 | 太原理工大学 | Unmanned workshop steel pipe length measuring method based on image distortion correction color separation processing |
| CN111854616A (en) * | 2020-05-21 | 2020-10-30 | 南京鸿亦沄智能科技有限公司 | Tree breast height diameter vision measurement method and system under assistance of laser |
| CN115294562A (en) * | 2022-07-19 | 2022-11-04 | 广西大学 | Intelligent sensing method for operation environment of plant protection robot |
| WO2024023935A1 (en) * | 2022-07-26 | 2024-02-01 | 日本電信電話株式会社 | Estimation device, estimation system, estimation method, and program |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE3322469A1 (en) * | 1982-06-22 | 1983-12-22 | Steyr-Daimler-Puch AG, 1011 Wien | Device for contactless thickness measurement of rod-shaped objects having an approximately circular cross-section, in particular tree trunks or the like |
| CN102538693A (en) * | 2012-02-14 | 2012-07-04 | 北京林业大学 | Vertical wood diameter detection method based on laser and vision technology |
| CN103697826A (en) * | 2013-12-27 | 2014-04-02 | 东北林业大学 | Stumpage DBH (diameter at breast height) 6-point measurement method on basis of laser distance measurement |
| EP2855107A1 (en) * | 2012-06-19 | 2015-04-08 | Centre De Recherche Industrielle Du Quebec | Method and system for detecting the quality of debarking at the surface of a wooden log |
| CN105066877A (en) * | 2015-07-16 | 2015-11-18 | 北京工业大学 | Tree measuring method based on intelligent terminal lens |
| CN105513078A (en) * | 2015-12-15 | 2016-04-20 | 浙江农林大学 | Standing tree information acquisition method and device based on images |
| CN105571561A (en) * | 2015-12-15 | 2016-05-11 | 浙江农林大学 | Method and device for collecting stumpage information |
-
2017
- 2017-09-20 CN CN201710849961.3A patent/CN107607053B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE3322469A1 (en) * | 1982-06-22 | 1983-12-22 | Steyr-Daimler-Puch AG, 1011 Wien | Device for contactless thickness measurement of rod-shaped objects having an approximately circular cross-section, in particular tree trunks or the like |
| CN102538693A (en) * | 2012-02-14 | 2012-07-04 | 北京林业大学 | Vertical wood diameter detection method based on laser and vision technology |
| EP2855107A1 (en) * | 2012-06-19 | 2015-04-08 | Centre De Recherche Industrielle Du Quebec | Method and system for detecting the quality of debarking at the surface of a wooden log |
| CN103697826A (en) * | 2013-12-27 | 2014-04-02 | 东北林业大学 | Stumpage DBH (diameter at breast height) 6-point measurement method on basis of laser distance measurement |
| CN105066877A (en) * | 2015-07-16 | 2015-11-18 | 北京工业大学 | Tree measuring method based on intelligent terminal lens |
| CN105513078A (en) * | 2015-12-15 | 2016-04-20 | 浙江农林大学 | Standing tree information acquisition method and device based on images |
| CN105571561A (en) * | 2015-12-15 | 2016-05-11 | 浙江农林大学 | Method and device for collecting stumpage information |
Non-Patent Citations (3)
| Title |
|---|
| 侯鑫新等: "《一种基于单CCD 与经纬仪的树木胸径测量方法》", 《计算机应用研究》 * |
| 张正友: "《A flexible new technique for camera calibration》", 《IEEE TRANS PATTERN ANALY MACH INTELL》 * |
| 阮晓晓: "《基于普通数码相机的测树因子识别研究》", 《中国优秀硕士学位论文全文数据库 农业科技辑》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109269430A (en) * | 2018-08-12 | 2019-01-25 | 浙江农林大学 | The more plants of standing tree diameter of a cross-section of a tree trunk 1.3 meters above the ground passive measurement methods based on depth extraction model |
| CN109269430B (en) * | 2018-08-12 | 2020-10-09 | 浙江农林大学 | Multi-standing-tree breast height diameter passive measurement method based on deep extraction model |
| CN110749283A (en) * | 2019-11-29 | 2020-02-04 | 杭州柏晗装饰设计工程有限公司 | Method for measuring house interior full decoration |
| CN111854616A (en) * | 2020-05-21 | 2020-10-30 | 南京鸿亦沄智能科技有限公司 | Tree breast height diameter vision measurement method and system under assistance of laser |
| CN111667879A (en) * | 2020-07-17 | 2020-09-15 | 浙江农林大学 | Moso bamboo forest stand structure optimization method based on minimum entropy |
| CN111798444A (en) * | 2020-07-17 | 2020-10-20 | 太原理工大学 | Unmanned workshop steel pipe length measuring method based on image distortion correction color separation processing |
| CN111667879B (en) * | 2020-07-17 | 2023-04-28 | 浙江农林大学 | Minimum entropy-based phyllostachys pubescens stand structure optimization method |
| CN111798444B (en) * | 2020-07-17 | 2023-06-27 | 太原理工大学 | Unmanned workshop steel pipe length measurement method based on image distortion correction color separation processing |
| CN115294562A (en) * | 2022-07-19 | 2022-11-04 | 广西大学 | Intelligent sensing method for operation environment of plant protection robot |
| WO2024023935A1 (en) * | 2022-07-26 | 2024-02-01 | 日本電信電話株式会社 | Estimation device, estimation system, estimation method, and program |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107607053B (en) | 2019-09-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107607053B (en) | A kind of standing tree tree breast diameter survey method based on machine vision and three-dimensional reconstruction | |
| CN109146948B (en) | Crop growth phenotype parameter quantification and yield correlation analysis method based on vision | |
| CN109285145B (en) | Multi-standing tree height measuring method based on smart phone | |
| CN107767442B (en) | Foot type three-dimensional reconstruction and measurement method based on Kinect and binocular vision | |
| KR102002745B1 (en) | Flight parameter measurement device and flight parameter measurement method | |
| CN107240129A (en) | Object and indoor small scene based on RGB D camera datas recover and modeling method | |
| CN106651942A (en) | Three-dimensional rotation and motion detecting and rotation axis positioning method based on feature points | |
| CN110610505A (en) | Image segmentation method fusing depth and color information | |
| CN109269430A (en) | The more plants of standing tree diameter of a cross-section of a tree trunk 1.3 meters above the ground passive measurement methods based on depth extraction model | |
| Cao et al. | Sparse photometric 3D face reconstruction guided by morphable models | |
| CN110672020A (en) | Stand tree height measuring method based on monocular vision | |
| CN110956661B (en) | Method for calculating dynamic pose of visible light and infrared camera based on bidirectional homography matrix | |
| CN112257676A (en) | Pointer instrument reading method and system and inspection robot | |
| Wu et al. | Passive measurement method of tree diameter at breast height using a smartphone | |
| CN107928675A (en) | A kind of trunk measuring method being combined based on deep learning and red dot laser | |
| CN107869954A (en) | A kind of binocular vision volume weight measuring system and its implementation | |
| CN109448043A (en) | Standing tree height extracting method under plane restriction | |
| CN106056121A (en) | Satellite assembly workpiece fast-identification method based on SIFT image feature matching | |
| CN111862315A (en) | Human body multi-size measuring method and system based on depth camera | |
| CN114332689A (en) | Citrus identification and positioning method, device, equipment and storage medium | |
| Sirmacek et al. | Accuracy assessment of building point clouds automatically generated from iphone images | |
| CN115375842A (en) | Plant three-dimensional reconstruction method, terminal and storage medium | |
| CN112200854A (en) | Leaf vegetable three-dimensional phenotype measurement method based on video image | |
| Xinmei et al. | Passive measurement method of tree height and crown diameter using a smartphone | |
| CN116189006A (en) | Remote sensing image building extraction method supporting three-dimensional data |
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 |