CN103793895A - Method for stitching fruit tree crown layer organ images - Google Patents
Method for stitching fruit tree crown layer organ images Download PDFInfo
- Publication number
- CN103793895A CN103793895A CN201410013218.0A CN201410013218A CN103793895A CN 103793895 A CN103793895 A CN 103793895A CN 201410013218 A CN201410013218 A CN 201410013218A CN 103793895 A CN103793895 A CN 103793895A
- Authority
- CN
- China
- Prior art keywords
- image
- fruit tree
- images
- canopy
- fruit
- 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
Images
Abstract
The invention discloses a method for stitching fruit tree crown layer organ images. The method comprises the following steps that A, series intensity images of crown layer organs of fruit trees in different growing periods are shot through a PMD camera and meanwhile, series color images of the crown layer organs of the fruit trees are obtained through a color camera; B, adjacent images of the intensity images and the color images are registered respectively; C, overlapping zones of the adjacent images of the intensity images and the color images are fused; D, a stitching effect image of the intensity images and a stitching effect image of the color images are obtained respectively through stitching; E, production of the fruit trees is instructed according to the stitching effect image of the intensity images and the stitching effect image of the color images. The method is suitable for image stitching of the crown layer organs in a year growing period of the fruit trees, the robustness, the instantaneity and the stitching accuracy of the algorithm can satisfy the requirement for three-dimensional reconstruction operation of a crown layer, and the method has a great significance in the improvement in an information level of fruit ranch management such as pruning, flower thinning, production predication and picking.
Description
Technical field
The present invention relates to IT application to agriculture field, be specifically related to a planting fruit-trees canopy organic image joining method.
Background technology
Refer to that in 1 year, along with climate change, fruit tree shows the vital movement overall process of certain regularity growth period in fruit tree year (annual grouth cycle).The key stage comprises: rest period, flower thinning phase, maturity stage etc.How to build the canopy three-dimensional configuration with color characteristic of different growth phases in year growth period is the informationalized research emphasis of fruit tree always, and its visual research is significant for the intelligent management that promotes the key links such as beta pruning, flower thinning, survey product, harvesting.
At present, mainly contain three kinds for building the means of fruit tree canopy three-dimensional configuration both at home and abroad: the one, stereovision technique, is subject to the impact of outdoor non-structured light when use; The 2nd, laser scanner, although can overcome largely above-mentioned factor, exists obtaining information speed slower, the drawback of data redundancy; The 3rd, 3D digitizer, but this equipment requirement for environmental conditions harshness to external world, and cannot obtain the colouring information amount of each organ of canopy.PMD (Photonic Mixer Device) video camera is based on ToF(Time of Flight) three-dimensional imaging device of technology, obtain the depth point cloud information of testee with the frames in high speed speed of 40fps, but this device resolution lower (200 pixel × 200 pixel), colour TV camera can obtain the abundant colouring information of measured object, although be subject to illumination effect larger, but be combined with PMD video camera, can complement one another, and being combined with in the multi-source image registration in non-structured light environment of the two obtained good application, for fruit tree canopy three-dimensional reconstruction provides reliable instrument.
Image Mosaics refers to the larger view that the picture construction that includes each other public domain more than two is become to comprise public domain, be widely used in fields such as computer vision, medical science, remote sensing, and have no research report for the canopy organic image in key stage in growth period in fruit tree year.The splicing of fruit tree canopy is the key link of canopy three-dimensional visualization, therefore adopt PMD video camera and build a fruit tree multi-source image acquisition system in conjunction with colour TV camera, carry out growth period in fruit tree year the canopy organic image splicing research of (comprising rest period, flower thinning phase, maturity stage), be intended to realize the canopy 3-dimensional reconstruction of the different growth phases of fruit tree in the future, for the orchard informationization management key links such as beta pruning, flower thinning, survey product, harvesting provide technical support.
Therefore,, for above present Research, the invention provides a kind of canopy organic image joining method that is suitable for fruit tree different growing stages.
Summary of the invention
In order to build the canopy three-dimensional panorama form of apple tree different growing stages, overcome single-sensor the drawback of obtaining image information, the invention provides a planting fruit-trees canopy organic image joining method.
A planting fruit-trees canopy organic image joining method provided by the invention, comprises the steps:
A, with the serial intensity image of PMD video camera picked-up fruit tree different growing stages canopy organ, obtains fruit tree canopy organ series coloured image with colour TV camera simultaneously, and selects respectively the adjacent image of intensity image, coloured image;
B is respectively by the adjacent image registration of intensity image, coloured image;
C merges the adjacent image overlapping region of intensity image, coloured image respectively;
D splices respectively and obtains intensity image splicing effect figure and coloured image splicing effect figure;
E instructs production of fruit trees with intensity image splicing effect figure and coloured image splicing effect figure.
Wherein, described in A, be one or more in rest period, flower thinning phase, maturity stage growth period.
Wherein, described adjacent image be shown in scene have 30%(to contain) two width images of above overlapping region.
Wherein, registration described in B, comprises the steps:
Serial intensity image or serial coloured image Gaussian filter G (λ) are carried out convolution algorithm by B1, build initial Gaussian pyramid, Gauss's group is carried out to 1/2 down-sampled computing, repeat this computing several times, until meet the threshold value of setting, obtain gaussian pyramid sequence image with this, set up metric space;
B2 utilizes the principal curvatures characteristic description characteristic point position distribution situation of DoG operator, obtains the Hessian matrix of the each point of original image;
The non-maximum value of utilizing B3 suppresses to find the local acknowledgement's extreme value in 3 × 3 × 3 three-dimensional neighborhood of each pixel in image, using this as unique point to be selected, adopts bilinear interpolation computing accurately to determine characteristic point position;
B4 utilizes statistics with histogram method, Gaussian function weighted feature sampled point in order to unique point yardstick as variance, the gradient of obtaining each pixel of image distributes and about the statistic histogram of characteristic direction, generate its characteristics of image describe adopt the BBF algorithm of KD_Tree feature structure to realize image to be spliced right unique point is slightly mated;
B5 employing RANSAC algorithm to purifying, is eliminated Mismatching point pair to unique point.
Wherein, adopt described in B5 RANSAC algorithm to unique point to purifying, comprise the steps:
In B5-1 sampling, some set, chooses 4 pairs of unique points pair, at random as the parameter of mapping model matrix H;
In B5-2 traversal, some set, is applied to transformation model H, calculates its error ε, in the time that ε is less than predetermined threshold value, this unique point, to as same place pair, is recorded to its quantity M; Otherwise deleted as non-same place;
B5-3 utilizes same place that quantity is M to again solving transformation matrix H, redefines same place quantity M ', at that time, when M ' >M, returns to previous step, otherwise continues to carry out; Repeat above step, until be recycled to preset times T;
B5-4 chooses M ' value corresponding same place pair set when maximum, obtains transformation matrix H.
Wherein, merge described in C, comprise the steps:
C1 carries out the tower decomposition of Gauss to adjacent image to be spliced, and to k layer G
kcarry out interpolation amplification;
C2 carries out Laplacian pyramid to adjacent image to be spliced, and fusion criterion is determined in layering, and high fdrequency component is selected the fusion method based on 3 × 3 provincial characteristicss, guarantees detailed information, and low frequency component adopts weighted mean fusion treatment;
C3 carries out recursive operation by Laplace pyramid according to level from top to bottom, can obtain the gaussian pyramid corresponding with it, reconstruct canopy original image, and then realize the accurate fusion in the doubling of the image to be spliced region.
Wherein, described in D, two adjacent images after the fusion of C overlapping region are incorporated in the design sketch obtaining in blank image by splicing effect figure.
Wherein, described in E, instructing production of fruit trees is one or several that instructs in the beta pruning of fruit tree, flower thinning, survey product, harvesting etc.
The present invention also provides the described application of planting fruit-trees canopy organic image joining method in production of fruit trees.
Wherein, described fruit tree preferably apple tree.
The present invention is for rebuilding fruit tree year growth canopy organ three-dimensional configuration, take rest period, flower thinning phase, maturity stage fruit tree canopy as research object, the intensity image and the coloured image that obtain for the PMD video camera based on photosynthetic mixing Detection Techniques and colour TV camera are respectively carried out canopy Study of Image Mosaics Technology.Utilize the yardstick invariant features of SIFT algorithm, and accurately determine image mapped model in conjunction with RANSAC algorithm, avoided the impact of the non-structured light in orchard and image scale transform; Based on this, fusion rule is determined in application Laplace pyramid decomposition and restructing algorithm, layering, has realized the canopy Image Mosaics of different growing stages, has effectively overcome traditional blending algorithm reflection detailed information ability, has spliced the shortcomings such as vestige is obvious; Joining method is suitable for the canopy organic image splicing in growth period in fruit tree year, and the robustness of algorithm, real-time and splicing precision all can meet the requirement of canopy three-dimensional reconstruction work, and achievement in research is significant to the level of IT application of the orchard managements such as lifting beta pruning, flower thinning, survey product, harvesting.
Accompanying drawing explanation
Fig. 1 is embodiment 1 canopy organic image splicing process flow diagram.
Fig. 2 is embodiment 1 rest period coloured image and intensity image adjacent image registration and splicing effect figure separately.
Fig. 3 is embodiment 1 flower thinning phase coloured image and intensity image adjacent image registration and splicing effect figure separately.
Fig. 4 is embodiment 1 maturity stage coloured image and intensity image adjacent image registration and splicing effect figure separately.
Wherein, (a) be the adjacent coloured image of canopy, (b) for coloured image splicing effect figure, (c) for the adjacent intensity image of canopy, (d) be intensity image splicing effect figure.
Embodiment
Following examples are used for illustrating the present invention, but are not used for limiting the scope of the invention.
Embodiment 1
For rebuilding apple tree year growth canopy organ three-dimensional configuration, take rest period, flower thinning phase, maturity stage apple tree canopy as research object, the intensity image and the coloured image that obtain for the PMD video camera based on photosynthetic mixing Detection Techniques and colour TV camera are respectively carried out canopy Study of Image Mosaics Technology.Apple tree canopy organic image acquisition system in the present embodiment is made up of PMD video camera and colour TV camera based on time-of-flight, wherein the wavelength of PMD CamCube3.0 video camera active light source transmitting is 870nm, utilize the optical sensor PhotonICs PMD41k-S2 of inner 200 × 200 resolution to obtain distance and the half-tone information of testee, frame speed is 40fps, canonical measure is apart from 0.3m to 7m, can quick obtaining static state and the depth information of mobile object, according to the distance apart from camera lens, imaging color by indigo plant to red, and can effectively remove the impact of the complicated chaff interference such as other trees and sky outside measurement range, color camera spatial resolution is 320 × 240, and largest frames speed is 30fps, colour TV camera (model is: Logitech C270) is screwed in PMD video camera top, multi-source image synchronous acquisition platform by researching and developing voluntarily (Zhou Wei, Feng Juan, Liu Gang, etc. the image registration techniques [J] in apple picking robot. Transactions of the Chinese Society of Agricultural Engineering, 2013,29 (11): 20-26.), synchronously obtain canopy organic image.
The canopy organic image joining method that is suitable for apple tree different growing stages in the present embodiment, is applicable to the canopy form three-dimensional reconstruction of the different growing stages under the non-structured light environment in orchard.
Originally be the flow process of the canopy organic image joining method that is suitable for apple tree different growing stages in example as shown in Figure 1, described in comprise the following steps:
Step S1000, choosing triennial free spindle individual plant apple tree is tested object, utilize multi-source shooting unit to obtain crucial growth phase in its year growth period: the canopy image in rest period, flower thinning phase, maturity stage, due to large (approximately 3 meters of the height of trees) of tree crown, need from top to bottom, take from left to right several parts of images and be used for splicing task, what PMD video camera photographed is intensity image, and what colour TV camera photographed is coloured image.In several parts of images that obtain, selecting captured part has 30%(to contain) two width series adjacent images of above overlapping region.Obtain after multi-source image, for guaranteeing not lose the colouring information of coloured image, build the canopy three-dimensional configuration with chromatic information, the image obtaining is not carried out to any pre-service work.
Step S2000, for accurately extracting the unique point of image, first, the image to be spliced of intensity image and coloured image is carried out to convolution algorithm with Gaussian filter G (λ) respectively, build initial Gaussian pyramid, λ=1.2 now, expand 2 times by λ, Gauss's group are carried out to 1/2 down-sampled computing, repeat this computing several times, until meet the threshold value of setting, obtain gaussian pyramid sequence image with this, set up metric space; For accurately Expressive Features point distribution, utilize the principal curvatures characteristic description characteristic point position distribution situation of DoG operator, the Hessian matrix that obtains the each point of original image is:
Wherein, λ is scale factor; D
xx(λ) be the convolution of x anisotropic filter and image; D
xy(λ) be the convolution of xy anisotropic filter and image; D
yy(λ) be the convolution of y anisotropic filter and image; Utilize non-maximum value to suppress to find the local acknowledgement's extreme value in 3 × 3 × 3 three-dimensional neighborhood of each pixel in image, using this as unique point to be selected, and then adopt bilinear interpolation computing accurately to determine characteristic point position; For improving the accuracy that unique point is described, utilize statistics with histogram method, Gaussian function weighted feature sampled point in order to unique point yardstick as variance, obtains the gradient distribution of each pixel of image and the statistic histogram about characteristic direction, generates its characteristics of image and describes.
Complete after above-mentioned steps, carry out unique point exact matching by treating stitching image, can obtain the mapping relations between image.The essence of Feature Points Matching is to find arest neighbors at higher dimensional space.First utilizing the BBF(Best in first of KD_Tree feature structure) arest neighbors of traversal proper vector completes thick coupling with inferior neighbour's ratio, and threshold value is made as 0.65, and what be less than this value is defined as a little interior point, the unique point pair that the match is successful.
For the unique point pair of repeated matching and mistake coupling, adopt RANSAC algorithm to purify to it, reduce mistake matching rate, key step is as follows: some set in sampling, choose at random 4 pairs of unique points pair, as the parameter of mapping model matrix H; In traversal, some set, is applied to transformation model H, calculates its error ε, in the time that ε is less than predetermined threshold value, this unique point, to as same place pair, is recorded to its quantity M; Otherwise deleted as non-same place; Utilize same place that quantity is M to again solving transformation matrix H, redefine same place quantity M ', in the time of M ' >M, return to previous step, otherwise continue to carry out; Repeat above step, until be recycled to preset times T; Choose M ' value corresponding same place pair set when maximum, obtain transformation matrix H.The adjacent coloured image of canopy of apple tree different growing stages and intensity image registration result are as Fig. 2 (a), (c), and Fig. 3 (a), (c), shown in Fig. 4 (a), (c).
Step S3000, carries out the tower decomposition of Gauss to original image, and its k tomographic image is G
k; For setting up Laplace pyramid, by above-mentioned formula G
kin interpolation amplify, be defined as G
k *, make it and G
k-1measure-alike, top layer G
k *the information comprising is less than G
k *, reason is to be G
k *down-sampled G
k-1, so far built the laplacian pyramid of canopy image.
For Accurate Reconstruction canopy original image, the Laplace pyramid of structure is carried out to recursive operation according to level from top to bottom, can obtain the gaussian pyramid corresponding with it, and then reconstruct canopy original image, complete Image Reconstruction.
Wherein, merge operator and select as follows: fusion operator has determined the quality of image syncretizing effect, significant for fusion mass.The details showing due to the canopy organ in each growth period is not quite similar, and especially, in the time that canopy floral leaf is in great numbers, outstanding fusion operator is even more important for clearly showing image detail.
The details of canopy image is included in laplacian pyramid high fdrequency component, therefore, selects the fusion method based on 3 × 3 provincial characteristicss, guarantees detailed information.Fusion rule is as follows:
First 3 × 3 template window are applied in high frequency imaging component; Secondly calculate respectively its variance yields; Finally utilize variance criterion to complete fusion.Fusion criterion is expressed as:
Wherein, O (i, j) is the image after merging; O
1(i, j), O
2(i, j), for treating fused images, σ is variance.
Adopt weighted mean fusion treatment for low frequency component.Expression formula is:
Wherein, ω
n(i, j) is weights,
Step S4000, is incorporated in two width images in the blank image that one width is new, to complete adjacent image splicing; The canopy organ coloured image of apple tree different growing stages and the splicing result of intensity image are respectively as Fig. 2 (b), (d), and Fig. 3 (b), (d), shown in Fig. 4 (b), (d).
With according to Fig. 2 (b), (d), Fig. 3 (b), (d), the information of Fig. 4 (b), (d), can instruct the operations such as the beta pruning carried out in apple tree production, flower thinning, survey product, harvesting.
The above is only the preferred embodiment of the present invention; it should be pointed out that for those skilled in the art, do not departing under the prerequisite of the technology of the present invention principle; can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.
Claims (10)
1. a planting fruit-trees canopy organic image joining method, comprises the steps:
A, with the serial intensity image of PMD video camera picked-up fruit tree different growing stages canopy organ, obtains fruit tree canopy organ series coloured image with colour TV camera simultaneously, and selects respectively the adjacent image of intensity image, coloured image;
B is respectively by the adjacent image registration of intensity image, coloured image;
C merges the adjacent image overlapping region of intensity image, coloured image respectively;
D splices respectively and obtains intensity image splicing effect figure and coloured image splicing effect figure;
E instructs production of fruit trees with intensity image splicing effect figure and coloured image splicing effect figure.
2. fruit tree canopy organic image joining method according to claim 1, is characterized in that, be one or more in dormant stage of fruit tree, flower thinning phase, maturity stage growth period described in A.
3. fruit tree canopy organic image joining method according to claim 1, is characterized in that, described adjacent image be shown in scene have 30%(to contain) two width images of above overlapping region.
4. fruit tree canopy organic image joining method according to claim 1, is characterized in that, registration described in B, comprises the steps:
Serial intensity image or serial coloured image Gaussian filter G (λ) are carried out convolution algorithm by B1, build initial Gaussian pyramid, Gauss's group is carried out to 1/2 down-sampled computing, repeat this computing several times, until meet the threshold value of setting, obtain gaussian pyramid sequence image with this, set up metric space;
B2 utilizes the principal curvatures characteristic description characteristic point position distribution situation of DoG operator, obtains the Hessian matrix of the each point of original image;
The non-maximum value of utilizing B3 suppresses to find the local acknowledgement's extreme value in 3 × 3 × 3 three-dimensional neighborhood of each pixel in image, using this as unique point to be selected, adopts bilinear interpolation computing accurately to determine characteristic point position;
B4 utilizes statistics with histogram method, Gaussian function weighted feature sampled point in order to unique point yardstick as variance, the gradient of obtaining each pixel of image distributes and about the statistic histogram of characteristic direction, generate its characteristics of image describe adopt the BBF algorithm of KD_Tree feature structure to realize image to be spliced right unique point is slightly mated;
B5 employing RANSAC algorithm to purifying, is eliminated Mismatching point pair to unique point.
5. fruit tree canopy organic image joining method according to claim 4, is characterized in that, adopt described in B5 RANSAC algorithm to unique point to purifying, comprise the steps:
In B5-1 sampling, some set, chooses 4 pairs of unique points pair, at random as the parameter of mapping model matrix H;
In B5-2 traversal, some set, is applied to transformation model H, calculates its error ε, in the time that ε is less than predetermined threshold value, this unique point, to as same place pair, is recorded to its quantity M; Otherwise deleted as non-same place;
B5-3 utilizes same place that quantity is M to again solving transformation matrix H, redefines same place quantity M ', at that time, when M ' >M, returns to previous step, otherwise continues to carry out; Repeat above step, until be recycled to preset times T;
B5-4 chooses M ' value corresponding same place pair set when maximum, obtains transformation matrix H.
6. fruit tree canopy organic image joining method according to claim 1, is characterized in that, merges described in C, comprises the steps:
C1 carries out the tower decomposition of Gauss to adjacent image to be spliced, and to k layer G
kcarry out interpolation amplification;
C2 carries out Laplacian pyramid to adjacent image to be spliced, and fusion criterion is determined in layering, and high fdrequency component is selected the fusion method based on 3 × 3 provincial characteristicss, guarantees detailed information, and low frequency component adopts weighted mean fusion treatment;
C3 carries out recursive operation by Laplace pyramid according to level from top to bottom, can obtain the gaussian pyramid corresponding with it, reconstruct canopy original image, and then realize the accurate fusion in the doubling of the image to be spliced region.
7. fruit tree canopy organic image joining method according to claim 1, is characterized in that, splicing effect figure is that two adjacent images after C overlapping region is merged are incorporated in the design sketch obtaining in blank image described in D.
8. fruit tree canopy organic image joining method according to claim 1, is characterized in that, instructing production of fruit trees described in E is one or several that instructs in the beta pruning of fruit tree, flower thinning, survey product, harvesting etc.
9. the application of the fruit tree canopy organic image joining method described in claim 1-8 any one in production of fruit trees.
10. application according to claim 9, is characterized in that, described fruit tree is apple tree.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410013218.0A CN103793895A (en) | 2014-01-10 | 2014-01-10 | Method for stitching fruit tree crown layer organ images |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410013218.0A CN103793895A (en) | 2014-01-10 | 2014-01-10 | Method for stitching fruit tree crown layer organ images |
Publications (1)
Publication Number | Publication Date |
---|---|
CN103793895A true CN103793895A (en) | 2014-05-14 |
Family
ID=50669520
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410013218.0A Pending CN103793895A (en) | 2014-01-10 | 2014-01-10 | Method for stitching fruit tree crown layer organ images |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103793895A (en) |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104270615A (en) * | 2014-10-16 | 2015-01-07 | 富盛科技股份有限公司 | Large-view-field monitoring device, control method and image splicing method used for image splicing |
CN104915982A (en) * | 2015-05-15 | 2015-09-16 | 中国农业大学 | Canopy layer illumination distribution prediction model construction method and illumination distribution detection method |
CN106204507A (en) * | 2015-05-28 | 2016-12-07 | 长沙维纳斯克信息技术有限公司 | A kind of unmanned plane image split-joint method |
CN106709898A (en) * | 2017-03-13 | 2017-05-24 | 微鲸科技有限公司 | Image fusing method and device |
CN107255446A (en) * | 2017-08-01 | 2017-10-17 | 南京农业大学 | A kind of Cold region apple fruit tree canopy three-dimensional map constructing system and method |
CN107358610A (en) * | 2017-06-08 | 2017-11-17 | 北京农业信息技术研究中心 | Crop high flux Phenotypic examination method and system |
CN107966944A (en) * | 2017-11-30 | 2018-04-27 | 贵州财经大学 | Smart greenhouse zone control system and subregion picking method |
CN109816680A (en) * | 2018-12-19 | 2019-05-28 | 黑龙江八一农垦大学 | A kind of high-throughput calculation method of crops plant height |
CN111539473A (en) * | 2020-04-23 | 2020-08-14 | 中国农业大学 | Orchard inter-row leading line extraction method based on 3D Lidar |
CN112102171A (en) * | 2020-09-18 | 2020-12-18 | 贝壳技术有限公司 | Image processing method, image processing device, computer-readable storage medium and electronic equipment |
CN112547576A (en) * | 2020-12-21 | 2021-03-26 | 山东大舜医药物流有限公司 | Vaccine sorting line ex-warehouse rechecking system |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102567979A (en) * | 2012-01-20 | 2012-07-11 | 南京航空航天大学 | Vehicle-mounted infrared night vision system and multi-source images fusing method thereof |
US8499293B1 (en) * | 2005-09-28 | 2013-07-30 | Oracle America, Inc. | Symbolic renaming optimization of a trace |
-
2014
- 2014-01-10 CN CN201410013218.0A patent/CN103793895A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8499293B1 (en) * | 2005-09-28 | 2013-07-30 | Oracle America, Inc. | Symbolic renaming optimization of a trace |
CN102567979A (en) * | 2012-01-20 | 2012-07-11 | 南京航空航天大学 | Vehicle-mounted infrared night vision system and multi-source images fusing method thereof |
Non-Patent Citations (3)
Title |
---|
宋延爽: ""基于网络摄像机视频流的全景图拼接技术研究"", 《中国优秀硕士学位论文全文数据库 信息科技辑》, 15 March 2013 (2013-03-15) * |
张瑜: ""微根窗根系的图像处理方法研究"", 《中国博士学位论文全文数据库 信息科技辑》, 15 November 2012 (2012-11-15) * |
潘子昂: ""基于SIFT算法的图像匹配研究"", 《中国优秀硕士学位论文全文数据库 信息科技辑》, 15 March 2013 (2013-03-15) * |
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104270615B (en) * | 2014-10-16 | 2018-11-13 | 富盛科技股份有限公司 | A kind of big visual field monitoring device, control method and image split-joint method for image mosaic |
CN104270615A (en) * | 2014-10-16 | 2015-01-07 | 富盛科技股份有限公司 | Large-view-field monitoring device, control method and image splicing method used for image splicing |
CN104915982A (en) * | 2015-05-15 | 2015-09-16 | 中国农业大学 | Canopy layer illumination distribution prediction model construction method and illumination distribution detection method |
CN106204507A (en) * | 2015-05-28 | 2016-12-07 | 长沙维纳斯克信息技术有限公司 | A kind of unmanned plane image split-joint method |
CN106709898A (en) * | 2017-03-13 | 2017-05-24 | 微鲸科技有限公司 | Image fusing method and device |
CN107358610A (en) * | 2017-06-08 | 2017-11-17 | 北京农业信息技术研究中心 | Crop high flux Phenotypic examination method and system |
CN107255446B (en) * | 2017-08-01 | 2020-01-07 | 南京农业大学 | Dwarfing close-planting fruit tree canopy three-dimensional map construction system and method |
CN107255446A (en) * | 2017-08-01 | 2017-10-17 | 南京农业大学 | A kind of Cold region apple fruit tree canopy three-dimensional map constructing system and method |
CN107966944A (en) * | 2017-11-30 | 2018-04-27 | 贵州财经大学 | Smart greenhouse zone control system and subregion picking method |
CN109816680A (en) * | 2018-12-19 | 2019-05-28 | 黑龙江八一农垦大学 | A kind of high-throughput calculation method of crops plant height |
CN111539473A (en) * | 2020-04-23 | 2020-08-14 | 中国农业大学 | Orchard inter-row leading line extraction method based on 3D Lidar |
CN111539473B (en) * | 2020-04-23 | 2023-04-28 | 中国农业大学 | 3D Lidar-based orchard inter-row navigation line extraction method |
CN112102171A (en) * | 2020-09-18 | 2020-12-18 | 贝壳技术有限公司 | Image processing method, image processing device, computer-readable storage medium and electronic equipment |
CN112102171B (en) * | 2020-09-18 | 2021-10-08 | 贝壳找房(北京)科技有限公司 | Image processing method, image processing device, computer-readable storage medium and electronic equipment |
CN112547576A (en) * | 2020-12-21 | 2021-03-26 | 山东大舜医药物流有限公司 | Vaccine sorting line ex-warehouse rechecking system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103793895A (en) | Method for stitching fruit tree crown layer organ images | |
Morgenroth et al. | Assessment of tree structure using a 3D image analysis technique—A proof of concept | |
Kazmi et al. | Indoor and outdoor depth imaging of leaves with time-of-flight and stereo vision sensors: Analysis and comparison | |
CN104156972B (en) | Perspective imaging method based on laser scanning distance measuring instrument and multiple cameras | |
Nielsen et al. | Vision-based 3D peach tree reconstruction for automated blossom thinning | |
CN104330074B (en) | Intelligent surveying and mapping platform and realizing method thereof | |
US20100322477A1 (en) | Device and method for detecting a plant | |
Rosell-Polo et al. | Kinect v2 sensor-based mobile terrestrial laser scanner for agricultural outdoor applications | |
CN106165387A (en) | Light field processing method | |
CN109816680A (en) | A kind of high-throughput calculation method of crops plant height | |
CN109816779A (en) | A method of artificial forest forest model, which is rebuild, using smart phone obtains single wooden parameter | |
CN112200854B (en) | Leaf vegetable three-dimensional phenotype measuring method based on video image | |
Peng et al. | Binocular-vision-based structure from motion for 3-D reconstruction of plants | |
CN115861546A (en) | Crop geometric perception and three-dimensional phenotype reconstruction method based on nerve body rendering | |
Clamens et al. | Real-time Multispectral Image Processing and Registration on 3D Point Cloud for Vineyard Analysis. | |
Potena et al. | Suckers emission detection and volume estimation for the precision farming of hazelnut orchards | |
CN116883480A (en) | Corn plant height detection method based on binocular image and ground-based radar fusion point cloud | |
CN105844700A (en) | System for acquiring three-dimensional point clouds in outdoor scene | |
Bahman | Height measurement of basil crops for smart irrigation applications in greenhouses using commercial sensors | |
CN105955079A (en) | Agricultural crop fine seed selection system based on internet of things | |
Drofova et al. | Use of scanning devices for object 3D reconstruction by photogrammetry and visualization in virtual reality | |
Kossieris et al. | Developing a low-cost system for 3d data acquisition | |
Hu et al. | High-fidelity 3D reconstruction of plants using Neural Radiance Fields | |
Zhao et al. | MHPCG: Multi-modal Hyperspectral Point Cloud Generation Based on Single RGB Image | |
Bayat et al. | The Architecture of a stereo image based system to measure tree geometric parameters |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20140514 |