CN104050670A - Complicated background blade image segmentation method combining simple interaction and mark watershed - Google Patents
Complicated background blade image segmentation method combining simple interaction and mark watershed Download PDFInfo
- Publication number
- CN104050670A CN104050670A CN201410285788.5A CN201410285788A CN104050670A CN 104050670 A CN104050670 A CN 104050670A CN 201410285788 A CN201410285788 A CN 201410285788A CN 104050670 A CN104050670 A CN 104050670A
- Authority
- CN
- China
- Prior art keywords
- image
- leaf
- point
- list
- mark
- 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 invention provides a complicated background blade image segmentation method combining simple interaction and a mark watershed. The method comprises the following steps that blade edge points and outer blade points are calibrated on a complicated background blade image in an odd-even alternating mode, the blade edge points and the outer blade points are numbered in sequence, and the points are stored in a point list L0; the calibrated points are processed to obtain a mark image; the complicated background blade image is transformed into a gray level image and an L*a*b* image; the mark image is adopted as a parameter, and mark watershed segmentation is carried out on the grey level image, an a* component image and a b* component image respectively; the final segmentation result is obtained comprehensively through voting. The simple user interaction mode is adopted, the mark is obtained, then mark watershed segmentation is carried out in a Lab space component image and the gray level image respectively, the final segmentation result is obtained comprehensively, and the segmentation accuracy is high.
Description
Technical field
The present invention relates to image processing field, more specifically, relate to a kind of combination simply alternately and the complex background leaf image dividing method of mark watershed divide.
Background technology
Over the past thousands of years plant closely bound up with the mankind, be closely connected.Along with the progress of human civilization, plant has but been suffered more and more serious destruction.The research of plant automatic recognition technology and application, can help general public to strengthen the cognition to plant, cultivates the awareness of the obligations of citizens of treasuring plant; Can promote again the classification of plant and build storehouse, to help reasonable utilization and the introducing and planting of plant, thereby having great importance.
Due to the flower of plant, really, there is complicated solid geometry feature in stem, branch, the robotization identification of blade is relatively effectively simple.Blade is taken off, then gathered simple background image, then identify, existing research widely.Yet priming leaf can cause damage to plant.Directly taking the blade on limb and identify, is desirable way.But this also takes other blades, limb, soil etc. to come in again unavoidably, caused the difficulty that in image, target blade and background are accurately cut apart, had a strong impact on the accuracy of identification.Existing dividing method is difficult to automatically obtain marking image, lacks more deep theoretical research and experimental exploring, cuts apart accuracy rate also not ideal enough.
Summary of the invention
The object of the present invention is to provide the complex background leaf image dividing method of the simple mutual and mark watershed divide of a kind of combination, realize complex background leaf image is cut apart accurately, and can retain blade detail section.
In order to achieve the above object, technical scheme of the present invention is as follows:
Complex background leaf image dividing method in conjunction with alternately simple and mark watershed divide, comprises the following steps:
S11: odd even is respectively demarcated leaf marginal point and leaf exterior point on original complex background leaf image, and leaf marginal point and leaf exterior point number consecutively, be stored in a list L0;
S12: process leaf marginal point and the leaf exterior point demarcated and obtain marking image, process is as follows:
S121: process leaf marginal point and the leaf exterior point demarcated and obtain prospect marking image;
S122: process leaf marginal point, leaf exterior point and the prospect marking image demarcated and obtain context marker image;
S123: overall treatment prospect marking image and context marker image obtain marking image;
S13: original complex background leaf image is converted to gray level image, L*a*b* image;
S14: take marking image as parameter, to gray level image, a* component image, b* component image, carry out mark watershed segmentation respectively;
S15: comprehensively obtain final segmentation result in ballot mode.
In the present invention, user calibrates corresponding leaf marginal point and leaf exterior point on original complex background leaf image, then the point calibrating, give computer system processor, by this simple man-machine interaction mode, obtain and then carry out mark watershed segmentation after marking image and obtain last segmentation result.The inventive method is exclusively used in the processing to complex background leaf image, with strong points, and treatment effect is good, and segmentation result is comparatively accurate.
In described step S121, process demarcating leaf marginal point and leaf exterior point, to obtain the step of prospect marking image as follows:
S21: the leaf marginal point in tie point list L0 forms polygon D1 successively, and the numbering of the leaf marginal point in L0 is stored in a list L1;
S22: whether the leaf exterior point in judging point list L0 in polygon D1, and is handled as follows the leaf exterior point p in polygon D1:
S221: the previous leaf marginal point i and the rear leaf marginal point j that find leaf exterior point p in a list L0;
S222: find through a p, and with i, straight line N that j place straight line M is vertical;
S223: the intersection point k of mark straight line M and N, calculation level i and the distance γ that puts j, and get δ=γ * ε, wherein ε is scale parameter, 0< ε <5;
S224: ask for and take a p as starting point, the directions of rays along some k to p, the some r that the δ distance of advancing arrives;
S225: some r is stored in and is inserted into a list L2, and some i is charged to insert at the Position Number of L1 and follow list of locations L3;
S23: judgement is inserted into whether non-NULL of a list L2; If so, turn to step S24; Otherwise, turn to step S25;
S24: successively the point being inserted in a list L2 is inserted in a list L1, then perform step S25, wherein, f the point being inserted in a list L2 is inserted into the z+f position in a list L1, and z is for inserting the value of following f element in list of locations L3;
S25: ask all bianry image U of the some polygon enclosing region that connection forms successively of a list L1, U is prospect marking image, and wherein polygon and inner point thereof are labeled as " 1 ", and all the other are labeled as " 0 ".
In the present invention, after step S24 shows that f point in L2 should be inserted into and be positioned at the leaf marginal point of z position in the L1 not carrying out before this update.
In the present invention, the point that plant leaf blade is carried out to mark is less, only needs to mark out simply on picture with mouse, can obtain and cut apart preferably accuracy rate.
It is as follows that leaf marginal point, leaf exterior point and the prospect marking image that in described step S122, processing is demarcated obtains context marker image step:
S31: ask all bianry image W1 of the some polygon enclosing region that connection forms successively of a list L0;
S32: ask
, to the morphologic erosion operation of image W2 doing mathematics or in W2 the eight point zero setting that are communicated with in neighborhoods of the every bit in some list L1, obtain image W3;
S33: the four limit frames of image W3 are all set to " 1 ", obtain context marker image W4.
In described step S123 overall treatment prospect marking image and context marker image obtain the method for marking image be by image U and W4 phase or.
In described step S13, adopt following conversion method to ask the gray level image of complex background leaf image:
Gray=0.299×Red+0.587×Green+0.114×Blue。
Described step S15 detailed process is as follows:
S41: the marking image of take carries out mark watershed segmentation to gray level image, a* component image, b* component image respectively as parameter, obtains respectively image J1, J2, J3;
S42: obtain the bianry image O1 of original complex background leaf image in ballot mode in conjunction with J1, J2, J3;
S43: bianry image O1 is first done to closed operation, after do opening operation, obtain final bianry image O2.
The detailed process of described step S42 is as follows:
S51: a newly-built random and identical bianry image O11 of original complex background leaf image size;
S52: any point t to image O11, if multipair point of answering in image J1, J2, J3 respectively, there are both or belong to above the foreground area of correspondence image, determine that some t belongs to the foreground area of image O11, put 1, otherwise this point belongs to the background area of O11, is set to 0, O11 resets and obtains new bianry image O1 the most at last.
Compared with prior art, the invention has the beneficial effects as follows: adopt manually mark limb edge point and leaf exterior point on original complex background image, and give these points this simple user interactions mode that computing machine is processed, obtain marking image, then in Lab spatial component image and gray level image, carry out respectively mark watershed segmentation, comprehensively obtain final segmentation result, cut apart accuracy rate high.
Accompanying drawing explanation
Fig. 1 is the process flow diagram of the inventive method;
Fig. 2 is that the inventive method is carried out former figure, signature and the segmentation result in processing procedure to Chinese yew blade;
Fig. 3-4 have gathered former figure and the segmentation result of the complex background leaf image of 20 Plants for the present invention.
Embodiment
Below in conjunction with drawings and Examples, the present invention is described further.
As shown in Figure 1, the present invention is in conjunction with the complex background leaf image dividing method of alternately simple and mark watershed divide, specific as follows:
(1) open original complex background leaf image IO, allow user's odd even respectively demarcate leaf marginal point and leaf exterior point.System successively from 1 open numbering, is stored in these points in a some list L0, and wherein the mode of numbering is arbitrarily, can, from any numeral, understand, from 1 open numbering for convenient herein.At this, the principle of user's reconnaissance is:
The 1st for demarcating phyllopodium point, and the 2nd outside of dropping on blade makes line segment between these 2 try one's best limb edge but this line segment remove the part of end points and do not intersect with limb edge as far as possible.The 3rd is selected in again on limb edge, be also make as far as possible the 2nd and the 3rd between the line segment part that limb edge but this line segment remove end points of trying one's best intersect with limb edge.The 4th outside of continuing to be selected in blade, so analogizes; Last point is demarcated again the 1st position.
System connects 2 adjacent points successively, forms the polygon of a sealing.As shown in Fig. 2 (a).Supplementary notes: connection blade and limb, normal cylindrical part is called petiole; The intersection of petiole and blade wheel profile is called phyllopodium point.
(2) in a list L0, be numbered the point of odd number, leaf marginal point, calls over, and successively from 1 open numbering, is stored in foreground point list L1.These points connect successively, form a polygon D1.
(3) each in checkpoint list L0 is numbered the some p of even number, i.e. leaf exterior point judges that it is whether in polygon D1.If proceed as follows:
(3.1) in L0, find previous leaf marginal point i and a rear leaf marginal point j of p.
(3.2) obtain through a p, with i, straight line N that j place straight line M is vertical.
(3.3) ask the intersection point k of straight line M and N.
(3.4) ask an i and the distance γ that puts j.
(3.5) get δ=γ * ε, wherein scale parameter ε span is: 0< ε <5, ε=0.8 o'clock optimum.
(3.6) ask for and take a p as starting point, the directions of rays along some k to p, the some r that the δ distance of advancing arrives.
(3.7) r is stored in and is inserted into a list L2, and some i is charged to insert at the Position Number of L1 and follow list of locations L3.
Does is (4) judgement inserted into whether non-NULL of a list L2? if so, turn to step (5); Otherwise, turn to step (6).
(5) successively the point being inserted in a list L2 is inserted in a list L1, then perform step S25, wherein, f the point being inserted in a list L2 is inserted into the z+f position in a list L1, and z is for inserting the value of following f element in list of locations L3.After f point in L2 being inserted into and being positioned at the leaf marginal point of z position in the L1 not carrying out before this update.
(6) ask the bianry image U in the foreground point list L1 region that all the some polygon that connection forms successively surrounds.Polygon and inner point thereof are labeled as " 1 ", and all the other are labeled as " 0 ".U is exactly will be for the prospect mark of mark watershed segmentation.As shown in Fig. 2 (b).
(7) ask the list L0 region that all the some polygon that connection forms successively surrounds, obtain bianry image W1.
(8) point overlapping with foreground area in order to remove background area, asks
.(1)
(9) to the morphologic erosion operation of W2 doing mathematics or in W2 the eight point zero setting that are communicated with in neighborhoods of the every bit in some list L1, obtain W3.The object of this step is to avoid background to be communicated with prospect.
(10) the four limit frames of W3 are all set to " 1 ", obtain W4.Object is that background area is communicated with four limit frames, avoids as much as possible in segmentation result also having except target blade and background the appearance in the 3rd region.W4 is exactly the context marker for mark watershed segmentation.As shown in Fig. 2 (c).
(11) image U and W4 phase or, obtain marking image V.
(12) ask the gray level image I1 of original complex background leaf image IO.In this method, what I0 adopted is RGB color model.Therefore change as follows: Gray=0.299 * Red+0.587 * Green+0.114 * Blue.(2)
(13) ask the L*a*b* image of original complex background leaf image IO, note a* component image is wherein I2, and b* component image is I3.
(14) take V as parameter, respectively I1, I2, I3 are carried out to mark watershed segmentation, obtain image J1, J2, J3.
(15), in ballot mode, in conjunction with J1, J2, J3, obtain bianry image O1.Be specially: a first newly-built bianry image O11 identical with complex background leaf image IO size, again the every bit of image O11 is processed, method is as follows: any point t to image O11, if corresponding point in image J1, J2, J3 respectively, has both or belong to above the foreground area of correspondence image, determine that some t belongs to the foreground area of image O11, put 1, otherwise this t belongs to the background area of O11, is set to 0, O11 resets and obtains new bianry image O1 the most at last.First to build at random a two-value O11 image herein, the image size of this image O11 is identical with complex background leaf image size, then by this image O11 and image J1, J2, J3 sentences ratio, as judged, whether point on the O11 position corresponding point in image J1 lay respectively at the foreground area of image J1, judge whether point on the O11 position corresponding point in image J2 lay respectively at the foreground area of image J2, judge whether point on the O11 position corresponding point in image J3 lay respectively at the foreground area of image J3, if the point on O11 is at image J1, J2, three position corresponding point in J3 have two or more to belong to the foreground area of correspondence image, determine that this point on O11 belongs to the foreground area of O11, this point is put to 1, otherwise this t belongs to the background area of O11, set to 0, O11 resets and obtains new bianry image O1 the most at last.
(16) O1 is first done to closed operation, after do opening operation, obtain final bianry image O2, as shown in Fig. 2 (d).Its objective is the tiny cavity that should not have in the leaf area of filling a vacancy, smoothing blade border, eliminates the tiny region outside blade.
In order to verify the validity of this dividing method, gathered the complex background leaf image of 20 Plants.10 every kind, amount to 200.
Take and do not affect follow-up shape, color and texture feature extraction as aim, formulate image and cut apart criterion of acceptability: 1. cut apart rear blade shape consistent with former figure blade; 2. limb edge sawtooth form is consistent; 3. allow to have mistake among a small circle to divide near limb edge, but a wrong minute area is significantly less than 5% of leaf area.
Finally, in 200 images, the segmentation result of 197 is qualified, and qualification rate is 98.5%.As shown in Figure 3-4, from each kind, choose randomly an image, 20 complex background leaf images that obtain.The right side of every former figure is the bianry image after it is cut apart.
Research and development of the present invention obtain the support of state natural sciences fund problem (61301294) and Guangdong Province's science of education " 12 " education planning infotech research special (13JXN021).
Obviously, the above embodiment of the present invention is only for example of the present invention is clearly described, and is not the restriction to embodiments of the present invention.For those of ordinary skill in the field, can also make other changes in different forms on the basis of the above description.Here exhaustive without also giving all embodiments.All any modifications of doing within the spirit and principles in the present invention, be equal to and replace and improvement etc., within all should being included in the protection domain of the claims in the present invention.
Claims (7)
1. in conjunction with the complex background leaf image dividing method of alternately simple and mark watershed divide, it is characterized in that, comprise the following steps:
S11: odd even is respectively demarcated leaf marginal point and leaf exterior point on original complex background leaf image, and leaf marginal point and leaf exterior point number consecutively, be stored in a list L0;
S12: process leaf marginal point and the leaf exterior point demarcated and obtain marking image, process is as follows:
S121: obtain prospect marking image after the leaf marginal point that processing is demarcated and leaf exterior point;
S122: process leaf marginal point, leaf exterior point and the prospect marking image demarcated and obtain context marker image;
S123: overall treatment prospect marking image and context marker image obtain marking image;
S13: original complex background leaf image is converted to gray level image, L*a*b* image;
S14: take marking image as parameter, to gray level image, a* component image, b* component image, carry out mark watershed segmentation respectively;
S15: comprehensively obtain final segmentation result in ballot mode.
2. the complex background leaf image dividing method of the simple mutual and mark watershed divide of combination according to claim 1, is characterized in that, in described step S121, processing demarcation leaf marginal point and leaf exterior point, to obtain the step of prospect marking image as follows:
S21: the leaf marginal point in tie point list L0 forms polygon D1 successively, and the leaf marginal point in L0 is stored in a list L1;
S22: whether the leaf exterior point in judging point list L0 in polygon D1, and is handled as follows the leaf exterior point p in polygon D1:
S221: the previous leaf marginal point i and the rear leaf marginal point j that find leaf exterior point p in a list L0;
S222: find through a p, and with i, straight line N that j place straight line M is vertical;
S223: the intersection point k of mark straight line M and N, calculation level i and the distance γ that puts j, and get δ=γ * ε, wherein ε is scale parameter, 0< ε <5;
S224: ask for and take a p as starting point, the directions of rays along some k to p, the some r that the δ distance of advancing arrives;
S225: some r is stored in and is inserted into a list L2, and some i is charged to insert at the Position Number of L1 and follow list of locations L3;
S23: judgement is inserted into whether non-NULL of a list L2; If so, turn to step S24; Otherwise, turn to step S25;
S24: successively the point being inserted in a list L2 is inserted in a list L1, then perform step S25, wherein, f the point being inserted in a list L2 is inserted into the z+f position in a list L1, and z is for inserting the value of following f element in list of locations L3;
S25: ask all bianry image U of the some polygon enclosing region that connection forms successively of a list L1, U is prospect marking image, and wherein polygon and inner point thereof are labeled as " 1 ", and all the other are labeled as " 0 ".
3. the complex background leaf image dividing method of the simple mutual and mark watershed divide of combination according to claim 2, it is characterized in that, it is as follows that leaf marginal point, leaf exterior point and the prospect marking image that in described step S122, processing is demarcated obtains context marker image step:
S31: ask all bianry image W1 of the some polygon enclosing region that connection forms successively of a list L0;
S32: ask
, to the morphologic erosion operation of image W2 doing mathematics or in W2 the eight point zero setting that are communicated with in neighborhoods of the every bit in some list L1, obtain image W3;
S33: the four limit frames of image W3 are all set to " 1 ", obtain context marker image W4.
4. the complex background leaf image dividing method of the simple mutual and mark watershed divide of combination according to claim 3, it is characterized in that, in described step S123 overall treatment prospect marking image and context marker image obtain the method for marking image be by image U and W4 phase or.
5. the complex background leaf image dividing method of the simple mutual and mark watershed divide of combination according to claim 4, is characterized in that, adopts following conversion method to ask the gray level image of original complex background leaf image in described step S13:
Gray=0.299×Red+0.587×Green+0.114×Blue。
6. the complex background leaf image dividing method of the simple mutual and mark watershed divide of combination according to claim 5, is characterized in that, described step S15 detailed process is as follows:
S41: the marking image of take carries out mark watershed segmentation to gray level image, a* component image, b* component image respectively as parameter, obtains respectively image J1, J2, J3;
S42: obtain the bianry image O1 of original complex background leaf image in ballot mode in conjunction with J1, J2, J3;
S43: bianry image O1 is first done to closed operation, after do opening operation, obtain final bianry image O2.
7. the complex background leaf image dividing method of the simple mutual and mark watershed divide of combination according to claim 6, is characterized in that, the detailed process of described step S42 is as follows:
S51: a newly-built random and identical bianry image O11 of original complex background leaf image size;
S52: any point t to image O11, if multipair point of answering in image J1, J2, J3 respectively, there are both or belong to above the foreground area of correspondence image, determine that some t belongs to the foreground area of image O11, put 1, otherwise this point belongs to the background area of O11, is set to 0, O11 resets and obtains new bianry image O1 the most at last.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410285788.5A CN104050670B (en) | 2014-06-24 | 2014-06-24 | In conjunction with simple mutual and the complex background leaf image dividing method in labelling watershed |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410285788.5A CN104050670B (en) | 2014-06-24 | 2014-06-24 | In conjunction with simple mutual and the complex background leaf image dividing method in labelling watershed |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104050670A true CN104050670A (en) | 2014-09-17 |
| CN104050670B CN104050670B (en) | 2016-08-17 |
Family
ID=51503453
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410285788.5A Active CN104050670B (en) | 2014-06-24 | 2014-06-24 | In conjunction with simple mutual and the complex background leaf image dividing method in labelling watershed |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104050670B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106910197A (en) * | 2017-01-13 | 2017-06-30 | 广州中医药大学 | A kind of dividing method of the complex background leaf image in single goal region |
| CN110443811A (en) * | 2019-07-26 | 2019-11-12 | 广州中医药大学(广州中医药研究院) | A kind of full-automatic partition method of complex background leaf image |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6985612B2 (en) * | 2001-10-05 | 2006-01-10 | Mevis - Centrum Fur Medizinische Diagnosesysteme Und Visualisierung Gmbh | Computer system and a method for segmentation of a digital image |
| CN102467740A (en) * | 2010-11-08 | 2012-05-23 | 北京大学 | Foreground and background interactive segmentation method of large-size color image and system thereof |
| CN102156984B (en) * | 2011-04-06 | 2013-03-06 | 南京大学 | Method for determining optimal mark image by adaptive threshold segmentation |
| CN103473550B (en) * | 2013-09-23 | 2016-04-13 | 广州中医药大学 | Based on the leaf image dividing method of Lab space and local dynamic threshold |
-
2014
- 2014-06-24 CN CN201410285788.5A patent/CN104050670B/en active Active
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106910197A (en) * | 2017-01-13 | 2017-06-30 | 广州中医药大学 | A kind of dividing method of the complex background leaf image in single goal region |
| CN106910197B (en) * | 2017-01-13 | 2019-05-28 | 广州中医药大学 | A kind of dividing method of the complex background leaf image in single goal region |
| CN110443811A (en) * | 2019-07-26 | 2019-11-12 | 广州中医药大学(广州中医药研究院) | A kind of full-automatic partition method of complex background leaf image |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104050670B (en) | 2016-08-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103824049A (en) | Cascaded neural network-based face key point detection method | |
| CN107368845A (en) | A kind of Faster R CNN object detection methods based on optimization candidate region | |
| CN108681692A (en) | Increase Building recognition method in a kind of remote sensing images based on deep learning newly | |
| CN103824295A (en) | Segmentation method of adhesion vascular pulmonary nodules in lung CT (computed tomography) image | |
| CN106022355B (en) | High spectrum image sky based on 3DCNN composes joint classification method | |
| CN104268526B (en) | A kind of Chinese character picture match and deformation method | |
| CN102592268A (en) | Method for segmenting foreground image | |
| CN103473785B (en) | A kind of fast multi-target dividing method based on three-valued image clustering | |
| CN102156882A (en) | Method for detecting airport target based on high-resolution remote sensing image | |
| CN105096311A (en) | Technology for restoring depth image and combining virtual and real scenes based on GPU (Graphic Processing Unit) | |
| CN105488844B (en) | The display methods of magnanimity model real-time shadow in a kind of three-dimensional scenic | |
| CN104103076A (en) | Nuclear power plant planned restricted zone remote sensing inspecting method based on high-resolution remote sensing images | |
| CN104915976A (en) | Image processing method and system for simulating pencil sketch | |
| CN108230376A (en) | Remote sensing image processing method, device and electronic equipment | |
| CN109741268A (en) | A kind of breakage image complementing method for mural painting | |
| CN104915975A (en) | Image processing method and system for simulating crayon colored drawing | |
| CN104050670A (en) | Complicated background blade image segmentation method combining simple interaction and mark watershed | |
| CN103885942A (en) | Rapid translation device and method | |
| CN103198465A (en) | Rotation error correcting method of CT (Computerized Tomography) scanned images | |
| Zhang | Practice Teaching of Landscape Survey Course Based on eCognition Remote Sensing Image Interpretation* Technology. | |
| CN106910197B (en) | A kind of dividing method of the complex background leaf image in single goal region | |
| CN109685773B (en) | Method for determining synapse position points in immune cell image | |
| CN102270343B (en) | Image segmentation method based on Ising graph model | |
| CN103294875A (en) | Group formation simulation method based on group intelligence and self-adaption evaluation | |
| CN106339985A (en) | Method for aviation image mosaic by selecting mosaic lines from vector building data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |