CN104361330B - A kind of crop row recognition methods of corn accurate dispenser system - Google Patents
A kind of crop row recognition methods of corn accurate dispenser system Download PDFInfo
- Publication number
- CN104361330B CN104361330B CN201410710905.8A CN201410710905A CN104361330B CN 104361330 B CN104361330 B CN 104361330B CN 201410710905 A CN201410710905 A CN 201410710905A CN 104361330 B CN104361330 B CN 104361330B
- Authority
- CN
- China
- Prior art keywords
- mtd
- corn
- crop row
- image
- algorithm
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/188—Vegetation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/68—Food, e.g. fruit or vegetables
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/09—Recognition of logos
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a kind of crop row recognition methods of corn accurate dispenser system, step is:Industrial camera and camera lens collection corn field RGB color image;RGB color image to acquisition utilizes improved excessively green gray processing algorithm gray processing;Picture noise is removed using the medium filtering of improved middle value-acquiring method;Maximum variance between clusters are to the image binaryzation after denoising;The noise of binary image is filtered out using Morphology Algorithm;Based on mahalanobis distance and corn vein Rule Extraction crop row skeleton;Main crop row is fitted to straight line by the Hough transform based on main framing point.The present invention at utmost retains crop row information, removes ambient interferences, improve arithmetic speed, based on mahalanobis distance and the accurate crop row skeleton of corn vein Rule Extraction, it is prevented effectively from the influence of the noises such as weeds, adapt to Different Crop and illumination condition, crop row accuracy rate is higher than 98.3%, and for the spray shower nozzle in precision agriculture system, alignment provides effective method automatically.
Description
Technical field
The present invention relates to the technical field of agriculture project, and in particular to a kind of crop row identification of corn accurate dispenser system
Method.
Background technology
The automatic alignment of dispenser shower nozzle is realized in accuracy pesticide applying system, it is important to the identification of crop row center line.Numeral
Image processing algorithm has very big advantage in terms of automatic identification, is basis and the key technology of modern precision agriculture.In the past
Research show that crop row extraction algorithm has the shortcomings that unicity and adaptability is not strong, different times, the light of crop growth
According to and crop species the realization of algorithm can be impacted.Designing a kind of crop row recognizer for meeting a variety of conditions is
The major issue of accuracy pesticide applying.
According to conventional research, the recognizer of field-crop row is usually with the center of field-crop row or crop ditch
Line is research object.Early in last century, the Marchant and Brivot and Sweden expert in Silsoe research centers
BjornAstrand and Belgian scholar V.Leemans etc. did corresponding research to row identification and navigation algorithm respectively,
And achieve certain achievement.Recent year also occurs in that the research of many this respects, Zhang Zhibin etc. by hough conversion with
Fisher criterions are combined, and are shown that many ridges recognize unified model according to ridge line space of points relation, are overcome traditional Hough transform to extract
The deficiency of many ridge lines;Sciagraphy and Direct Hough Transform method are combined by Zhao Ruijiao etc., are proposed a kind of based on vertical histogram throwing
The method that the Improved Hough Transform of shadow detects crop row center line;Gray level image is divided into several horizontal bars by Ma Hongxia etc., is used
Vertical projection method finds out navigator fix point, and sets area-of-interest, and anchor point is fitted using Hough transform in region
Go out the datum line that navigates.In crop row extraction, skeletal extraction algorithm is also an important ring, and the accuracy of extraction directly affects work
The order of accuarcy that thing row is extracted.Traditional skeletal extraction algorithm has three kinds, and one is the method for topological thinning, and two be to be based on distance to become
The method changed, three be the method based on Voronoi diagram, but all there are some drawbacks.Slowly it is superfine propose it is a kind of based on it is European away from
From novel framework extraction algorithm, with certain improvement effect.These algorithms all have one to field-crop row extraction algorithm
Determine reference value, but algorithm is complicated, and the need for fundamentally can not meeting agricultural machinery, therefore also need to further research
And experiment.
The content of the invention
The technical problem to be solved in the present invention is to provide a kind of crop row recognition methods of corn accurate dispenser system, the party
Method can be prevented effectively from the influence of the noises such as weeds, and be adapted to Different Crop and illumination condition, and its crop row accuracy rate is higher than
98.3%。
The technical scheme is that:A kind of crop row recognition methods of corn accurate dispenser system, comprises the following steps:
S1:Utilize industrial camera and camera lens collection corn field RGB color image;
S2:RGB color image to acquisition utilizes improved excessively green gray processing algorithm progress gray processing;Wherein, it is improved
Crossing green gray processing algorithm is:Gray value Gary after gray processing is
S3:Gray level image noise is removed using the medium filtering of improved middle value-acquiring method;
S4:Binary conversion treatment is carried out to the image after denoising using maximum variance between clusters;
S5:The noise of binary image is filtered out using Morphology Algorithm;
S6:Based on mahalanobis distance and corn vein Rule Extraction crop row skeleton;
S7:Main crop row is fitted to by straight line using the Hough transform algorithm based on main framing point.
The corn field RGB color image is the image of the true complex environment of corn growth mid-term.
The improved middle value-acquiring method is:For 3*3 array of pixels M:
Wherein, a>b>C, d>e>F, g>h>I, e>b>H, d>a>G, i<f<c;Each row vector to matrix carries out descending row
Row:a>b>C, d>e>F, g>h>I, e>b>H, then the maximum d in a, d, g is the minimum in matrix M maximum, c, f, i
Value i is M minimum value, then
Due to the intermediate value of 9 elements, four elements must be more than, while again smaller than four elements, so d, i, a, f, e, h are
It is unlikely to be median.More than excluding after six pixels, minimum value g in comparison window matrix in maximum, intermediate value are only needed
In median b and minimum value in maximum c, then take the intermediate value of three, the intermediate value is array M median.
The method based on mahalanobis distance and corn vein Rule Extraction crop row skeleton is:Based on corn vein rule
Corn central point is determined, this central point is regard as main framing point;By mahalanobis distance convert other point obtain range image, to away from
Skeletal extraction is carried out from image.
The algorithm of the mahalanobis distance conversion is as follows:, wherein,Represent other points two
Tie up information,The two-dimensional signal of main framing point is represented,The inverse matrix of initial data covariance matrix is represented,Table
Show other points to the distance of main framing point.
The present invention at utmost retains crop row information by improved excessively green gray processing algorithm, removes ambient interferences, changes
The middle value-acquiring method entered improves arithmetic speed, and standard is obtained based on mahalanobis distance and corn vein Rule Extraction crop row skeleton
True skeleton, it can be prevented effectively from the influence of the noises such as weeds, be adapted to Different Crop and illumination condition, its crop row accuracy rate
Higher than 98.3%, for the spray shower nozzle in precision agriculture system, alignment provides effective method automatically.
Brief description of the drawings
Fig. 1 is flow chart of the invention.
Fig. 2(a)It is the gray-scale map using traditional excessively green algorithm process;(b)Be the present invention excessively green algorithm gray processing after
Image.
Fig. 3 is the image that the improved medium filtering of the present invention is obtained.
The maximum variance between clusters that Fig. 4 is the present invention obtain binary image.
The morphologic filtering that Fig. 5 is the present invention filters out the image after noise.
Fig. 6(a)It is the image of the method refinement based on pixel edge;(b)It is based on mahalanobis distance and corn vein rule
Crop row skeletal extraction image.
Fig. 7 is the image of the crop row center line of the Hough transform fitting based on main framing point.
Embodiment
Illustrated below in conjunction with the accompanying drawings with specific embodiment of the invention.
As shown in figure 1, a kind of crop row recognition methods of corn accurate dispenser system, comprises the following steps:
S1:Corn field RGB color image is gathered using industrial camera.
Using MV-VD030SC models industrial camera and AFT-0814MP camera lenses collection corn field RGB color image, and will
It is stored in computer program, and preserving type is BMP forms, and image size is 640*480.Pass through the image procossing based on MFC
Visual program, picture is directly displayed on program runnable interface.Corn field RGB color image is that corn growth mid-term is truly multiple
The image in heterocycle border.Generally, the corn map picture of growth medium has:Leaf present strip, intersect serious shielding, weeds and
Soil Background noise is all larger, crop row center not substantially and width compared with the features such as.
S2:RGB color image to acquisition utilizes improved excessively green gray processing algorithm progress gray processing.
Because the G components of crop plant will typically be more than R and B component, but background is not necessarily, traditional excess green
Algorithm is Gray=2G-R-B, without crop row information is farthest retained, removes ambient interferences.In traditional excess green
Adding conditional in algorithm implementation process, if G<R or G<B, Gray=255, obtaining improved excessively green gray processing algorithm is:Gray processing
Gray value Gary afterwards is
Gray level image is obtained to the RGB color image processing of acquisition using improved excessively green gray processing algorithm, and in program
Shown on interface, such as Fig. 2(b)It is shown.Image such as Fig. 2 of traditional super green image partitioning algorithm processing(a)It is shown.By figure figure
2(a)With(b)Contrast understands that improved excessively green gray processing algorithm is according to the color character of image, crop in obtained gray level image
The segmentation of row and background is more obvious.
S3:Gray level image noise is removed using the medium filtering of improved middle value-acquiring method.
Picture noise after gray processing is larger, disturbs the processing to target line, it is impossible to significantly isolate gray level image
In crop row, accordingly, it would be desirable to be filtered operation to image, and the effect of filtering process will be to processing result image
Reliability causes direct influence.The corn field picture of shooting needs continuous four filtering just to make effect more preferably, therefore, improves
Filtering speed has very big effect to reducing Riming time of algorithm.
Original median filtering algorithm, uses and carries out whole sequences to 3*3 array window, then obtain median
Substitute the gray value of the central point of matrix.And to expect the median of 9 pixels, it is necessary to which data are carried out with the ratio of 36 times
Compared with this will consume the substantial amounts of time, be unfavorable for the real-time processing of image.Therefore the present invention is to the intermediate value acquisition side of medium filtering
Method is improved, by taking 3*3 array of pixels M as an example:
Each row vector first to matrix carries out descending arrangement, it is assumed that a>b>C, d>e>F, g>h>I, e>b>H, then a, d,
Maximum in g is matrix M maximum, it is assumed that be d>a>g;The minimum value that minimum value in c, f, i is M, it is assumed that be i<f
<c;Then
Due to the intermediate value of 9 elements, four elements must be more than, while again smaller than four elements, so d, i, a, f, e, h are
It is unlikely to be median.More than excluding after six pixels, minimum value g in comparison window matrix in maximum, intermediate value are only needed
In median b and minimum value in maximum c, then take the intermediate value of three, the intermediate value is the median of 9 pixels.
Sort algorithm after improvement only needs 21 minor sorts just can find intermediate value, and algorithm speed is improved.
Using the gray level image after the median filter method process step S2 processing of improved middle value-acquiring method, use
The noise that 3*3 array window removes gray level image is as shown in Figure 3.
S4:Binary conversion treatment is carried out to the image after denoising using maximum variance between clusters.
Using maximum variance between clusters, to enter row threshold division, it calculates simple, and not dry by information such as brightness of image
Disturb, with preferable treatment effect, obtain clearly binary image after treating, as shown in Figure 4.
S5:The noise of binary image is filtered out using Morphology Algorithm.
Due to crops in the ranks with have a little relative to the less noise of crop row area on crop row, therefore use shape
State algorithm filters out noise to bianry image.Here it is main to be carried out using corrosion and expansion algorithm, crop can be made by corrosion
Row gradually shrinks to its center of gravity, and the object removal less than structural element, and the effect expanded is then on the contrary, hole can be filled up
Hole, increases the width of crop row.In order to not change and eliminate crop row useful information, Morphological scale-space is template used to be closed
Suitable, the present invention is operated using 3*3 templates to binary image, and being determined by experiment first to be corroded, and then expanded, and followed
The denoising effect that ring is 3 times is best.The result for filtering out binary image noise using this method is as shown in Figure 5.As shown in Figure 5, it is sharp
After the noise of the obtained bianry images of this method removal step S4, the noise between crop row is eliminated so that later place
Manage result more accurate.
S6:Based on mahalanobis distance and corn vein Rule Extraction crop row skeleton.
In order to extract the framework information of crop row, micronization processes are generally carried out to it using the method based on pixel edge,
Such as Fig. 6(a)It is shown.Because camera is installed on the middle part of agricultural machinery shower nozzle, the captured image of normal walking is also pair
Claim.If there is deviation, the position of middle two rows crop has more noticeable change.In order to reduce amount of calculation, letter is being considered
On the premise of breath amount and accuracy, the unnecessary lines in image are removed, only retain near two rows at center to represent crop row
Center line.In the process, traditional skeletal extraction algorithm uses Euclidean distance or other conversion close to Euclidean distance
Method generates range image, but applies the effect on corn map picture bad, and the skeleton error of extraction is more than 10%.
Method based on mahalanobis distance and corn vein Rule Extraction crop row skeleton is:Determined based on corn vein rule
Corn central point, regard this central point as main framing point;Other points are converted by mahalanobis distance and obtain range image, it is specific to become
Scaling method is as follows:, wherein,Other two-dimensional signals are represented,Represent main framing point
Two-dimensional signal,The inverse matrix of initial data covariance matrix is represented,Represent other points to main framing point away from
From.Range image will be finally obtained after conversion, traditional skeletal extraction is carried out, obtains framework information.Based on mahalanobis distance and jade
The method of rice vein Rule Extraction crop row skeleton extracts result such as Fig. 6 of crop row skeleton(b)It is shown.
S7:Main crop row is fitted to by straight line using the Hough transform algorithm based on main framing point.
Using main framing point as the main node of crop row, it is fitted and is in line using Hough transform algorithm.According to generation
The transformation relation of boundary's coordinate system and image coordinate system, it is inclined by the pixel for calculating crop row bottom centre and picture centre in image
Gap from, we can draw the size of the actual geographic deviation corresponding to image pixel deviation, Real-time Feedback control information, from
And realize the automatic alignment of accuracy pesticide applying system mechanics shower nozzle.Make using the Hough transform algorithm fitting based on main framing point is main
The result of thing row is as shown in Figure 7.
As shown in Figure 7, the crop row and the goodness of fit of the actual crop row of corn extracted using the method for the present invention is higher,
Capable trend and the trend of actual crop row are consistent substantially.Found by actual test data comparison, method of the invention
The crop row of extraction and the actual goodness of fit are more than 99.3%, and worst error is no more than 1mm.This is reduced to a certain extent
System error rate, improves system accuracy, reliable deviation data is provided for follow-up accuracy pesticide applying.
Present invention could apply to the extraction of the crop row of other Different Crops, such as wheat.
The foregoing is only a preferred embodiment of the present invention, but protection scope of the present invention be not limited thereto,
Any one skilled in the art the invention discloses technical scope in, the change or replacement that can be readily occurred in,
It should all be included within the scope of the present invention.Therefore, protection scope of the present invention should be with the protection model of claims
Enclose and be defined.
Claims (2)
1. a kind of crop row recognition methods of corn accurate dispenser system, it is characterised in that its step is as follows:
S1:Utilize industrial camera and camera lens collection corn field RGB color image;
S2:RGB color image to acquisition utilizes improved excessively green gray processing algorithm progress gray processing;Wherein, it is improved excessively green
Gray processing algorithm is:Gray value Gary after gray processing is
S3:Gray level image noise is removed using the medium filtering of improved middle value-acquiring method;
S4:Binary conversion treatment is carried out to the image after denoising using maximum variance between clusters;
S5:The noise of binary image is filtered out using Morphology Algorithm;
S6:Based on mahalanobis distance and corn vein Rule Extraction crop row skeleton;
S7:Main crop row is fitted to straight line by the Hough transform algorithm based on main framing point;
The improved middle value-acquiring method is:For 3*3 array of pixels M:
<mrow>
<mi>M</mi>
<mo>=</mo>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<mi>a</mi>
</mtd>
<mtd>
<mi>b</mi>
</mtd>
<mtd>
<mi>c</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mi>d</mi>
</mtd>
<mtd>
<mi>e</mi>
</mtd>
<mtd>
<mi>f</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mi>g</mi>
</mtd>
<mtd>
<mi>h</mi>
</mtd>
<mtd>
<mi>i</mi>
</mtd>
</mtr>
</mtable>
</mfenced>
</mrow>
Wherein, a>b>C, d>e>F, g>h>I, e>b>H, d>a>G, i<f<c;Each row vector to matrix carries out descending arrangement:a
>b>C, d>e>F, g>h>I, e>b>H, then the minimum value i that the maximum d in a, d, g is in matrix M maximum, c, f, i are
M minimum value, then
Due to the intermediate value of 9 elements, four elements must be more than, while again smaller than four elements, so d, i, a, f, e, h can not
It can be median;More than excluding after six pixels, minimum value g in comparison window matrix in maximum is only needed, in intermediate value
Maximum c in median b and minimum value, then takes the intermediate value of three, and the intermediate value is array M median;
The method that the Morphology Algorithm filters out the noise of binary image is:It is advanced to binary image using 3*3 template
Row corrosion, then expands, and circulates 3 times;
The method based on mahalanobis distance and corn vein Rule Extraction crop row skeleton is:Determined based on corn vein rule
Corn central point, regard this central point as main framing point;Other points are converted by mahalanobis distance and obtain range image, figure of adjusting the distance
As carrying out skeletal extraction;
The method of the mahalanobis distance conversion is as follows:Wherein,Represent other point two dimension letters
Breath,Represent the two-dimensional signal of main framing point, Σ-1The inverse matrix of initial data covariance matrix is represented,Represent other
Distance of the point to main framing point.
2. the crop row recognition methods of corn accurate dispenser system according to claim 1, it is characterised in that the corn
Field RGB color image is the image of the true complex environment of corn growth mid-term.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410710905.8A CN104361330B (en) | 2014-12-01 | 2014-12-01 | A kind of crop row recognition methods of corn accurate dispenser system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410710905.8A CN104361330B (en) | 2014-12-01 | 2014-12-01 | A kind of crop row recognition methods of corn accurate dispenser system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104361330A CN104361330A (en) | 2015-02-18 |
CN104361330B true CN104361330B (en) | 2017-09-22 |
Family
ID=52528588
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410710905.8A Active CN104361330B (en) | 2014-12-01 | 2014-12-01 | A kind of crop row recognition methods of corn accurate dispenser system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104361330B (en) |
Families Citing this family (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105117701B (en) * | 2015-08-21 | 2018-06-15 | 郑州轻工业学院 | Corn crop row framework extraction method based on largest square principle |
CN105389581B (en) * | 2015-10-15 | 2019-08-06 | 哈尔滨工程大学 | A kind of rice germ plumule integrity degree intelligent identifying system and its recognition methods |
CN105987684A (en) * | 2015-12-18 | 2016-10-05 | 中国科学院合肥物质科学研究院 | Monocular vision-based agricultural vehicle navigation line detection system and method |
CN106909881A (en) * | 2017-01-16 | 2017-06-30 | 中国农业大学 | The method and system of corn breeding base ridge number are extracted based on unmanned aerial vehicle remote sensing images |
CN110288608B (en) * | 2018-03-19 | 2024-04-12 | 北京京东尚科信息技术有限公司 | Crop row center line extraction method and device |
CN109190452B (en) * | 2018-07-09 | 2021-03-02 | 北京农业智能装备技术研究中心 | Crop row identification method and device |
CN110178605B (en) * | 2019-06-28 | 2021-03-30 | 重庆文理学院 | Automatic sprinkling system of accurate pesticide to kiwi fruit leaf disease |
CN110516563A (en) * | 2019-08-06 | 2019-11-29 | 西安电子科技大学 | Agriculture transplanter intelligence method for path navigation based on DSP |
CN110765927B (en) * | 2019-10-21 | 2022-11-25 | 广西科技大学 | Identification method of associated weeds in vegetation community |
CN113469917A (en) * | 2021-07-14 | 2021-10-01 | 上海影卓信息科技有限公司 | Method, system and medium for removing image noise based on morphology |
WO2023090988A1 (en) | 2021-11-16 | 2023-05-25 | Sime Darby Plantation Intellectual Property Sdn Bhd | Method for detecting and/or counting trees |
CN116168193B (en) * | 2023-02-22 | 2024-03-19 | 七海行(深圳)科技有限公司 | Calibration method, device and equipment of medicine application equipment |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103530606A (en) * | 2013-09-30 | 2014-01-22 | 中国农业大学 | Agricultural machine navigation path extraction method under weed environment |
CN103559493A (en) * | 2013-10-21 | 2014-02-05 | 中国农业大学 | Method for extracting linear ground objects |
CN103914836A (en) * | 2014-03-24 | 2014-07-09 | 郑州轻工业学院 | Farmland machine leading line extraction algorithm based on machine vision |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP4659631B2 (en) * | 2005-04-26 | 2011-03-30 | 富士重工業株式会社 | Lane recognition device |
-
2014
- 2014-12-01 CN CN201410710905.8A patent/CN104361330B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103530606A (en) * | 2013-09-30 | 2014-01-22 | 中国农业大学 | Agricultural machine navigation path extraction method under weed environment |
CN103559493A (en) * | 2013-10-21 | 2014-02-05 | 中国农业大学 | Method for extracting linear ground objects |
CN103914836A (en) * | 2014-03-24 | 2014-07-09 | 郑州轻工业学院 | Farmland machine leading line extraction algorithm based on machine vision |
Non-Patent Citations (2)
Title |
---|
基于Hough变换的农业机械视觉导航基准线识别;马红霞 等;《农机化研究》;20130430(第4期);第37页第2段 * |
田间路径识别算法和基于立体视觉的车辆自动导航方法研究;张方明;《中国优秀博硕士学位论文全文数据库 (博士) 信息科技辑》;20060915(第09期);第40页 * |
Also Published As
Publication number | Publication date |
---|---|
CN104361330A (en) | 2015-02-18 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104361330B (en) | A kind of crop row recognition methods of corn accurate dispenser system | |
CN110232389B (en) | Stereoscopic vision navigation method based on invariance of green crop feature extraction | |
Maldonado Jr et al. | Automatic green fruit counting in orange trees using digital images | |
Li et al. | Robust visual tracking based on convolutional features with illumination and occlusion handing | |
CN105021196B (en) | Crop row detection method based on minimum tangent circle and morphology principle | |
CN106327507B (en) | A kind of color image conspicuousness detection method based on background and foreground information | |
Fernandez et al. | Automatic detection of field-grown cucumbers for robotic harvesting | |
CN109961049A (en) | Cigarette brand recognition methods under a kind of complex scene | |
CN105279769B (en) | A kind of level particle filter tracking method for combining multiple features | |
CN107886086A (en) | A kind of target animal detection method and device based on image/video | |
CN103440667B (en) | The automaton that under a kind of occlusion state, moving target is stably followed the trail of | |
CN103218605A (en) | Quick eye locating method based on integral projection and edge detection | |
CN102117487A (en) | Scale-direction self-adaptive Mean-shift tracking method aiming at video moving object | |
US10089712B2 (en) | System and method providing automatic alignment of aerial/satellite imagery to known ground features | |
CN107895379A (en) | The innovatory algorithm of foreground extraction in a kind of video monitoring | |
CN109685045A (en) | A kind of Moving Targets Based on Video Streams tracking and system | |
CN106803257A (en) | The dividing method of scab in a kind of crop disease leaf image | |
CN103714556A (en) | Moving target tracking method based on pyramid appearance model | |
CN105117701A (en) | Corn crop row skeleton extraction method based on largest square principle | |
CN108596920A (en) | A kind of Target Segmentation method and device based on coloured image | |
CN103413312A (en) | Video target tracking method based on neighborhood components analysis and scale space theory | |
CN109359549A (en) | A kind of pedestrian detection method based on mixed Gaussian and HOG_LBP | |
CN106529441B (en) | Depth motion figure Human bodys' response method based on smeared out boundary fragment | |
CN111199245A (en) | Rape pest identification method | |
CN111582198B (en) | Automatic sea-land segmentation method for remote sensing image |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |