CN103198298A - Mild insect pest lamina vein recognition method based on derivative spectrum method - Google Patents
Mild insect pest lamina vein recognition method based on derivative spectrum method Download PDFInfo
- Publication number
- CN103198298A CN103198298A CN2013100865365A CN201310086536A CN103198298A CN 103198298 A CN103198298 A CN 103198298A CN 2013100865365 A CN2013100865365 A CN 2013100865365A CN 201310086536 A CN201310086536 A CN 201310086536A CN 103198298 A CN103198298 A CN 103198298A
- Authority
- CN
- China
- Prior art keywords
- image
- vein
- derivative
- recognition methods
- hsi
- 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 provides a mild insect pest lamina vein recognition method based on a derivative spectrum method. The mild insect pest lamina vein recognition method based on the derivative spectrum method includes: a first derivative spectrum of an original hyperspectral image is calculated, a first derivative spectrum image at the position of 680 mm is obtained, and vein extraction calculation is performed for the first derivative spectrum image to obtain a first vein image; single band images at the position of 640mm, 550mm and 460mm in the original hyperspectral image are acquired to organize an RGB image, then the RGB image is converted into HSI space to obtain an HSI image, and the vein extraction calculation is performed for an H component in the HSI image to obtain a second vein image; and the first vein image and the second vein image are fused to obtain a vein image after recognition. Compared with the prior art, with the mild insect pest lamina vein recognition method based on the derivative spectrum method, first derivative transforming and HSI color space transforming are respectively performed for the original hyperspectral image, then the vein extraction calculation is performed, vein recognition can be performed for a mild insect pest lamina, and a recognized lamina image is clear.
Description
Technical field
The present invention relates to a kind of vein recognition methods, relate in particular to a kind of vein recognition methods of the slight insect pest blade based on derivative spectrophotometry.
Background technology
Because the complicacy of plant, the plant modeling is more more difficult than Building Modeling.The three-dimensional model inhibition of setting up plant is that botany and computer graphics etc. are studied focus easily, generally can be divided into microcosmic and two yardsticks of macroscopic view.
The micro-scale modeling is to set up model according to cell, nutrient and plant growth rhythm etc., perhaps sets up model by concrete definition limb, leaf and apparatus derivatorius.The macro-scale modeling refers to set up scenes such as the forest that comprises plant, crops, meadow.Wherein, the accuracy of model is paid attention in the micro-scale modeling, requires to tally with the actual situation.
The blade vein has comprised very important plant physiology information, and extracting plant leaf vein is the committed step of carrying out plant modeling and identification, and vein identification also is the focus of research always.
Along with the development of Computer Processing technology, extract and the vein modeling has become the research focus based on the vein of image processing techniques.Document (2006.Leaf vein extraction using independent component analysis.Proceedings of IEEE Conference on Systems, Man, and Cybernetics.5:3890-3894.) carries out the extraction of vein with independent component analysis; Document (2008. plant leaf veins based on the Hough conversion detect new method [J]. communication and computing machine .5 (8): 58-60.) utilize gray scale stretching, Hough conversion and marginal growth, image corrosion to carry out vein with expansion and detect.
Document (2009.Fast leaf vein extraction using hue and intensity information.Proceedings of IEEE International Conference on Information Engineering and Computer Society.1-4.) proposes the fast look vein extraction algorithm based on the SHI color space, this algorithm is divided into vein and mesophyll color similarity and different two classes to leaf, for the leaf with different colours, the tone of HSI color space is divided between 12 chromatic zoneses, be the mesophyll pixel between the maximum chromatic zones of pixel, remove the mesophyll pixel and obtain the vein image; The leaf single for color extracts in two steps, and the first step is divided vein according to the extracting method of different colours leaf to H component image extraction unit, and second step strengthened the I component image earlier, carried out the extraction of most of vein again.
Document (the 2011. vein extraction algorithms that combine based on improved Sobel operator and hue information. Transactions of the Chinese Society of Agricultural Engineering .27 (7): 196～199.) then proposed a kind of vein extraction algorithm that combines based on improved Sobel operator and hue information, the vein that image is carried out improved Sobel operator extracts and extracts based on the vein of hue information, and the result that both are extracted merges and obtains final vein image then.
In correlative study in the past, the experimental subjects of employing all is healthy blade, except the intrinsic aberration of vein and mesophyll, does not have other aberration zone, is conducive to the identification of vein.But in actual applications, blade often all can suffer the invasion of various disease and pests.The leaf chlorophyll of insect infestation harm distributes can be inhomogeneous, form aberration, have tiny worm channel simultaneously, when carrying out existing edge recognizer, the border of these color non-uniform areas and the border of tiny worm channel are easy to the edge of vein simultaneously identified, are difficult to avoid.
Summary of the invention
The invention provides a kind of vein recognition methods of the slight insect pest blade based on derivative spectrophotometry, utilize this algorithm can obtain the vein image of slight insect pest blade.
Vein recognition methods based on the slight insect pest blade of derivative spectrophotometry comprises:
(1) former high spectrum image is asked first derivative spectrum, obtain the first derivative spectrum image at 680nm place, this first derivative spectrum image is carried out vein extract computing, obtain the first vein image;
(2) get 640nm, 550nm in the former high spectrum image, 460nm place single band image composition RGB image, again this RGB image transitions is obtained the HSI image to the HSI space, the H component in the HSI image is carried out vein extract computing, obtain the second vein image;
(3) the first vein image and the second vein image are merged the vein image after obtaining identifying.
Slight insect pest blade described in the present invention refers to that vein is complete, mesophyll is corroded by insect but still keeps the blade of at least 80% mesophyll area.
It is for mesophyll pixel and vein pixel region are separated that former high spectrum image is asked first derivative spectrum.Because the curve of spectrum of single pixel is the curve that is linked to be by a series of discrete spectrum data, therefore, to high-spectral data carry out the derivative conversion be not equal on the mathematical meaning to continuously, function that can be little carries out derivative operation, but realizes approximate replacement to first derivation by the single order difference coefficient under the differential window of a dimensioning.Therefore need be to the high-spectral data log-transformation of going ahead of the rest, go the differentiate conversion again.The formula of log-transformation and derivative conversion is respectively suc as formula (1), shown in (2).
A(λ)＝Ln[1R(λ)] （1）；
D(λ)＝[A(λ)-A(λ+ω)]ω （2）；
Wherein, λ represents wavelength location, the original spectrum reflected value of R (λ) expression af at wavelength lambda, and A (λ) is the numerical value after R (λ) the process log-transformation, ω represents the yardstick of differential window, the first derivative values of D (λ) expression af at wavelength lambda spectral reflectance value.
In the differentiate conversion, the size of ω has played crucial effects to the validity of information extraction.Choose less differential window meticulous spectrum metamorphosis information can be provided, but the while also can be amplified the high frequency noise in the spectrum; When choosing bigger differential window, curve there is certain smoothing denoising function (it is level and smooth to be similar to weighting), but the differential window is excessive also can be level and smooth with the flex point on the curve of spectrum and extreme point, lose the important information that carry in place, concavo-convex peak on certain spectrum metamorphosis information, the especially curve of spectrum.
As preferably, the yardstick of differential window is 60 wave bands.This moment, the curve of spectrum was comparatively level and smooth, and several main absorption peaks and reflection peak also can completely keep, mesophyll and vein pixel curve are all obviously distinguished at the 680nm place, therefore the present invention adopts derivative spectrophotometry to extract the first derivative spectrum image at 680nm place, carries out vein on the basis of this first derivative spectrum image and extracts computing.
In the step (1), the first derivative spectrum image at 680nm place is carried out vein when extracting computing, comprise that successively rim detection, linear space filtering and mathematical morphology handle.
Rim detection is the uncontinuity at the brightness of image value, detects this uncontinuity by first order derivative and second derivative.The first order derivative of selecting during image is handled is the amplitude of image gradient, and second derivative then Laplace operator commonly used is calculated.The basic purpose of rim detection is to use one of following two basic norms to find the fast-changing place of brightness in image:
(1) first order derivative that finds brightness value place bigger than specified threshold value on amplitude;
(2) find the second derivative of brightness value that the place of zero crossing is arranged.
Rim detection mainly adopts various operators to come saliency maps so as the pixel of fringe region, and makes image binaryzation.Edge detection operator commonly used has Sobel operator, Gaussian Laplce (LoG) operator, Canny operator etc.
The Sobel operator uses the approximate derivative of Sobel mask to search the edge, is suitable for containing the image of the more and gray scale gradual change of noise; The LoG operator is to use the Laplace operator of Gaussian function that image is carried out convolution to produce the dual edge image, locatees final edge with zero crossing again; The Canny operator uses the Gaussian filter have the specified value deviation to come smoothly to determine marginal point to reduce noise by partial gradient and the edge direction of every bit, then with these edge thinnings and carry out edge link and obtained final edge image.
As preferably, rim detection described in the present invention adopts the Canny operator, and threshold value is [0.04,0.1], and sigma is 1.5.Has the Canny operator of this threshold value and sigma value to the extraction effect the best at vein edge.
Slight insect pest blade is carried out in the image that rim detection obtains, and it is borders that blade is subjected to the unhealthy zone that forms after the insect pest that many tiny boundary curves are arranged, and need remove by linear space filtering.
Described linear space filtering adopts formula (3) to calculate:
Wherein, R is the response of the linear space filtering of certain pixel, and w is the mask coefficient, and z is and this coefficient corresponding gray that mn is the pixel sum that comprises in the mask.
The concept of filtering comes from the Fourier transform of signal being handled at frequency domain.Spatial filtering refers to the direct operation that image pixel is handled, and normally mask is moved in pointwise in pending image, comes calculated response at each pixel place wave filter by the relation of predefined.Linear space filtering is carried out linear operation at image pixel, and the response of linear space filtering is worth the sum of products to provide by the respective pixel in wave filter and the inswept zone of filtering mask.
Carry out the mathematical morphology processing after linear space filtering is finished again and obtain the first vein image.
The main application of mathematical morphology is to obtain image topological sum structural information, and interacting by image and structural element obtains the most essential form, and the main application in image is handled is:
(1) utilizes the mathematical morphology fundamental operation, image is observed and handled, to improve picture quality;
(2) geometric properties and the parameter of description and definition image are as girth, area, skeleton, directivity, interconnectedness etc.
Expanding with corrosion is the basis that morphology is handled, and all needs to set a structural element, does AND-operation by structural element and original image and realizes.Expansion has the effect of " lengthening " or " chap " according to the size of structural element to former figure, and corrosion then has the effect of " contraction " or " refinement ", and degree and the mode of " lengthening " and " contraction " are all controlled by structural element.
Particularly, expansion procedure is as follows:
(1) with each pixel of structural element scanning original image;
(2) do AND-operation with the bianry image of structural element and its covering;
(3) if be 0 all, this pixel is 0 in the result images, otherwise is 1.
The corrosion treatment process is as follows:
(1) with each pixel of structural element scanning original image;
(2) do AND-operation with the bianry image of structural element and its covering;
(3) if be 1 all, this pixel is 1 in the result images, otherwise is 0.
Expansion process can keep original shape of image, and corrosion treatment can be removed details less in the image usually.With expansion process and corrosion treatment combination, can carry out complicated morphology such as opening operation, closed operation, top cap conversion and handle.
In the step of the present invention (1), mathematical morphology is handled and is adopted expansion process and the corrosion treatment that hockets.The number of times that expansion process and corrosion treatment replace is preferably 3～5 times.The final first vein image that obtains.
In the first vein image that obtains, the vein edge is still clear inadequately, need be undertaken perfect by the HSI color space.The HSI color space is made of H component (tone), S component (saturation degree) and I component (brightness), by tone, saturation degree and brightness color is described, this model separates luminance component and the color information in the coloured image, is a kind of desirable image processing algorithm based on color description.
Because the former high spectrum image of collection of the present invention has only the RGB type, there is not the HSI type, need in the step (2) 640nm, 550nm, 460nm place single band image in the former high spectrum image are formed the RGB image, again this RGB image transitions is obtained the HSI image to the HSI space, transfer equation is as follows:
Wherein, H, S, I represent the value of 3 components in the HSI space respectively, and R, G, B represent the value of 3 components in the rgb space respectively.Near people's visually-perceptible, differentiation power is stronger, can be used for extracting the trunk vein to the ability of color description for H component in the HSI space.
In the step (2), the H component in the HSI image is carried out vein when extracting computing, comprise successively that rim detection, threshold value are cut apart, linear space filtering and mathematical morphology handle.
Wherein, rim detection can adopt the algorithm identical with step (1); Carry out threshold value when cutting apart, threshold value is preferably 120～130, and more preferably 120; When carrying out linear space filtering, choose the template that size is 30 * 30 pixels; Adopting expansion process and hole to be filled into line number morphology after filtration is finished handles; Obtain the second vein image.
The second vein image that the first vein image that step (1) is obtained and step (2) obtain merges the vein image after namely obtaining identifying.
Compared with prior art, beneficial effect of the present invention is:
The present invention carries out first order derivative conversion and HSI color space transformation respectively to former high spectrum image, carries out vein again and extracts computing, can carry out vein identification to slight insect pest blade, and the vein image of identification is comparatively clear.
Description of drawings
Fig. 1 is the operational flowchart of vein recognition methods that the present invention is based on the slight insect pest blade of derivative spectrophotometry;
Fig. 2 is the first derivative spectrum figure at the 680nm place of former high spectrum image; ω=60 represent that namely the yardstick of differential window is 60 wave bands;
Fig. 3 (a) is the edge detection results of Canny operator (threshold value is [0.0187,0.0468], and sigma gets default value 1);
Fig. 3 (b) is the edge detection results of Sobel operator (threshold value is 0.072);
Fig. 3 (c) is the edge detection results of LoG operator (threshold value is 0.00237, and sigma gets default value 2);
Fig. 3 (d) is the edge detection results of Canny operator (threshold value is [0.04,0.1], and sigma gets 1.5);
Fig. 3 (e) is the edge detection results of Sobel operator (threshold value is 0.05);
Fig. 3 (f) is the edge detection results of LoG operator (threshold value is made as 0.003, and sigma gets 2.2);
Fig. 4 (a) is for removing the vein edge image that the imperfect blade of left and right sides inferior horn obtains among Fig. 3 (d);
Fig. 4 (b) is the first vein image;
Fig. 4 (c) is the H component image of HSI image;
Fig. 4 (d) is the second vein image;
Fig. 4 (e) is the vein image of final identification.
Embodiment
Below in conjunction with the drawings and specific embodiments the present invention is described in further detail.
This embodiment is example with slight insect pest rape leaf, describes the step (as shown in Figure 1) of the vein recognition methods of the slight insect pest blade that the present invention is based on derivative spectrophotometry in detail, specifically comprises:
(1) former high spectrum image is asked first derivative spectrum, obtain the first derivative spectrum image at 680nm place, this first derivative spectrum image is carried out vein extraction computing obtain the first vein image;
First derivative spectrum through type (1) (2) obtains:
A(λ)＝Ln[1R(λ)] （1）；
D(λ)＝[A(λ)-A(λ+ω)]ω （2）；
Wherein λ represents wavelength location, the original spectrum reflected value of R (λ) expression af at wavelength lambda, and A (λ) is the numerical value after R (λ) the process log-transformation, ω represents the yardstick of differential window, the first derivative values of D (λ) expression af at wavelength lambda spectral reflectance value.
In this embodiment, the yardstick of differential window is chosen for 60 wave bands, under this differential window, mesophyll and vein pixel curve all can significantly be distinguished at the 680nm place, as shown in Figure 2.
The first derivative spectrum image at 680nm place is carried out vein when extracting computing, comprise that successively rim detection, linear space filtering and mathematical morphology handle, detailed process is as follows:
Rim detection is used the Canny operator, and threshold value is made as [0.04,0.1], and sigma gets 1.5; Testing result is seen Fig. 3 (d);
As a comparison, use respectively adaptive threshold Canny, Sobel and LoG operator the single-range first derivative spectrum image of 680nm is done the rim detection computing; Wherein:
When using the Canny operator, threshold value is [0.0187,0.0468], and sigma gets default value 1;
When using the Sobel operator, threshold value is 0.072;
When using the LoG operator, threshold value is that 0.00237, sigma gets default value 2;
(b) shown in (c), as seen from the figure, the edge of Canny operator and the identification of LoG operator is too much as Fig. 3 (a) for the edge detection results of these three kinds of operators, and the edge of Sobel operator identification is very few.
Be re-used as contrast, use the Sobel of self-defined threshold value and LoG respectively the single-range first derivative spectrum image of 680nm to be done the rim detection computing respectively, wherein:
When using the Sobel operator, threshold value was made as 0.05 o'clock;
When using the LoG operator, threshold value is made as 0.003, sigma and gets 2.2;
The edge detection results of these two kinds of operators as Fig. 3 (e) (f) shown in.
By above contrast as seen, use the Canny operator, threshold value is made as [0.04,0.1], and sigma gets 1.5 o'clock edge extracting best results.
Obtain Fig. 3 (d) after utilizing Canny operator extraction edge, the more imperfect blade of left and right sides inferior horn among the figure is removed, obtain the vein edge with the identification of Canny operator shown in Fig. 4 (a).
Many fine edge curves are arranged among Fig. 4 (a), are because blade is caused by the border in the unhealthy zone that forms after the insect pest.Need remove by linear space filtering.
Linear space filtering adopts formula (3) to calculate:
Wherein, R is the linear filtering response of certain pixel, and w is the mask coefficient, and z is and this coefficient corresponding gray that mn is the pixel sum that mask comprises.
Carry out expansion process and corrosion treatment (triplicate) after linear space filtering is finished again and obtain the first vein image such as Fig. 4 (b).Shown in Fig. 4 (b), the master pulse stage casing is jagged, and this breach is difficult to fill to fill up with expansion or hole.
(2) 640nm, 550nm, the 460nm place single band image of getting in the former high spectrum image formed the RGB image, again this RGB image transitions obtained the HSI image to the HSI space, the H component in the HSI image carried out vein extract computing, obtains the second vein image;
Three single band images of 640nm in the former high spectrum image, 550nm and 460nm are the red, green, blue component of corresponding rgb spaces respectively, and the RGB image transitions that these three single band figure are formed obtains the HSI image to the HSI color space.
The H component of HSI image is shown in Fig. 4 (c), and the stage casing of trunk vein is obviously darker than blade remainder color among the figure, can carry out vein extraction computing it is separated.Vein in the step (2) extracts computing and comprises: rim detection, threshold value are cut apart, linear space filtering and mathematical morphology are handled.Concrete operations are as follows:
Rim detection also adopts the Canny operator, and threshold value is made as [0.04,0.1], and sigma gets 1.5; Carrying out threshold value (threshold value adopts 120) after rim detection is finished cuts apart; When carrying out linear space filtering, choose size and be 30 * 30 template and slip over all pixels on the image, the noise object that specific filtration resistance filtering template is little; After linear space filtering is finished again doing mathematics morphology handle, mathematical morphology is handled and is comprised that expansion process and hole fill, wherein expansion process is 3 times, it is 2 square structure element that structural element is chosen the length of side; When hole was filled, the hole stuffing function can adopt prior art; Obtain the second vein image shown in Fig. 4 (d).
(3) the first vein image and the second vein image are merged vein image after obtaining identifying;
At last, with Fig. 4 (b) and Fig. 4 (d) addition, the vein image after namely obtaining identifying is shown in Fig. 4 (e).
Claims (9)
1. the vein recognition methods based on the slight insect pest blade of derivative spectrophotometry is characterized in that, comprising:
(1) former high spectrum image is asked first derivative spectrum, obtain the first derivative spectrum image at 680nm place, this first derivative spectrum image is carried out vein extract computing, obtain the first vein image;
(2) get 640nm, 550nm in the former high spectrum image, 460nm place single band image composition RGB image, again this RGB image transitions is obtained the HSI image to the HSI space, the H component in the HSI image is carried out vein extract computing, obtain the second vein image;
(3) the first vein image and the second vein image are merged the vein image after obtaining identifying.
2. vein recognition methods as claimed in claim 1 is characterized in that, when asking first derivative spectrum, the yardstick of differential window is 60 wave bands.
3. vein recognition methods as claimed in claim 1 is characterized in that, in the step (1), the first derivative spectrum image at 680nm place is carried out vein when extracting computing, comprises that successively rim detection, linear space filtering and mathematical morphology handle.
4. vein recognition methods as claimed in claim 3 is characterized in that, described rim detection adopts the Canny operator, and threshold value is [0.04,0.1], and sigma is 1.5.
5. vein recognition methods as claimed in claim 3 is characterized in that, in the step (1), mathematical morphology is handled and adopted expansion process and the corrosion treatment that hockets.
6. vein recognition methods as claimed in claim 4 is characterized in that, the number of times that expansion process and corrosion treatment replace is 3～5 times.
7. vein recognition methods as claimed in claim 1 is characterized in that, in the step (2), the H component in the HSI image is carried out vein when extracting computing, comprises successively that rim detection, threshold value are cut apart, linear space filtering and mathematical morphology handle.
8. vein recognition methods as claimed in claim 7 is characterized in that, carries out threshold value when cutting apart, and threshold value is 120～130.
9. vein recognition methods as claimed in claim 7 is characterized in that, chooses size and is the template of 30 * 30 pixels and carry out linear space filtering.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2013100865365A CN103198298A (en) | 2013-03-18 | 2013-03-18 | Mild insect pest lamina vein recognition method based on derivative spectrum method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2013100865365A CN103198298A (en) | 2013-03-18 | 2013-03-18 | Mild insect pest lamina vein recognition method based on derivative spectrum method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103198298A true CN103198298A (en) | 2013-07-10 |
Family
ID=48720834
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2013100865365A Pending CN103198298A (en) | 2013-03-18 | 2013-03-18 | Mild insect pest lamina vein recognition method based on derivative spectrum method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103198298A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105005813A (en) * | 2015-06-26 | 2015-10-28 | 广州铁路职业技术学院 | Insect pest analyzing and counting method and system |
| CN105160676A (en) * | 2015-08-31 | 2015-12-16 | 中国烟草总公司广东省公司 | Cured tobacco rib image extraction method |
| CN106778888A (en) * | 2016-12-27 | 2017-05-31 | 浙江大学 | A kind of orchard pest and disease damage survey system and method based on unmanned aerial vehicle remote sensing |
| CN107369176A (en) * | 2017-07-14 | 2017-11-21 | 华南理工大学 | A kind of flexible IC substrates oxidation area detection system and method |
| CN107367515A (en) * | 2017-07-14 | 2017-11-21 | 华南理工大学 | A kind of ultrathin flexible IC substrate ink foreign matter detection systems and method |
| CN109948631A (en) * | 2019-03-26 | 2019-06-28 | 山东山大新元易通信息科技有限公司 | A kind of junior tobacco leaf blue veins cigarette automatic identifying method and blue veins cigarette detect alarm system |
| CN113065589A (en) * | 2021-03-26 | 2021-07-02 | 北京林业大学 | Method and device for prejudging whether potential parent relation exists in sample based on leaf characteristics |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6282301B1 (en) * | 1999-04-08 | 2001-08-28 | The United States Of America As Represented By The Secretary Of The Army | Ares method of sub-pixel target detection |
| CN101881726A (en) * | 2010-06-18 | 2010-11-10 | 北京农业智能装备技术研究中心 | Nondestructive detection method for comprehensive character living bodies of plant seedlings |
-
2013
- 2013-03-18 CN CN2013100865365A patent/CN103198298A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6282301B1 (en) * | 1999-04-08 | 2001-08-28 | The United States Of America As Represented By The Secretary Of The Army | Ares method of sub-pixel target detection |
| CN101881726A (en) * | 2010-06-18 | 2010-11-10 | 北京农业智能装备技术研究中心 | Nondestructive detection method for comprehensive character living bodies of plant seedlings |
Non-Patent Citations (4)
| Title |
|---|
| 刘炜等: "不同尺度的微分窗口下土壤有机质的一阶导数光谱响应特征分析", 《红外与毫米波学报》 * |
| 李灿灿等: "基于改进的Sobel算子和色调信息的叶脉提取方法", 《农业工程学报》 * |
| 田有文等: "高光谱图像技术诊断温室黄瓜病害的方法", 《农业工程学报》 * |
| 黄木易等: "冬小麦条锈病的光谱特征及遥感监测", 《农业工程学报》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105005813A (en) * | 2015-06-26 | 2015-10-28 | 广州铁路职业技术学院 | Insect pest analyzing and counting method and system |
| CN105005813B (en) * | 2015-06-26 | 2017-11-10 | 广州铁路职业技术学院 | Method of counting and system are analyzed in insect pest |
| CN105160676A (en) * | 2015-08-31 | 2015-12-16 | 中国烟草总公司广东省公司 | Cured tobacco rib image extraction method |
| CN106778888A (en) * | 2016-12-27 | 2017-05-31 | 浙江大学 | A kind of orchard pest and disease damage survey system and method based on unmanned aerial vehicle remote sensing |
| CN107369176A (en) * | 2017-07-14 | 2017-11-21 | 华南理工大学 | A kind of flexible IC substrates oxidation area detection system and method |
| CN107367515A (en) * | 2017-07-14 | 2017-11-21 | 华南理工大学 | A kind of ultrathin flexible IC substrate ink foreign matter detection systems and method |
| CN107367515B (en) * | 2017-07-14 | 2019-11-15 | 华南理工大学 | A kind of ultrathin flexible IC substrate ink foreign matter detecting method |
| CN107369176B (en) * | 2017-07-14 | 2021-07-20 | 华南理工大学 | System and method for detecting oxidation area of flexible IC substrate |
| CN109948631A (en) * | 2019-03-26 | 2019-06-28 | 山东山大新元易通信息科技有限公司 | A kind of junior tobacco leaf blue veins cigarette automatic identifying method and blue veins cigarette detect alarm system |
| CN109948631B (en) * | 2019-03-26 | 2019-11-05 | 山东山大新元易通信息科技有限公司 | A kind of junior tobacco leaf blue veins cigarette automatic identifying method and blue veins cigarette detect alarm system |
| CN113065589A (en) * | 2021-03-26 | 2021-07-02 | 北京林业大学 | Method and device for prejudging whether potential parent relation exists in sample based on leaf characteristics |
| CN113065589B (en) * | 2021-03-26 | 2023-08-15 | 北京林业大学 | Method and device for pre-judging whether potential parent relationship exists in sample based on blade characteristics |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103198298A (en) | Mild insect pest lamina vein recognition method based on derivative spectrum method | |
| CN103310218B (en) | A kind of overlap blocks fruit precise recognition method | |
| Liu et al. | Detection of citrus fruit and tree trunks in natural environments using a multi-elliptical boundary model | |
| CN113837974B (en) | NSST domain power equipment infrared image enhancement method based on improved BEEPS filtering algorithm | |
| CN109238182B (en) | Fourier spectrum characteristic-based objective fabric flatness rating method | |
| CN103247049A (en) | SMT (Surface Mounting Technology) welding spot image segmentation method | |
| CN106815819A (en) | Many strategy grain worm visible detection methods | |
| CN101833658A (en) | Illumination invariant extracting method for complex illumination face recognition | |
| CN104331683A (en) | Facial expression recognition method with noise robust | |
| CN103593652A (en) | Cucumber disease identification method based on cucumber leaf symptom image processing | |
| CN103793709A (en) | Cell recognition method and device, and urine analyzer | |
| CN105654099A (en) | Sugarcane segmentation and identification method based on improved vision | |
| CN106683098A (en) | Segmentation method of overlapping leaf images | |
| CN103207989B (en) | A kind of vein recognition methods of the slight insect pest blade based on principal component analysis | |
| CN106529535A (en) | License plate positioning method in complex lighting environment based on wavelet transform | |
| CN103971369B (en) | The optic disk localization method of retinal images | |
| CN108182399A (en) | Refer to vein characteristic comparison method, apparatus, storage medium and processor | |
| CN107784639A (en) | A kind of polygon filtering and noise reduction method of unmanned aerial vehicle remote sensing image improvement | |
| da Silva Oliveira et al. | Automatic counting of stomata in epidermis microscopic images | |
| CN105930811A (en) | Palm texture feature detection method based on image processing | |
| CN104809733A (en) | Ancient building wall polluted inscription character image edge extraction method | |
| CN103226695A (en) | Multi-scale intelligent Chinese red pine identification model based on selective vision attention mechanism | |
| CN106960422B (en) | A kind of filtering method of the discontinuous fringe figure based on orientation consistency | |
| CN103886291B (en) | Palmprint image wrinkle detection method based on dynamic template | |
| Patil et al. | Identification of growth rate of plant based on leaf features using digital image processing techniques |
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: 20130710 |