CN105261045A - Digital method for rapidly evaluating loss severity of alfalfa diseases and insects - Google Patents

Abstract

The invention discloses a digital method for rapidly evaluating the loss severity of alfalfa diseases and insects. According to the technical scheme of the method, firstly, the focal regions and the non-focal regions of alfalfa leaves are analyzed and then the common disease degree and the RGB color component law thereof are generalized. After that, a color decision-making tree component system for the focal regions and the non-focal regions of alfalfa leaves is established accordingly. The visual processing flow of digital pictures is integrated as one, so that the method is suitable for non-computer professionals to use. At the same time, the functions of batch processing, data exporting and the like are realized, so that the method can be used for processing volume data. Based on the geometric projection imaging rule, the inscribed circle calculation manner is adopted, so that the geometric deformation avoided. The calculation accuracy is improved. Moreover, binary pictures for the healthy regions and the non-healthy regions of alfalfa leaves are constructed, and then the pictures are displayed and compared. In this way, the method can be applied to alfalfa leaves of smaller than 0.5 m in both length and width, with the diameters of the focal regions thereof to be larger than 1mm, and the area statistical precision of the focal regions thereof to be larger than 95%.

Description

A kind of digitizing solution of the rapid evaluation alfalfa disease and pest extent of damage

Technical field

The invention belongs to grass disease, digitized fine grass cultivation that pest injurious loss estimating and measuring method combines with digital imagery and computer technology and agricultural technology field, be specifically related to a kind of digitizing solution of the rapid evaluation alfalfa disease and pest extent of damage.

Background technology

China is one of large agricultural country in the world, but crops are but faced with various agricultural disaster.Diseases and pests of agronomic crop is one of Main Agricultural disaster of China, and it has that kind is many, impact is large and feature that is population outbreak often, and its occurrence scope and the order of severity often cause heavy losses to Chinese national economy, particularly agricultural production.Going deep in recent years along with China's opening, exotic invasive disease pest and weed constantly increases, along with Planter industry structure adjustment, new quality variety is promoted, the enforcement of high-yield and high-efficiency facility cultivation measure and diversification plantation be extended to more disease, worm, grass, the plague of rats population outbreak create suitable condition, new disease pest and weed constantly occurs, the occurring and damage of diseases and pests of agronomic crop is serious ascendant trend, therefore the integrated control work carrying out crop diseases and pest crop smothering, for the important means of the yield and quality of raising crops, must cause enough attention.

The growth of green crops all depends on photosynthesis.Green plants carries out photosynthesis by leaf.Take off when green plants leaf color green and just mean that plant health is out of joint.So, when research plant growth situation, can't do without area and the color of the leaf will considering it.Especially when research crop yield, often weather will be concerned about, insects etc., to the infringement caused of crops blade, therefore calculate the area of leaf, number percent that leaf damaged area accounts for the leaf total area, thus comparatively accurately measure crop pest degree and be necessary.This is significant to estimation crop yield, minimizing pesticide dosage, production cost and environmental pollution, and promotes that preventing and treating efficiency, prevention effect and Controlling benefit improves, and is conducive to ensureing agricultural production, agricultural product quality and farmland ecological environment safety.

Animal husbandry is in recent years able to tremendous expansion, and the Forage grass industry that guarantee animal husbandry is greatly developed obtains the most attention of various circles of society.Clover is planted on a large scale as a kind of forage grass of high-quality, in the process that Alfalfa planting technology is constantly perfect, rarely have report about the quick estimating and measuring method of alfalfa disease and pest loss.In order to adapt to modernization, digitizing ALFALFA PRODUCTION present situation, be necessary that very much finding a kind of alfalfa disease and pest loses quick estimating and measuring method, to serve production.

Summary of the invention

For solving the problem, the invention provides a kind of digitizing solution of the rapid evaluation alfalfa disease and pest extent of damage.

For achieving the above object, the technical scheme that the present invention takes is:

A digitizing solution for the rapid evaluation alfalfa disease and pest extent of damage, comprises the steps:

S1, alfalfa-leaves focal zone and non-focal zone to be analyzed, summarize common degree of disease and RGB color component rule thereof, and construct blade focal zone and non-focal zone color decision tree component system accordingly;

S2, calculating blade disease rate

S21, calculate the capable and wide L row of whole figure length of a film H (be a row (column) with a pixel unit), then find that H is capable, the capable and L row of L row, H, calculate mean value avgR, avgG, avgB of three components (RGB) of this four row (column) respectively;

S22, three components of each for picture pixel and four mean values of respective components are made difference and make ratio with mean value again, as long as there is a ratio to be less than 0.15, just this pixel is become gray-scale pixels;

If in three of S23 pixel components the difference of largest component value and minimum component value be greater than 180 or the difference of second largest component value and minimum component value be greater than 160, then this pixel is background (background colour is white), is set to gray-scale pixels;

S24, find out the max-thresholds of three components of picture pixels:

S241, value according to RGB tri-components, scan each pixel one by one, add up the value number that its three components are corresponding;

S242, remember that all pixel numbers are N, order i calculates successively from 0 to 255 $\begin{array}{c}\{W_{X0}=\underset{j=0}{\overset{i}{\Σ}}{P}_X\[j\],W_{X1}=\underset{j=i}{\overset{255}{\Σ}}{P}_X\[j\],\\ U_{X0}=\underset{j=0}{\overset{i}{\Σ}}(j\×P_X\[j\]),U_{X1}=\underset{j=i}{\overset{255}{\Σ}}(j\×P_X\[j\]),\\ V_{X0}=\frac{U_{X0}}{W_{X0}},V_{X1}=\frac{U_{X1}}{W_{X1}},\\ T_X=(W_{X0}\×W_{X1}){(V_{X0}-V_{X1})}^2\}\end{array}$ The T more at every turn calculated _x, get its maximal value and be required max-thresholds, be respectively T _r, T _g, T _b;

S243, when the max-thresholds of three components is all greater than 100, illustrate background color close to white; If each component of pixel is all greater than its corresponding max-thresholds or pixel each component when being all greater than 130, then this pixel three components are made all to become white; If the blue color component value of pixel is greater than green component values, then this pixel three components are made all to become white; If the difference of maxima and minima is less than 30 in pixel three components, then this pixel three components are made all to become white; When being greater than 100 when the max-thresholds of three components is different, illustrate that background color keeps off white; If each component of pixel is all less than its corresponding max-thresholds or pixel each component when being all greater than 130, then this pixel three components are made all to become white; If the blue color component value of pixel is greater than green component values, then this pixel three components are made all to become white; If the difference of maxima and minima is less than 30 in pixel three components, then this pixel three components are made all to become white;

S25, picture carried out respectively from left to right, from right to left, from top to bottom, from top to bottom to four corrosion, concrete:

S251, scan whole blade from upper and lower, left and right four direction respectively, as downward scanning, scan each pixel from left to right successively, if this pixel is white, then continues to scan to the right, until scanned this row, then carry out next line scanning; If scanning certain point is not white, then scan num pixel downwards, note sum is the pixel of non-white in num pixel, if then this point is square boundary point;

S252, be scanned across the blade rectangle after background process successively from four direction up and down, if pixel is white, then goes out in the same coordinate of undressed former figure, its pixel is set to redness.Running into first is not white pixel, then stop this row (column) to scan.The background color of former figure is finally made to become redness;

S253, from top to bottom, from left to right successively the picture of above-mentioned process to be scanned, if pixel is not red, then predicate blade, and by the number of pixels of sum counter-blade; Meanwhile, if this pixel G > R and B > G, then this point is impaired loci, counts with num; If G < R & & G > B & & (G-R)≤10, this point is also impaired loci, counts with num; Finally use try to achieve blade injury rate;

S3, extraction scale mark, and calculate length and width, the area of blade;

S31, get part (1) district in part (2) district of below untreated picture Leaf rectangle or more, get all pixels in this district, obtain the mean value Ravg of three components of these pixels, Bavg, Gavg, and then each pixel is scanned, if this pixel meets ((R/Ravg < 0.2) or (1 < R/Ravg < 1.2)) and ((G/Gavg < 0.2) or (1 < G/Gavg < 1.2)) and ((B/Bavg < 0.2) or (1 < B/Bavg < 1.2)), this pixel is become white (background colour).If the minimum value of the component of the RGB of this pixel is greater than 110, also become white;

S32, this part to be corroded, the same with the method for step S25, then this part of picture is become gray-scale map;

S33, remove this part impurity, get lower 1/4 place of this part, scan this all pixels of 1/5, the pixel of non-background color is added up, if this part of non-background color is more than 0.5, then has impurity; Lower 1/32 place is become white, from lower 1/32 upwards individual element spot scan, all non-background colors is become background, until run into first background color;

S34, with the method for recurrence by all be not background color, the point be connected is classified as a set, get that maximum set, scanning from right to left must obtain straight line, if do not obtained, so just judge that scale mark is not or not this region, the continuation that upwards (downwards) scanning runs into non-background color from first pixel from the position obtaining straight line upwards scans and record upwards (downwards) and scans how many individual putting and be designated as N1, stop until running into background color, get second point again, same method obtains N2, .... until scanning complete strips straight line, then two maximum N are got, the distance of these two N just represents 1cm,

S35, how many pixels are had to calculate the area of each pixel and then obtain blade area according to 1 square centimeter.

Decision tree is a kind of instrument representing processing logic with binary tree figure, intuitively, clearly can express the logical requirements of processing.Be particularly suitable for that factor of judgment is fewer, the uncomplicated situation of logical combination relation.It provide and a kind ofly show the similar method that what can obtain under what conditions and be worth this rule-like.RGB is a kind of color standard of industry member, by obtaining color miscellaneous to the change of red (R), green (G), blue (B) three Color Channels and their superpositions each other, namely RGB represents the color of red, green, blue three passages, and the RGB component that its use RGB model is each pixel in image distributes the intensity level within the scope of a 0-255.Under RGB pattern, often kind of RGB composition all can use the value from 0 (black) to 255 (whites).

The present invention has following beneficial effect:

Integrate the visualization processing flow process of digital pictures, namely photo import, process, contrast, and friendly interface, easy and simple to handle, be applicable to Non-computer Majors human users and use; There is the function such as batch processing, statistical conversion simultaneously, be applicable to mass data process; Digital pictures form comprises jpg, tiff, tif and bmp general format, and it is imported by multi-dimensional matrix mode; According to geometric projection imaging law, adopt incircle account form, cut down geometry deformation, improve computational accuracy; Construct the bianry image in blade health and non-health district, and display and contrast are provided; The dll file of user VB, BC secondary development is provided, user is provided shielded .P file, prevents code from distorting and leak; Be applicable to the blade that blade length and width is all less than 0.5m, scope and precision are alfalfa-leaves focal zone diameter is the scope being greater than 1mm, and alfalfa-leaves lesion area statistical precision reaches more than 95%.

Accompanying drawing explanation

Fig. 1 is embodiment of the present invention Leaf focal zone and non-focal zone color decision tree component system schematic diagram.

Fig. 2 is the schematic diagram of caustic solution in the embodiment of the present invention.

Fig. 3 extracts scale mark in the embodiment of the present invention, and calculates length and width, the area schematic diagram of blade;

Fig. 4 is that in the embodiment of the present invention, the non-background color of recurrence method connects a some normalizing set schematic diagram.

Fig. 5 is that embodiment of the present invention alfalfa-leaves size of tumor results of measuring exports schematic diagram.

Fig. 6 is each high-level schematic functional block diagram of software kit of the present invention.

Embodiment

In order to make objects and advantages of the present invention clearly understand, below in conjunction with embodiment, the present invention is further elaborated.Should be appreciated that specific embodiment described herein only in order to explain the present invention, be not intended to limit the present invention.

Embodiment

S1, alfalfa-leaves focal zone and non-focal zone to be analyzed, summarize common degree of disease and RGB color component rule thereof, and construct blade focal zone and non-focal zone color decision tree component system accordingly; As shown in Figure 1.

S2, calculating blade disease rate

S21, calculate the capable and wide L row of whole figure length of a film H (be a row (column) with a pixel unit), then find that H is capable, the capable and L row of L row, H, calculate mean value avgR, avgG, avgB of three components (RGB) of this four row (column) respectively;

S22, three components of each for picture pixel and four mean values of respective components are made difference and make ratio with mean value again, as long as there is a ratio to be less than 0.15 (setting x as certain pixel), just this pixel is become gray-scale pixels, namely each value of pixel is identical;

If in three of S23 pixel components the difference of largest component value and minimum component value be greater than 180 or the difference of second largest component value and minimum component value be greater than 160, then this pixel is background (background colour is white), is set to gray-scale pixels;

S24, find out the max-thresholds of three components of picture pixels:

According to the value (0-255) of RGB tri-components, scan each pixel one by one, add up value (0-255) number corresponding to its three components (X [i]=n (X ∈ { R, G, B}, 0≤i≤255 and i ∈ Z), represent that X component pixel is that the pixel of i has n).

Remember that all pixel numbers are N, order i calculates successively from 0 to 255 $\begin{array}{c}\{W_{X0}=\underset{j=0}{\overset{i}{\&Sigma;}}{P}_X\&lsqb;j\&rsqb;,W_{X1}=\underset{j=i}{\overset{255}{\&Sigma;}}{P}_X\&lsqb;j\&rsqb;,\\ U_{X0}=\underset{j=0}{\overset{i}{\&Sigma;}}(j\&times;P_X\&lsqb;j\&rsqb;),U_{X1}=\underset{j=i}{\overset{255}{\&Sigma;}}(j\&times;P_X\&lsqb;j\&rsqb;),\\ V_{X0}=\frac{U_{X0}}{W_{X0}},V_{X1}=\frac{U_{X1}}{W_{X1}},\\ T_X=(W_{X0}\&times;W_{X1}){(V_{X0}-V_{X1})}^2\}\end{array}$ The T more at every turn calculated _x, get its maximal value and be required max-thresholds, be respectively T _r, T _g, T _b.

When the max-thresholds of three components is all greater than 100, illustrate that background color is close to white:

If each component of pixel is all greater than its corresponding max-thresholds or pixel each component when being all greater than 130, then this pixel three components are made all to become white (255).

If the blue color component value of pixel is greater than green component values, then this pixel three components are made all to become white (255).

If the difference of maxima and minima is less than 30 in pixel three components, then this pixel three components are made all to become white (255).

When being greater than 100 when the max-thresholds of three components is different, illustrate that background color keeps off white: if each component of pixel is all less than its corresponding max-thresholds or pixel each component when being all greater than 130, then make this pixel three components all become white (255).

If the blue color component value of pixel is greater than green component values, then this pixel three components are made all to become white (255).

If the difference of maxima and minima is less than 30 in pixel three components, then this pixel three components are made all to become white (255).

S25, as shown in Figure 2, carries out four corrosion respectively from left to right, from right to left, from top to bottom, from top to bottom to picture, concrete:

S251, scan whole blade from upper and lower, left and right four direction respectively, as downward scanning, scan each pixel from left to right successively, if this pixel is white, then continues to scan to the right, until scanned this row, then carry out next line scanning; If scanning certain point is not white, then scan num pixel downwards, note sum is the pixel of non-white in num pixel, if then this point is square boundary point;

S252, be scanned across the blade rectangle after background process successively from four direction up and down, if pixel is white, then goes out in the same coordinate of undressed former figure, its pixel is set to redness.Running into first is not white pixel, then stop this row (column) to scan.The background color of former figure is finally made to become redness;

S253, from top to bottom, from left to right successively the picture of above-mentioned process to be scanned, if pixel is not red, then predicate blade, and by the number of pixels of sum counter-blade; Meanwhile, if this pixel G > R and B > G, then this point is impaired loci, counts with num; If G < R & & G > B & & (G-R)≤10, this point is also impaired loci, counts with num; Finally use try to achieve blade injury rate;

S3, extraction scale mark, and calculate length and width, the area of blade;

S31, as shown in Figure 3, get part (1) district in part (2) district of below untreated picture Leaf rectangle or more, get all pixels in this district, obtain the mean value Ravg of three components of these pixels, Bavg, Gavg, and then each pixel is scanned, if this pixel meets ((R/Ravg < 0.2) or (1 < R/Ravg < 1.2)) and ((G/Gavg < 0.2) or (1 < G/Gavg < 1.2)) and ((B/Bavg < 0.2) or (1 < B/Bavg < 1.2)), this pixel is become white (background colour).If the minimum value of the component of the RGB of this pixel is greater than 110, also become white;

S32, this part to be corroded, the same with the method for step S25, then this part of picture is become gray-scale map (R=G=B=0.11 × B+0.59 × G+0.3 × R);

S33, remove this part impurity (the black circle as on Fig. 3), get lower 1/4 place of this part, scan this all pixels of 1/5, the pixel of non-background color is added up, if this part of non-background color is more than 0.5, then has impurity; Lower 1/32 place is become white, from lower 1/32 upwards individual element spot scan, all non-background colors is become background, until run into first background color;

S34, as shown in Figure 4, with the method for recurrence by all be not background color, the point be connected is classified as a set, get that maximum set, scanning from right to left must obtain straight line, if do not obtained, so just judge that scale mark is not or not this region, the continuation that upwards (downwards) scanning runs into non-background color from first pixel from the position obtaining straight line upwards scans and record upwards (downwards) and scans how many individual putting and be designated as N1, stop until running into background color, get second point again, same method obtains N2, .... until scanning complete strips straight line, then two maximum N are got, the distance of these two N just represents 1cm, (picture can be become binary map by this step, easier)

S35, how many pixels are had to calculate the area of each pixel and then obtain blade area according to 1 square centimeter.

The measuring and calculating of individual digital pictures of this method exports as shown in Figure 5, and the functional module of this method software kit as shown in Figure 6.

The above is only the preferred embodiment of the present invention; it should be pointed out that for those skilled in the art, under the premise without departing from the principles of the invention; can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.

Claims (1)

1. a digitizing solution for the rapid evaluation alfalfa disease and pest extent of damage, is characterized in that, comprises the steps:

S1, alfalfa-leaves focal zone and non-focal zone to be analyzed, summarize common degree of disease and RGB color component rule thereof, and construct blade focal zone and non-focal zone color decision tree component system accordingly;

S2, calculating blade disease rate

S21, calculate the capable and wide L row of whole figure length of a film H (be a row (column) with a pixel unit), then find that H is capable, the capable and L row of L row, H, calculate mean value avgR, avgG, avgB of three components (RGB) of this four row (column) respectively;

S22, three components of each for picture pixel and four mean values of respective components are made difference and make ratio with mean value again, as long as there is a ratio to be less than 0.15, just this pixel is become gray-scale pixels;

If in three of S23 pixel components the difference of largest component value and minimum component value be greater than 180 or the difference of second largest component value and minimum component value be greater than 160, then this pixel is background (background colour is white), is set to gray-scale pixels;

S24, find out the max-thresholds of three components of picture pixels:

S241, value according to RGB tri-components, scan each pixel one by one, add up the value number that its three components are corresponding;

S242, remember that all pixel numbers are N, order i calculates successively from 0 to 255

{ $\begin{array}{cc}{W}_{X0}=\underset{j=0}{\overset{j}{\&Sigma;}}{P}_X\&lsqb;j\&rsqb;,& W_{X1}=\underset{j=i}{\overset{255}{\&Sigma;}}{P}_X\&lsqb;j\&rsqb;\end{array},$

$\begin{array}{cc}{U}_{X0}=\underset{j=0}{\overset{j}{\&Sigma;}}(j\&times;P_X\&lsqb;j\&rsqb;),& U_{X1}=\underset{j=i}{\overset{255}{\&Sigma;}}(j\&times;P_X\&lsqb;j\&rsqb;)\end{array},$

$\begin{array}{cc}{V}_{X0}=\frac{U_{X0}}{W_{X0}},& V_{X1}=\frac{U_{X1}}{W_{X1}}\end{array},$

$T_X=(W_{X0}\&times;W_{X1}){(V_{X0}-V_{X1})}^2$ The T that more at every turn calculates _x, get its maximal value and be required max-thresholds, be respectively T _r, T _g, T _b;

S243, when the max-thresholds of three components is all greater than 100, illustrate background color close to white; If each component of pixel is all greater than its corresponding max-thresholds or pixel each component when being all greater than 130, then this pixel three components are made all to become white; If the blue color component value of pixel is greater than green component values, then this pixel three components are made all to become white; If the difference of maxima and minima is less than 30 in pixel three components, then this pixel three components are made all to become white; When being greater than 100 when the max-thresholds of three components is different, illustrate that background color keeps off white; If each component of pixel is all less than its corresponding max-thresholds or pixel each component when being all greater than 130, then this pixel three components are made all to become white; If the blue color component value of pixel is greater than green component values, then this pixel three components are made all to become white; If the difference of maxima and minima is less than 30 in pixel three components, then this pixel three components are made all to become white;

S25, picture carried out respectively from left to right, from right to left, from top to bottom, from top to bottom to four corrosion, concrete:

S251, scan whole blade from upper and lower, left and right four direction respectively, as downward scanning, scan each pixel from left to right successively, if this pixel is white, then continues to scan to the right, until scanned this row, then carry out next line scanning; If scanning certain point is not white, then scan num pixel downwards, note sum is the pixel of non-white in num pixel, if then this point is square boundary point;

S252, be scanned across the blade rectangle after background process successively from four direction up and down, if pixel is white, then goes out in the same coordinate of undressed former figure, its pixel is set to redness.Running into first is not white pixel, then stop this row (column) to scan.The background color of former figure is finally made to become redness;

S253, from top to bottom, from left to right successively the picture of above-mentioned process to be scanned, if pixel is not red, then predicate blade, and by the number of pixels of sum counter-blade; Meanwhile, if this pixel G > R and B > G, then this point is impaired loci, counts with num; If G<R & & G>B & & (G-R)≤10, this point is also impaired loci, counts with num; Finally use try to achieve blade injury rate;

S3, extraction scale mark, and calculate length and width, the area of blade;

S31, get part (1) district in part (2) district of below untreated picture Leaf rectangle or more, get all pixels in this district, obtain the mean value Ravg of three components of these pixels, Bavg, Gavg, and then each pixel is scanned, if this pixel meets ((R/Ravg < 0.2) or (1 < R/Ravg < 1.2)) and ((G/Gavg < 0.2) or (1 < G/Gavg < 1.2)) and ((B/Bavg < 0.2) or (1 < B/Bavg < 1.2)), this pixel is become white (background colour).If the minimum value of the component of the RGB of this pixel is greater than 110, also become white;

S32, this part to be corroded, the same with the method for step S25, then this part of picture is become gray-scale map;

S33, remove this part impurity, get lower 1/4 place of this part, scan this all pixels of 1/5, the pixel of non-background color is added up, if this part of non-background color is more than 0.5, then has impurity; Lower 1/32 place is become white, from lower 1/32 upwards individual element spot scan, all non-background colors is become background, until run into first background color;

S34, with the method for recurrence by all be not background color, the point be connected is classified as a set, get that maximum set, scanning from right to left must obtain straight line, if do not obtained, so just judge that scale mark is not or not this region, the continuation that upwards (downwards) scanning runs into non-background color from first pixel from the position obtaining straight line upwards scans and record upwards (downwards) and scans how many individual putting and be designated as N1, stop until running into background color, get second point again, same method obtains N2, .... until scanning complete strips straight line, then two maximum N are got, the distance of these two N just represents 1cm,

S35, how many pixels are had to calculate the area of each pixel and then obtain blade area according to 1 square centimeter.