CN117893541A - Fruit tree leaf mosaic analysis method based on edge detection - Google Patents
Fruit tree leaf mosaic analysis method based on edge detection Download PDFInfo
- Publication number
- CN117893541A CN117893541A CN202410303304.9A CN202410303304A CN117893541A CN 117893541 A CN117893541 A CN 117893541A CN 202410303304 A CN202410303304 A CN 202410303304A CN 117893541 A CN117893541 A CN 117893541A
- Authority
- CN
- China
- Prior art keywords
- leaf
- edge
- acquiring
- curve
- disease
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 235000013399 edible fruits Nutrition 0.000 title claims abstract description 50
- 238000003708 edge detection Methods 0.000 title claims abstract description 28
- 238000004458 analytical method Methods 0.000 title claims abstract description 11
- 201000010099 disease Diseases 0.000 claims abstract description 84
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 claims abstract description 84
- 238000001514 detection method Methods 0.000 claims abstract description 9
- 238000004891 communication Methods 0.000 claims abstract description 7
- 238000000034 method Methods 0.000 claims description 41
- 230000005484 gravity Effects 0.000 claims description 37
- 238000005452 bending Methods 0.000 claims description 7
- 238000013528 artificial neural network Methods 0.000 description 4
- 230000001788 irregular Effects 0.000 description 4
- 241000700605 Viruses Species 0.000 description 2
- 208000015181 infectious disease Diseases 0.000 description 2
- 230000003746 surface roughness Effects 0.000 description 2
- 210000003462 vein Anatomy 0.000 description 2
- 238000005299 abrasion Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000002146 bilateral effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000003306 harvesting Methods 0.000 description 1
- 230000002458 infectious effect Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000010223 real-time analysis Methods 0.000 description 1
- 238000011897 real-time detection Methods 0.000 description 1
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention relates to the technical field of image data processing, and provides a fruit tree leaf mosaic analysis method based on edge detection, which comprises the following steps: acquiring a leaf image of the fruit tree; acquiring a main edge, a secondary edge and a leaf contour edge for the fruit leaf part image; acquiring a region and a plurality of connected regions, acquiring a plurality of matching straight lines of the connected regions, acquiring the surface flatness of blades of the matching straight lines according to the gray value of each pixel point on the matching straight lines, and acquiring the bifurcation tendency tortuosity of blades of the connected regions according to the slope, the length and the surface flatness of the blades of the matching straight lines; acquiring a root point, acquiring a disease coefficient of a region according to the distance between the root point and all the communication regions, the angle of a deflection angle and the bifurcation trend tortuosity of the leaf and stem, and acquiring the leaf disease degree of the leaf according to the disease coefficient of the region; and finishing the analysis and detection of the mosaic disease according to the disease degree of the leaf blade. The invention considers a plurality of characteristics to improve the detection precision of the mosaic disease.
Description
Technical Field
The invention relates to the technical field of image data processing, in particular to a fruit tree leaf mosaic analysis method based on edge detection.
Background
The planted products can be affected variously due to the difference of climate and geographical location. Among them, besides natural disasters, diseases are extremely harmful to crops. If no intervention measures are taken in time for diseases, the leaves and fruits fall slightly, the whole plant dies seriously, and even the harvest in the current year is influenced. Therefore, the importance of real-time analysis and detection of crop disease conditions is not small.
The fruit trees have more yield in China, and diseases are frequently generated, so that the yield of the fruits is greatly reduced. Among these diseases, mosaic disease is a worldwide virus disease, which is an infectious emergency disease that can be caused by infection of surrounding plants by leaf abrasion and aphid-borne viruses if measures are not taken in time.
At present, when the leaf mosaic condition of the fruit tree is analyzed and detected by using an image processing method, the leaf mosaic condition is easily influenced by background factors, and the leaf mosaic condition is similar to other interference objects, so that the detection precision of the mosaic is lower.
Disclosure of Invention
The invention provides an edge detection-based analysis method for a mosaic disease of a fruit tree leaf, which aims to solve the problem of lower detection precision of the mosaic disease, and adopts the following technical scheme:
the embodiment of the invention provides a method for analyzing a mosaic disease of a fruit tree leaf based on edge detection, which comprises the following steps:
Acquiring a leaf image of the fruit tree;
obtaining a main edge, a secondary edge and a leaf contour edge of the fruit tree leaf by using edge detection; acquiring a first area and a second area according to a main edge, acquiring a plurality of connected areas for each area, acquiring a first curve and a second curve of each connected area by analyzing the acquired edges, acquiring a matching straight line according to the first curve and the second curve, acquiring the blade surface flatness of the matching straight line according to the gray value of each pixel point on the matching straight line and the difference between the maximum gray value and the minimum gray value, and acquiring the blade bifurcation trend tortuosity of the connected areas according to the slope, the length and the blade surface flatness of all the matching straight lines;
Obtaining root points on the edges of the leaf profile, obtaining a disease coefficient of each region according to the distance between the root points and all the communicating regions, the angle of the offset angle between the communicating regions and the leaf stem bifurcation tendency tortuosity of the communicating regions, and obtaining leaf disease degree of the leaf according to the disease coefficients of the first region and the second region;
And finishing the analysis and detection of the mosaic disease according to the leaf disease degree of the leaves.
Preferably, the method for obtaining the main edge, the secondary edge and the leaf contour edge comprises the following steps:
And (3) performing edge detection by using a Canny operator, marking edge lines at the outermost periphery in the fruit tree leaf part image as leaf contour edges, marking edge lines of two endpoints on the leaf contour edges as connecting edge lines, marking the longest connecting edge lines as main edges, and marking edge lines connecting the main edges and the leaf contour edges as secondary edges.
Preferably, the method for acquiring the first area and the second area according to the main edge includes:
the area inside the leaf contour edge is marked as a leaf area, the main edge divides the leaf area into two parts, one part with large area is a first area, and the other part with small area is a second area.
Preferably, the method for obtaining the first curve and the second curve of each connected domain by analyzing the obtained edges includes:
The method comprises the steps of firstly removing leaf contour edges from radicals in a communication domain, wherein if only two edges exist in the communication domain, the two edges are used as a first curve and a second curve, and if a plurality of edges exist in the communication domain, the main edges are removed, the remaining two secondary edges are used as the first curve and the second curve, and the edges only comprise the main edges and the secondary edges.
Preferably, the method for obtaining the matching straight line according to the first curve and the second curve includes:
And matching the first curve with the second curve by using a DTW algorithm, wherein each pixel point in the first curve and the second curve have at least one matching point, the pixel points of the first curve and the matched pixel points form a matching sequence pair, and the connecting line of the two corresponding pixel points of the matching sequence pair is marked as a matching straight line.
Preferably, the method for obtaining the surface flatness of the blade of the matching straight line according to the gray value of each pixel point on the matching straight line and the difference between the maximum gray value and the minimum gray value comprises the following steps:
The gray value of each pixel point on the matching straight line is obtained, the maximum gray value and the minimum gray value are selected, and the difference between the maximum gray value and the minimum gray value is used as the gray change reference of the matching straight line;
Obtaining the gray value average value of all pixel points on the matching straight line, and summing the difference between the gray values of all pixel points and the gray value average value to obtain the gray difference coefficient of the matching straight line;
and taking the gray level change reference of the matching straight line as a base number and taking the negative number of the gray level difference coefficient as an index to obtain the surface flatness of the blade of the matching straight line.
Preferably, the method for obtaining the bifurcation of the stems and the leaves of the connected domain according to the slope, the length and the surface flatness of the blades of all the matching straight lines comprises the following steps:
Where denotes the length of the jth matching straight line,/> denotes the length average of all the matching straight lines,/> denotes the slope of the end point of the jth matching straight line in the first curve,/> denotes the slope of the end point of the jth matching straight line in the second curve,/> denotes the blade surface flatness of the jth matching straight line,/> denotes the number of matching straight lines in the connected domain,/> denotes the blade bifurcation tendency tortuosity in the connected domain.
Preferably, the root point obtaining method includes:
And acquiring the intersection point of the secondary edge and the leaf edge profile, calculating the Euclidean distance between the end points at the two ends of the main edge and the intersection point respectively, acquiring the maximum Euclidean distance, and marking the end point of the main edge corresponding to the maximum Euclidean distance as a root point.
Preferably, the method for obtaining the disease coefficient of each region according to the distance between the root point and all the communicating regions, the angle of the offset angle between the communicating regions and the leaf-stem bifurcation tendency tortuosity of the communicating regions comprises the following steps:
The method comprises the steps of obtaining the gravity center of each communicating domain, connecting root points with the gravity centers of the communicating domains to obtain gravity center straight lines, wherein the length of each gravity center straight line is the gravity center distance, the slope of each two adjacent gravity center straight lines is differenced to obtain the angle of the offset angle of the corresponding two communicating domains, the length of each two adjacent gravity center straight lines is differenced to obtain the gravity center distance difference between the communicating domains, the mean value of the offset angles of all the communicating domains and the mean value of the gravity center distance difference are obtained, and the disease coefficient of the area is obtained according to the offset angle between all the adjacent communicating domains, the difference of the gravity center distance and the mean value and the leaf-stalk bifurcation tendency tortuosity.
Preferably, the method for obtaining the disease coefficient of the area according to the deviation angle, the gravity center distance and the mean value difference between all adjacent connected areas comprises the following steps:
Wherein denotes an angle of an offset angle between the r-1 th connected domain and the r-1 th connected domain,/> denotes an average value of all the offset angles,/> denotes a difference in center of gravity distances between the r-1 th connected domain and the r-1 th connected domain,/> denotes an average value of center of gravity distances between all the connected domains,/> denotes a leaf-stem bifurcation tendency tortuosity of the r-th connected domain,/> denotes the number of connected domains, and/> denotes a disease coefficient.
The beneficial effects of the invention are as follows: according to the method, different connected domains are extracted through edge detection on the images of the fruit tree leaves, the leaf stem bifurcation tendency tortuosity of each connected domain is obtained according to the gray level characteristic and the curve characteristic of each connected domain, the leaf stem bifurcation tendency tortuosity reflects the bending degree of the leaves, the greater the bending degree is, the more likely the leaves have mosaic disease, the leaf disease degree is obtained through comparing disease differences on the left side and the right side of the leaves, the interference of other objects is eliminated according to the image characteristic of the leaves by the mosaic disease through a pushing analysis method, and the detection accuracy of the mosaic disease is greatly improved.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions of the prior art, the drawings which are used in the description of the embodiments or the prior art will be briefly described, it being obvious that the drawings in the description below are only some embodiments of the invention, and that other drawings can be obtained according to these drawings without inventive faculty for a person skilled in the art.
Fig. 1 is a schematic flow chart of a method for analyzing a mosaic disease of a fruit tree leaf based on edge detection according to an embodiment of the present invention;
Fig. 2 is a schematic diagram of leaf edge detection.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Referring to fig. 1, a flowchart of a method for analyzing a mosaic disease of a fruit tree leaf based on edge detection according to an embodiment of the present invention is shown, where the method includes the following steps:
Step S001, acquiring a leaf image of the fruit.
The method comprises the steps of flatly placing fruit tree leaves on a tabletop, placing a light source on the side face above the fruit tree leaves, shooting an image right above the fruit tree leaves through a CCD camera, acquiring an RGB image, acquiring a gray image by using weighted graying conversion of the acquired RGB image, denoising the image by using a bilateral filtering algorithm, and eliminating the influence of noise to acquire a fruit tree leaf image.
So far, the image of the leaves of the fruit tree is obtained.
Step S002, obtaining a main edge, a secondary edge and a leaf contour edge by using edge detection for the fruit leaf part image; acquiring a first area and a second area according to the main edge, acquiring a plurality of connected areas for each area, acquiring a first curve and a second curve of each connected area by analyzing the obtained edges, acquiring a matching straight line according to the first curve and the second curve, acquiring the blade surface flatness of the matching straight line according to the gray value of each pixel point on the matching straight line and the difference between the maximum gray value and the minimum gray value, and acquiring the blade bifurcation trend tortuosity of the connected areas according to the slope, the length and the blade surface flatness of all the matching straight lines.
When the mosaic disease occurs on the leaves of the fruit trees, the surfaces of the leaves of the fruit trees are deformed in a concave-convex manner, and uneven distribution of the colors of the leaves is shown in images of the leaves of the fruit trees; meanwhile, some serious mosaic diseases can cause edge curling of leaf edges, and the edge is in a concave-convex smooth shape in a fruit tree leaf image instead of a smooth arc under normal conditions; and the bifurcation of the leaf stems may become uneven.
Therefore, edge detection is performed on the fruit tree leaf image by using a Canny operator, the edge line at the outermost periphery in the fruit tree leaf image is marked as a leaf contour edge, a main leaf stem exists in the fruit tree leaf, the main leaf stem is arranged inside the leaf contour edge, two endpoints of the edge line corresponding to the main leaf stem are arranged on the leaf contour edge, the edge line with the largest length of the edge lines of the two endpoints on the leaf contour edge line is the main edge, namely the edge corresponding to the main leaf stem is the main edge, the edge line connecting the main edge and the leaf contour edge is marked as the secondary edge, and the obtained edge detection result is shown in fig. 2.
The edge detection image at this time is analyzed by using the connected domain to obtain a plurality of connected domains, the main edge divides the fruit tree leaf into two parts, namely a first area and a second area, wherein the area of the first area is larger than that of the second area, and the operation methods of the first area and the second area are the same, so the first area is taken as an example for description of the embodiment.
In each connected domain, as the concave-convex deformation of the leaf surface is caused by the disease, uneven distribution of leaf color is shown in the leaf image of the fruit tree, namely, the gray value of pixel points in one connected domain also has larger difference, for each acquired connected domain, the edge lines belonging to the leaf outline edge and the main edge in the outline edge of each connected domain are eliminated, the left side edge and the right side edge of each connected domain are taken as the remaining two edge lines, if the leaf outline edge and the main edge are eliminated in the outline edge of the connected domain, the leaf outline edge is only eliminated, and the left side edge and the right side edge are formed by the main edge and the secondary edge.
And marking the left side edge and the right side edge as a first curve and a second curve, matching by using a DTW algorithm to obtain matched pixel points of each pixel point on the first curve and the second curve, forming a matching sequence pair by the pixel points and the matched pixel points, and analyzing gray level change by analyzing all the matching sequences in the connected domain in order to represent the concave-convex change degree of the leaf in the connected domain.
And connecting each matching sequence pair corresponding to two pixel points in the connected domain to obtain a matching straight line, acquiring the gray value of each pixel point in the matching straight line, selecting the maximum gray value and the minimum gray value on the matching straight line, and making the difference between the acquired maximum gray value and the minimum gray value to obtain the gray change reference coefficient of the matching straight line. The larger the gradation change reference coefficient is, the larger the gradation change on the straight line is, the more the leaf unevenness is likely to appear, and the more serious the leaf unevenness phenomenon is, the more serious the mosaic is.
For each matching straight line, calculating the gray average value of all pixel points on the matching straight line, and acquiring the gray difference coefficient of the matching straight line according to the gray value of each pixel point on the matching straight line and the gray average value of the pixel points on the matching straight line, wherein the formula is as follows:
Where denotes the gray value of the i-th pixel on the matching line,/> denotes the gray average value of the pixels on the matching line,/> denotes the number of pixels on the matching line, and/> denotes the gray gap coefficient of the matching line. The larger the sum of the difference between the gray value of each pixel point on the straight line and the gray average value is, the larger the gray value variation and the floating of the pixel points on the matched straight line are, the more the gray value variation and the floating of the pixel points are, the more the convex-concave phenomenon of the leaves is serious, and the more the mosaic is serious.
Constructing the blade surface flatness of each matching straight line according to the obtained gray level change reference coefficient and gray level difference coefficient of each matching straight line, wherein the formula is as follows:
Where denotes a gradation change reference coefficient of the matching straight line,/> denotes a gradation gap coefficient of the matching straight line, and/> denotes a blade surface flatness of each matching straight line. Blade surface flatness/> represents a mosaic disease feature that matches straight lines exhibit concave-convex deformation. The smaller/> , the more severe the mosaic at this line.
If mosaic disease occurs, not only the surface roughness of the blade is different in the communicating region, but also the distance and the slope of the corresponding matching pair between the two edges forming the communicating region and the irregularity thereof, and meanwhile, the blade flatness, namely the surface roughness of the blade, also influences the distribution among the branches of the blade and the stem.
Acquiring the length of each matching straight line in the connected domain and the Euclidean distance between two corresponding pixel points of the matching sequence pairs, acquiring the average value of the lengths of all the matching straight lines in the connected domain, acquiring the slope of each pixel point on the first curve and the second curve, and acquiring the bifurcation and the bending degree of the leaf stem in the connected domain according to the slope difference of the two corresponding pixel points of all the matching sequence pairs, the length of the matching straight line corresponding to the matching sequence pairs and the surface flatness of the leaf, wherein the formula is as follows:
Where denotes the length of the jth matching straight line,/> denotes the length average of all the matching straight lines,/> denotes the slope of the end point of the jth matching straight line in the first curve,/> denotes the slope of the end point of the jth matching straight line in the second curve,/> denotes the blade surface flatness of the jth matching straight line,/> denotes the number of matching straight lines in the connected domain,/> denotes the blade bifurcation tendency tortuosity in the connected domain.
The smaller the surface flatness of the blade is, the more obvious the concave-convex phenomenon is, and the concave-convex phenomenon can lead to the deformation of the communicating region, so that the smaller the surface flatness of the blade is, the more irregular the communicating region is, the larger the length change of the matching straight line is, the more irregular the communicating region is, the larger the gradient difference of pixel points at two ends of the matching straight line is, the larger the difference of bending degrees of two edges is, the more irregular the communicating region is, and the greater the bending degree of the bifurcation trend of the blade stem of the communicating region is.
Thus, the leaf and stem bifurcation tendency tortuosity of each connected domain is obtained.
Step S003, obtaining root points on the edges of the leaf profile, obtaining the disease coefficient of each region according to the distance between the root points and all the communicating regions, the angle of the offset angle between the communicating regions and the leaf stem bifurcation tendency tortuosity of the communicating regions, and obtaining the leaf disease degree of the leaf according to the disease coefficients of the first region and the second region.
If mosaic disease occurs, larger difference exists between adjacent connected domains; because the leaf basically presents a shape with a narrow upper part and a wide lower part, and the veins in the leaf can be shortened slowly from the bottom to the head, the pixel points of the root of the leaf stem are obtained based on the veins; acquiring all secondary edges in the images of the leaves of the fruits, acquiring intersection points of the secondary edges and the outlines of the leaves, calculating Euclidean distances between the end points of the two ends of the leaves and the intersection points respectively, and acquiring maximum distances, wherein the end points of the leaves and the stems corresponding to the maximum distances are the pixel points of the roots of the leaves and the stems and are recorded as root points.
Acquiring the gravity center points of all the communicating domains in the fruit tree leaf image, connecting the root point with the gravity center point of each communicating domain to obtain gravity center straight lines, acquiring the gravity center distances of the root point and the gravity center point of each communicating domain, recording the included angles on the adjacent two gravity center straight lines as offset angles, acquiring the angle of each offset angle, obtaining the average value of the angles of all the offset angles by taking the difference of slopes of the two gravity center straight lines forming the offset angles, and acquiring disease coefficients according to the difference of the offset angles of the adjacent communicating domains, the gravity center distances of the communicating domains and the difference of the branching tendency tortuosity of the leaf stems of the communicating domains, wherein the formula is as follows:
Wherein denotes an angle of an offset angle between the r-1 th connected domain and the r-1 th connected domain,/> denotes an average value of all the offset angles,/> denotes a difference in center of gravity distances between the r-1 th connected domain and the r-1 th connected domain,/> denotes an average value of center of gravity distances between all the connected domains,/> denotes a leaf-stem bifurcation tendency tortuosity of the r-th connected domain,/> denotes the number of connected domains, and/> denotes a disease coefficient.
If the deviation angle between two connected domains and the average value of the deviation angle are larger, it is indicated that the position change exists between the two connected domains and possibly caused by the concave-convex of the leaves, similarly, if the difference of the center of gravity distance between the two connected domains and the average value are too large, it is indicated that the uneven distribution of the two connected domains is possibly caused by the concave-convex of the leaves, and the bending degree of the bifurcation of the stems of the connected domains reflects the shape rule of the connected domains, if the value is larger, it is indicated that the shape of the connected domains is more irregular, the concave-convex of the leaves is more likely to appear, and therefore, the formula polynomial is larger and mosaic is more likely to exist.
The disease coefficient of the first area is obtained, the disease coefficient of the second area is obtained in the same way, and the disease degree of the blade is obtained according to the disease coefficients of the first area and the second area, wherein the formula is as follows:
Wherein denotes a disease coefficient of the first region,/> denotes a disease coefficient of the second region, and/> denotes a leaf disease degree.
The difference between the left and right leaves is characterized by calculation to influence the disease degree of the whole leaf. The larger the leaf, the larger the average disease degree of the left and right leaves, and the larger the difference between the disease degrees of the left and right leaves, the larger the disease degree of the leaf mosaic disease can be indicated.
So far, the leaf disease degree of the leaves is obtained.
And S004, completing the analysis and detection of the mosaic disease according to the leaf disease degree of the leaves.
The method comprises the steps of obtaining a large amount of training data, using a VGG neural network, marking leaves artificially, marking the leaves with mosaic disease as 1, marking the leaves without mosaic disease as 0, inputting the leaves with mosaic disease degree into the neural network, outputting the leaves with mosaic disease degree or not, obtaining the leaves with disease degree of each leaf after training the neural network, and inputting the leaves with mosaic disease degree into the neural network to finish analysis and detection of the mosaic disease of the leaves of the fruit tree.
It should be noted that: the sequence of the embodiments of the present invention is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments.
Claims (10)
1. The method for analyzing the mosaic disease of the fruit tree leaf based on the edge detection is characterized by comprising the following steps of:
Acquiring a leaf image of the fruit tree;
obtaining a main edge, a secondary edge and a leaf contour edge of the fruit tree leaf by using edge detection; acquiring a first area and a second area according to a main edge, acquiring a plurality of connected areas for each area, acquiring a first curve and a second curve of each connected area by analyzing the acquired edges, acquiring a matching straight line according to the first curve and the second curve, acquiring the blade surface flatness of the matching straight line according to the gray value of each pixel point on the matching straight line and the difference between the maximum gray value and the minimum gray value, and acquiring the blade bifurcation trend tortuosity of the connected areas according to the slope, the length and the blade surface flatness of all the matching straight lines;
Obtaining root points on the edges of the leaf profile, obtaining a disease coefficient of each region according to the distance between the root points and all the communicating regions, the angle of the offset angle between the communicating regions and the leaf stem bifurcation tendency tortuosity of the communicating regions, and obtaining leaf disease degree of the leaf according to the disease coefficients of the first region and the second region;
And finishing the analysis and detection of the mosaic disease according to the leaf disease degree of the leaves.
2. The method for analyzing the leaf mosaic of the fruit tree based on the edge detection according to claim 1, wherein the method for acquiring the main edge, the secondary edge and the leaf contour edge is as follows:
And (3) performing edge detection by using a Canny operator, marking edge lines at the outermost periphery in the fruit tree leaf part image as leaf contour edges, marking edge lines of two endpoints on the leaf contour edges as connecting edge lines, marking the longest connecting edge lines as main edges, and marking edge lines connecting the main edges and the leaf contour edges as secondary edges.
3. The method for analyzing the mosaic disease of the fruit tree leaf based on the edge detection according to claim 1, wherein the method for acquiring the first area and the second area according to the main edge is as follows:
the area inside the leaf contour edge is marked as a leaf area, the main edge divides the leaf area into two parts, one part with large area is a first area, and the other part with small area is a second area.
4. The method for analyzing the mosaic disease of the fruit tree leaf based on the edge detection according to claim 1, wherein the method for obtaining the first curve and the second curve of each connected domain by analyzing the obtained edge is as follows:
The method comprises the steps of firstly removing leaf contour edges from radicals in a communication domain, wherein if only two edges exist in the communication domain, the two edges are used as a first curve and a second curve, and if a plurality of edges exist in the communication domain, the main edges are removed, the remaining two secondary edges are used as the first curve and the second curve, and the edges only comprise the main edges and the secondary edges.
5. The method for analyzing the mosaic disease of the fruit tree leaf based on the edge detection according to claim 1, wherein the method for obtaining the matching straight line according to the first curve and the second curve is as follows:
And matching the first curve with the second curve by using a DTW algorithm, wherein each pixel point in the first curve and the second curve have at least one matching point, the pixel points of the first curve and the matched pixel points form a matching sequence pair, and the connecting line of the two corresponding pixel points of the matching sequence pair is marked as a matching straight line.
6. The method for analyzing the mosaic disease of the fruit tree leaf based on the edge detection according to claim 1, wherein the method for obtaining the surface flatness of the leaf of the matching straight line according to the gray value of each pixel point on the matching straight line and the difference between the maximum gray value and the minimum gray value is as follows:
The gray value of each pixel point on the matching straight line is obtained, the maximum gray value and the minimum gray value are selected, and the difference between the maximum gray value and the minimum gray value is used as the gray change reference of the matching straight line;
Obtaining the gray value average value of all pixel points on the matching straight line, and summing the difference between the gray values of all pixel points and the gray value average value to obtain the gray difference coefficient of the matching straight line;
and taking the gray level change reference of the matching straight line as a base number and taking the negative number of the gray level difference coefficient as an index to obtain the surface flatness of the blade of the matching straight line.
7. The method for analyzing the leaf mosaic disease of the fruit tree based on the edge detection according to claim 1, wherein the method for acquiring the bifurcation and the bending degree of the leaves and stems of the connected domain according to the slope, the length and the surface flatness of the leaves of all the matching straight lines is as follows:
Where denotes the length of the jth matching straight line,/> denotes the length average of all the matching straight lines,/> denotes the slope of the end point of the jth matching straight line in the first curve,/> denotes the slope of the end point of the jth matching straight line in the second curve,/> denotes the blade surface flatness of the jth matching straight line,/> denotes the number of matching straight lines in the connected domain,/> denotes the blade bifurcation tendency tortuosity in the connected domain.
8. The method for analyzing the mosaic disease of the fruit tree leaf based on the edge detection according to claim 1, wherein the root point obtaining method is as follows:
And acquiring the intersection point of the secondary edge and the leaf edge profile, calculating the Euclidean distance between the end points at the two ends of the main edge and the intersection point respectively, acquiring the maximum Euclidean distance, and marking the end point of the main edge corresponding to the maximum Euclidean distance as a root point.
9. The method for analyzing the leaf mosaic disease of the fruit tree based on the edge detection according to claim 1, wherein the method for acquiring the disease coefficient of each region according to the distance between the root point and all the connected domains, the angle of the offset angle between the connected domains and the leaf and stem bifurcation tendency tortuosity of the connected domains is as follows:
The method comprises the steps of obtaining the gravity center of each communicating domain, connecting root points with the gravity centers of the communicating domains to obtain gravity center straight lines, wherein the length of each gravity center straight line is the gravity center distance, the slope of each two adjacent gravity center straight lines is differenced to obtain the angle of the offset angle of the corresponding two communicating domains, the length of each two adjacent gravity center straight lines is differenced to obtain the gravity center distance difference between the communicating domains, the mean value of the offset angles of all the communicating domains and the mean value of the gravity center distance difference are obtained, and the disease coefficient of the area is obtained according to the offset angle between all the adjacent communicating domains, the difference of the gravity center distance and the mean value and the leaf-stalk bifurcation tendency tortuosity.
10. The method for analyzing the leaf mosaic disease of the fruit tree based on the edge detection according to claim 9, wherein the method for acquiring the disease coefficient of the area according to the deviation angle, the difference between the center of gravity distance and the mean value and the leaf-stem bifurcation trend tortuosity between all adjacent connected domains is as follows:
Wherein denotes an angle of an offset angle between the r-1 th connected domain and the r-1 th connected domain,/> denotes an average value of all the offset angles,/> denotes a difference in center of gravity distances between the r-1 th connected domain and the r-1 th connected domain,/> denotes an average value of center of gravity distances between all the connected domains,/> denotes a leaf-stem bifurcation tendency tortuosity of the r-th connected domain,/> denotes the number of connected domains, and/> denotes a disease coefficient.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410303304.9A CN117893541B (en) | 2024-03-18 | 2024-03-18 | Fruit tree leaf mosaic analysis method based on edge detection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410303304.9A CN117893541B (en) | 2024-03-18 | 2024-03-18 | Fruit tree leaf mosaic analysis method based on edge detection |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117893541A true CN117893541A (en) | 2024-04-16 |
| CN117893541B CN117893541B (en) | 2024-05-28 |
Family
ID=90641684
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410303304.9A Active CN117893541B (en) | 2024-03-18 | 2024-03-18 | Fruit tree leaf mosaic analysis method based on edge detection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117893541B (en) |
Citations (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004272313A (en) * | 2003-03-05 | 2004-09-30 | Ricoh Co Ltd | Image analysis apparatus |
| US20130136312A1 (en) * | 2011-11-24 | 2013-05-30 | Shih-Mu TSENG | Method and system for recognizing plant diseases and recording medium |
| KR20160049117A (en) * | 2014-10-24 | 2016-05-09 | 주식회사 포스코 | Apparatus for detecting flaw depht and flaw shape of steel plate and detecting method using thereof |
| AU2020103131A4 (en) * | 2020-10-30 | 2021-01-07 | Nanjing Forestry University | Leaf surface reconstruction and physically based deformation simulation based on the point cloud data |
| CN113989689A (en) * | 2021-11-29 | 2022-01-28 | 沈阳农业大学 | Crop pest and disease identification method and system based on unmanned aerial vehicle |
| EP3989161A1 (en) * | 2020-10-23 | 2022-04-27 | Tata Consultancy Services Limited | Method and system for leaf age estimation based on morphological features extracted from segmented leaves |
| CN115908371A (en) * | 2022-12-14 | 2023-04-04 | 南京信息工程大学 | Plant leaf disease and insect pest degree detection method based on optimized segmentation |
| WO2023134789A1 (en) * | 2022-10-25 | 2023-07-20 | 苏州德斯米尔智能科技有限公司 | Automatic inspection method for belt-type conveying device |
| CN116563295A (en) * | 2023-07-12 | 2023-08-08 | 无锡康贝电子设备有限公司 | Visual detection method for cutting chip winding state |
| CN116563282A (en) * | 2023-07-10 | 2023-08-08 | 东莞市博思特数控机械有限公司 | Drilling tool detection method and system based on machine vision |
| CN116664557A (en) * | 2023-07-28 | 2023-08-29 | 无锡市明通动力工业有限公司 | Visual detection method for surface defects of fan blade |
| CN116740054A (en) * | 2023-08-08 | 2023-09-12 | 天筛(聊城)生物科技有限公司 | Tongue image tooth trace detection method based on image processing |
| CN116824516A (en) * | 2023-08-30 | 2023-09-29 | 中冶路桥建设有限公司 | Road construction safety monitoring and management system |
| CN116958572A (en) * | 2023-09-18 | 2023-10-27 | 济宁市林业保护和发展服务中心 | Leaf disease and pest area analysis method in fruit tree breeding |
| CN117152137A (en) * | 2023-10-30 | 2023-12-01 | 江苏高特高金属科技有限公司 | Welded pipe corrosion state detection method based on image processing |
| WO2024040856A1 (en) * | 2022-08-24 | 2024-02-29 | 广东拓斯达科技股份有限公司 | Defect detection method and apparatus, and electronic device and storage medium |
| CN117689659A (en) * | 2024-02-02 | 2024-03-12 | 深圳市未尔科技有限公司 | Production quality monitoring method based on flat electronic product |
-
2024
- 2024-03-18 CN CN202410303304.9A patent/CN117893541B/en active Active
Patent Citations (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004272313A (en) * | 2003-03-05 | 2004-09-30 | Ricoh Co Ltd | Image analysis apparatus |
| US20130136312A1 (en) * | 2011-11-24 | 2013-05-30 | Shih-Mu TSENG | Method and system for recognizing plant diseases and recording medium |
| KR20160049117A (en) * | 2014-10-24 | 2016-05-09 | 주식회사 포스코 | Apparatus for detecting flaw depht and flaw shape of steel plate and detecting method using thereof |
| EP3989161A1 (en) * | 2020-10-23 | 2022-04-27 | Tata Consultancy Services Limited | Method and system for leaf age estimation based on morphological features extracted from segmented leaves |
| AU2020103131A4 (en) * | 2020-10-30 | 2021-01-07 | Nanjing Forestry University | Leaf surface reconstruction and physically based deformation simulation based on the point cloud data |
| CN113989689A (en) * | 2021-11-29 | 2022-01-28 | 沈阳农业大学 | Crop pest and disease identification method and system based on unmanned aerial vehicle |
| WO2024040856A1 (en) * | 2022-08-24 | 2024-02-29 | 广东拓斯达科技股份有限公司 | Defect detection method and apparatus, and electronic device and storage medium |
| WO2023134789A1 (en) * | 2022-10-25 | 2023-07-20 | 苏州德斯米尔智能科技有限公司 | Automatic inspection method for belt-type conveying device |
| CN115908371A (en) * | 2022-12-14 | 2023-04-04 | 南京信息工程大学 | Plant leaf disease and insect pest degree detection method based on optimized segmentation |
| CN116563282A (en) * | 2023-07-10 | 2023-08-08 | 东莞市博思特数控机械有限公司 | Drilling tool detection method and system based on machine vision |
| CN116563295A (en) * | 2023-07-12 | 2023-08-08 | 无锡康贝电子设备有限公司 | Visual detection method for cutting chip winding state |
| CN116664557A (en) * | 2023-07-28 | 2023-08-29 | 无锡市明通动力工业有限公司 | Visual detection method for surface defects of fan blade |
| CN116740054A (en) * | 2023-08-08 | 2023-09-12 | 天筛(聊城)生物科技有限公司 | Tongue image tooth trace detection method based on image processing |
| CN116824516A (en) * | 2023-08-30 | 2023-09-29 | 中冶路桥建设有限公司 | Road construction safety monitoring and management system |
| CN116958572A (en) * | 2023-09-18 | 2023-10-27 | 济宁市林业保护和发展服务中心 | Leaf disease and pest area analysis method in fruit tree breeding |
| CN117152137A (en) * | 2023-10-30 | 2023-12-01 | 江苏高特高金属科技有限公司 | Welded pipe corrosion state detection method based on image processing |
| CN117689659A (en) * | 2024-02-02 | 2024-03-12 | 深圳市未尔科技有限公司 | Production quality monitoring method based on flat electronic product |
Non-Patent Citations (1)
| Title |
|---|
| 赵方;石晟;闫民;: "普通光照下叶片图像特征信息抽取", 计算机工程与应用, no. 05, 31 May 2015 (2015-05-31), pages 160 - 170 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117893541B (en) | 2024-05-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109308697B (en) | Leaf disease identification method based on machine learning algorithm | |
| Miri et al. | Retinal image analysis using curvelet transform and multistructure elements morphology by reconstruction | |
| CN110120042B (en) | Crop image pest and disease damage area extraction method based on SLIC super-pixel and automatic threshold segmentation | |
| CN110705468B (en) | Eye movement range identification method and system based on image analysis | |
| CN112435239A (en) | Scindapsus aureus leaf shape parameter estimation method based on MRE-PointNet and self-encoder model | |
| CN112907545B (en) | Seed bud length and root length detection method based on image processing | |
| CN113298777A (en) | Cotton leaf blight detection method and system based on color features and super-pixel clustering | |
| CN110751029B (en) | Maximum curvature-based adaptive finger vein line extraction method | |
| CN110827273A (en) | Tea disease detection method based on regional convolution neural network | |
| CN106683098B (en) | A kind of dividing method of overlapped blades image | |
| CN115082477B (en) | Semiconductor wafer processing quality detection method based on light reflection removing effect | |
| CN114049267A (en) | Improved neighborhood search based statistical and bilateral filtering point cloud denoising method | |
| CN111145205B (en) | Pig body temperature detection method based on infrared image under multiple pig scenes | |
| CN115908371A (en) | Plant leaf disease and insect pest degree detection method based on optimized segmentation | |
| CN110610438B (en) | Crop canopy petiole included angle calculation method and system | |
| CN116523898A (en) | Tobacco phenotype character extraction method based on three-dimensional point cloud | |
| CN117893541B (en) | Fruit tree leaf mosaic analysis method based on edge detection | |
| CN109871827B (en) | Footprint expression method combining region confidence coefficient and pressure distribution direction intensity | |
| CN111291686A (en) | Method and system for extracting crop root phenotype parameters and judging root phenotype | |
| CN117315289B (en) | Aeroengine blade contour edge detection method based on image processing | |
| CN111369497B (en) | Walking type tree fruit continuous counting method and device | |
| CN115937151B (en) | Method for judging curling degree of crop leaves | |
| CN108447038A (en) | A kind of mesh denoising method based on non local full variation operator | |
| CN115953686B (en) | Peanut pest detection method and system based on image processing | |
| CN116596879A (en) | Grape downy mildew self-adaptive identification method based on quantiles of boundary samples |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant |