CN110274557B - Blade area measuring device and method based on computer vision - Google Patents
Blade area measuring device and method based on computer vision Download PDFInfo
- Publication number
- CN110274557B CN110274557B CN201910522389.9A CN201910522389A CN110274557B CN 110274557 B CN110274557 B CN 110274557B CN 201910522389 A CN201910522389 A CN 201910522389A CN 110274557 B CN110274557 B CN 110274557B
- Authority
- CN
- China
- Prior art keywords
- servo electric
- measuring
- box
- blade
- magnetic ring
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 22
- NIXOWILDQLNWCW-UHFFFAOYSA-N acrylic acid group Chemical group C(C=C)(=O)O NIXOWILDQLNWCW-UHFFFAOYSA-N 0.000 claims description 35
- 238000003825 pressing Methods 0.000 claims description 31
- 238000007789 sealing Methods 0.000 claims description 29
- 238000012549 training Methods 0.000 claims description 26
- 238000013527 convolutional neural network Methods 0.000 claims description 14
- 239000011324 bead Substances 0.000 claims description 8
- 238000012706 support-vector machine Methods 0.000 claims description 8
- 238000012417 linear regression Methods 0.000 claims description 5
- 238000010845 search algorithm Methods 0.000 claims description 5
- 238000012545 processing Methods 0.000 claims description 4
- 230000001678 irradiating effect Effects 0.000 claims description 3
- 238000005259 measurement Methods 0.000 claims description 2
- 238000003780 insertion Methods 0.000 claims 1
- 230000037431 insertion Effects 0.000 claims 1
- 238000000691 measurement method Methods 0.000 description 3
- 238000005303 weighing Methods 0.000 description 3
- 210000000056 organ Anatomy 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 244000038559 crop plants Species 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001066 destructive effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000003337 fertilizer Substances 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 235000015097 nutrients Nutrition 0.000 description 1
- 239000011049 pearl Substances 0.000 description 1
- 230000029553 photosynthesis Effects 0.000 description 1
- 238000010672 photosynthesis Methods 0.000 description 1
- 230000000243 photosynthetic effect Effects 0.000 description 1
- 230000029058 respiratory gaseous exchange Effects 0.000 description 1
- 230000005068 transpiration Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/28—Measuring arrangements characterised by the use of optical techniques for measuring areas
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computing Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Medical Informatics (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The invention discloses a blade area measuring device and method based on computer vision, and particularly relates to the field of machine vision, wherein the blade area measuring device comprises a workbench, a measuring machine body is arranged at the top of the workbench, a first servo electric cylinder is arranged at the top of the measuring machine body, the number of the first servo electric cylinders is two, ejector rods are connected to the top ends of output shafts of the two first servo electric cylinders, push rods are arranged at the bottoms of the ejector rods, first connecting rods are arranged on two sides of each push rod, connecting pipes are arranged at the bottom ends of the first connecting rods, a measuring box is arranged at the bottom end of each push rod, and an image collecting device is arranged inside the measuring machine body; the measuring box comprises a box body, a limiting pipe is arranged at the top of the box body, and a first magnetic ring is embedded on the outer wall of the box body. The method is simple to operate and high in precision, effectively solves the problems that the traditional measuring method is large in damage to the blade and is easy to generate errors, can accurately measure the slight change of the blade area, and obtains high-precision blade area information.
Description
Technical Field
The invention relates to the field of machine vision, in particular to a blade area measuring device and method based on computer vision.
Background
The leaves are important physiological organs of crops, photosynthesis, respiration and transpiration are all performed by the leaf organs, and the change of the area of the leaves can reflect the quantity of photosynthetic products accumulated by the leaves and the utilization condition of crop plants on nutrients such as light, temperature, water and fertilizer. Therefore, the area of the leaf can be rapidly, accurately, simply and conveniently obtained without damage, and the high-yield and high-efficiency cultivation strategy of the crop can be promoted.
The traditional blade area measurement method comprises an empirical formula method, a nine-sense-of-the-world method and a weighing method, wherein the empirical formula method estimates an area value through the length and the width of a blade and an empirical coefficient, and the accuracy is not high. The nine-officer lattice method is characterized in that leaves are flatly laid on a flat plate with fixed-size lattices, the number of the lattices covered by the leaves is calculated, the judgment subjectivity on the area of the lattices which cannot be completely covered by the edges of the leaves is too strong, and the repeatability of a measuring result is poor. The weighing method is a method in which a blade region with a known area is taken on a blade, and the area value is converted by weight proportion after weighing, and the method is destructive and cannot be continuously carried out on the same blade.
Disclosure of Invention
In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide a device and a method for measuring a leaf area based on computer vision, wherein red-light images acquired by a plurality of camera units are input into a computer and are distributed into a plurality of groups of images, the plurality of groups of images form a training set, each group of images is labeled as a category, the training set is used for performing extrinsic feature recognition, in R-CNN, a selective search algorithm is used to scan the input training set to find possible objects therein, a plurality of region suggestions are produced, then a convolutional neural network is run on the region suggestions, finally, each convolutional neural network is output to an SVM, a linear regression is used to tighten a bounding box of the objects, the bounding box of the object in the training set is read, the area of the leaf is calculated by using the bounding box of the object, the operation is simple and accurate, the problem that the traditional measuring method is large in damage to the blades and prone to generating errors is effectively solved, and small changes of the blade area can be accurately measured.
In order to achieve the purpose, the invention provides the following technical scheme: a blade area measuring device and method based on computer vision comprises a workbench, wherein a measuring machine body is arranged at the top of the workbench, first servo electric cylinders are arranged at the top of the measuring machine body, the number of the first servo electric cylinders is two, ejector rods are connected to the top ends of output shafts of the two first servo electric cylinders, push rods are arranged at the bottoms of the ejector rods, first connecting rods are arranged on two sides of each push rod, connecting pipes are arranged at the bottom ends of the first connecting rods, a measuring box is arranged at the bottom ends of the push rods, and an image collecting device is arranged inside the measuring machine body;
the measuring box comprises a box body, a limiting pipe is arranged at the top of the box body, a first magnetic ring is embedded on the outer wall of the box body, a pressing block is arranged in an inner cavity of the box body, a connecting spring is arranged at the top of the pressing block, a light chamber is formed at the bottom end of the pressing block, a plurality of red lamp beads are arranged in the light chamber, a first acrylic plate is arranged at the bottom of the light chamber, a plurality of connecting holes are formed in the bottom wall of the box body, an electromagnet is arranged at the top of the inner cavity of the connecting holes, a placing cover plate is arranged at the bottom of the box body, a plurality of magnetic rods are arranged at the edge of the surface of the placing cover plate, a placing groove is formed in the middle of the placing cover;
the image collecting device comprises a sealing pipe, a second magnetic ring is sleeved outside the sealing pipe, an image table is arranged at the bottom end of the sealing pipe, a collecting cavity is formed in the top of the image table, a plurality of camera units are arranged at the bottom of the inner cavity of the collecting cavity, a third acrylic plate is arranged at the top of each camera unit, second servo electric cylinders are arranged on two sides of the image table, and second connecting rods are arranged at the top ends of output shafts of the second servo electric cylinders.
In a preferred embodiment, the bottom of the workbench is provided with a support rod, and the bottom end of the support rod is provided with a movable wheel.
In a preferred embodiment, the measuring box is movably connected with the measuring machine body through a push rod, a push rod and a first servo electric cylinder, the top end of the first connecting rod is fixedly connected with the push rod, the bottom end of the connecting pipe is fixedly arranged at the top of the measuring box, and the first connecting rod is movably connected with the connecting pipe.
In a preferred embodiment, the limiting pipe penetrates through the top of the box body, the bottom end of the push rod penetrates through the limiting pipe and is fixedly connected with the top of the pressing block, the top end of the connecting spring is fixedly connected with the top of the inner cavity of the box body, the outer wall of the first magnetic ring and the outer wall of the box body are arranged in a coplanar manner, and the pressing block is movably connected with the box body through the connecting spring.
In a preferred embodiment, the sealing tube is embedded in the top of the measuring machine body, the top of the sealing tube is coplanar with the top of the measuring machine body, the third acrylic plate is embedded in the top end of the imaging table, and the sealing tube is arranged right above the third acrylic plate.
In a preferred embodiment, the bottom end of the second servo electric cylinder is fixedly connected with the bottom of the inner cavity of the measuring machine body, the end of the second connecting rod is fixedly connected with a second magnetic ring, the second magnetic ring is movably connected with the outer wall of the sealing pipe through a second connecting rod and the second servo electric cylinder, and a control console is arranged on one side of the measuring machine body.
A leaf area measurement method based on computer vision specifically comprises the following steps
Step one, a control command is sent by using a control console, the electromagnet is controlled to be powered off, the placing cover plate is taken out from the bottom of the measuring box, the picked blade is placed on a second acrylic plate in the placing groove and is paved, then the electromagnet is controlled to be powered on, a magnetic rod at the top of the placing cover plate is inserted into a connecting hole in an aligning mode, the magnetic rod is adsorbed to the electromagnet, and the placing cover plate is fixed to the bottom of the measuring box;
setting the extension height of an output shaft of a second servo electric cylinder, fixing the height of a second magnetic ring, controlling a first servo electric cylinder to work, setting the moving distance of the first servo electric cylinder according to the distance between the first magnetic ring and the second magnetic ring, driving an ejector rod to move by the first servo electric cylinder, moving a measuring box at the bottom end of a push rod to the inside of a sealing pipe, stopping the first servo electric cylinder when the first magnetic ring and the second magnetic ring are adsorbed, controlling the first servo electric cylinder to continuously push downwards for 1cm, enabling the push rod to extrude a pressing block in a limiting rod, enabling the pressing block to move downwards to enable a first acrylic plate at the bottom of the pressing block to be pressed on the surface of a blade on the second acrylic plate, and enabling the blade to be flattened in a placing groove;
controlling a plurality of red lamp beads to work, irradiating the blades by using red light with long wavelength, injecting the red light which is not shielded by the blades into a collecting cavity through a second acrylic plate and a third acrylic plate, controlling a plurality of camera units to work at the moment, and capturing and shooting images of the blades;
inputting red-light image pictures acquired by a plurality of camera units into a computer, distributing the red-light image pictures into a plurality of groups of images, enabling the plurality of groups of images to form a training set, marking each group of images as a category, and using the training set to perform external feature recognition;
step five, acquiring a training set by a convolutional neural network in an image processing unit, scanning and inputting the training set by using a selective search algorithm in an R-CNN (R-CNN), searching possible objects in the training set, producing a plurality of regional suggestions, operating a convolutional neural network on the regional suggestions, finally outputting each convolutional neural network to an SVM (support vector machine), tightening a bounding box of the object by using a linear regression, and reading the bounding box of the object contained in the training set;
and step six, calculating the area of each object frame by using a Helen formula according to the acquired length and width data of the object frames, and then obtaining the area of the measured blade after superposition.
The invention has the technical effects and advantages that:
1. the height of a second magnetic ring is fixed by setting the extension height of an output shaft of a second servo electric cylinder, the first servo electric cylinder is controlled to work, the moving distance of the first servo electric cylinder is set according to the distance between the first magnetic ring and the second magnetic ring, the first servo electric cylinder drives an ejector rod to move, a measuring box at the bottom end of the push rod moves to the inside of a sealing pipe, when the first magnetic ring and the second magnetic ring are adsorbed, the first servo electric cylinder stops working, then the first servo electric cylinder is controlled to continue to push downwards for 1cm, the push rod is enabled to extrude a pressing block in a limiting rod, the pressing block moves downwards a first acrylic plate at the bottom of the pressing block to be pressed on the surface of a blade on a second acrylic plate, the blade is flattened in a placing groove, a plurality of red light beads are controlled to work, the blade is irradiated by red light of the blade, the red light which is not shielded by the blade is injected into a collecting cavity through the second acrylic plate and a third acrylic plate, at the moment, the plurality of camera units are controlled to work, the blade images are captured and shot, so that the error between the obtained image information and the real outline of the blade is minimum, the obtained image information is more real, and high-precision blade area information is obtained when the convolution neural network is used for calculating the blade area;
2. the red light image pictures acquired by a plurality of camera units are input into a computer and distributed into a plurality of groups of images, the plurality of groups of images form a training set, each group of images is marked as a category, the training set is used for external feature recognition, in R-CNN, a selective search algorithm is used to scan the input training set for possible objects therein, to produce a plurality of regional suggestions, then, a convolutional neural network is operated on the region suggestions, finally, each convolutional neural network is output to the SVM, a linear regression is used for tightening the bounding box of the object, the method has the advantages that the object frame containing the object in the training set is read, the area of the blade is calculated by using the object frame, the operation is simple, the precision is high, the problems that the traditional measuring method is large in damage to the blade and error is easy to generate are effectively solved, and the slight change of the area of the blade can be accurately measured.
Drawings
Fig. 1 is a schematic view of the overall structure of the present invention.
Fig. 2 is a schematic structural view of the measuring box of the present invention when moving down.
Fig. 3 is a schematic view of the overall structure of the present invention.
Fig. 4 is a schematic view of the measuring chamber structure of the present invention.
Fig. 5 is a schematic diagram of a detailed structure a in fig. 2 according to the present invention.
Fig. 6 is a schematic view of the structure of the placing cover plate of the present invention.
The reference signs are: 1 workstation, 2 measure the organism, 3 first servo electronic jar, 4 ejector pins, 5 push rods, 6 head rod, 7 connecting pipes, 8 measurement casees, 9 image collection device, 10 boxes, 11 spacing pipes, 12 first magnetic rings, 13 pressfitting piece, 14 connecting spring, 15 lamp light rooms, 16 red lamp pearls, 17 first ya keli board, 18 connecting holes, 19 electro-magnets, 20 place the apron, 21 magnetic poles, 22 standing grooves, 23 second ya keli board, 24 sealed tubes, 25 second magnetic rings, 26 image platform, 27 collection chamber, 28 camera units, 29 third ya keli board, 30 second servo electronic jar, 31 second connecting rods.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The blade area measuring device and method based on computer vision as shown in fig. 1-6 comprise a workbench 1, wherein a measuring machine body 2 is arranged at the top of the workbench 1, a first servo electric cylinder 3 is arranged at the top of the measuring machine body 2, the number of the first servo electric cylinders 3 is two, top ends of output shafts of the two first servo electric cylinders 3 are connected with ejector rods 4, the bottoms of the ejector rods 4 are provided with push rods 5, two sides of the push rods 5 are provided with first connecting rods 6, the bottom ends of the first connecting rods 6 are provided with connecting pipes 7, the bottom ends of the push rods 5 are provided with measuring boxes 8, and an image collecting device 9 is arranged inside the measuring machine body 2;
the measuring box 8 comprises a box body 10, a limiting pipe 11 is arranged at the top of the box body 10, a first magnetic ring 12 is embedded on the outer wall of the box body 10, a pressing block 13 is arranged in the inner cavity of the box body 10, a connecting spring 14 is arranged at the top of the pressing block 13, a light chamber 15 is arranged at the bottom end of the pressing block 13, a plurality of red lamp beads 16 are arranged in the light chamber 15, a first acrylic plate 17 is arranged at the bottom of the light chamber 15, a plurality of connecting holes 18 are formed in the bottom wall of the box body 10, an electromagnet 19 is arranged at the top of the inner cavity of the connecting hole 18, a placing cover plate 20 is arranged at the bottom of the box body 10, a plurality of magnetic rods 21 are arranged at the edge of the surface of the placing cover plate 20, a placing groove 22 is formed in the middle of the placing cover plate 20;
the image collecting device 9 comprises a sealing pipe 24, a second magnetic ring 25 is sleeved outside the sealing pipe 24, an image table 26 is arranged at the bottom end of the sealing pipe 24, a collecting cavity 27 is formed in the top of the image table 26, a plurality of camera units 28 are arranged at the bottom of the inner cavity of the collecting cavity 27, a third acrylic plate 29 is arranged at the top of each camera unit 28, second servo electric cylinders 30 are arranged on two sides of the image table 26, and second connecting rods 31 are arranged at the top ends of output shafts of the second servo electric cylinders 30.
The bottom of the workbench 1 is provided with a support rod, the bottom end of the support rod is provided with a movable wheel, the measuring box 8 is movably connected with the measuring machine body 2 through a push rod 5, a push rod 4 and a first servo electric cylinder 3, the top end of a first connecting rod 6 is fixedly connected with the push rod 4, the bottom end of a connecting pipe 7 is fixedly arranged at the top of the measuring box 8, and the first connecting rod 6 is movably connected with the connecting pipe 7;
the limiting pipe 11 penetrates through the top of the box body 10, the bottom end of the push rod 5 penetrates through the limiting pipe 11 and is fixedly connected with the top of the pressing block 13, the top end of the connecting spring 14 is fixedly connected with the top of the inner cavity of the box body 10, the outer wall of the first magnetic ring 12 is arranged in a coplanar manner with the outer wall of the box body 10, and the pressing block 13 is movably connected with the box body 10 through the connecting spring 14;
the sealing tube 24 is embedded at the top of the measuring machine body 2, the top of the sealing tube 24 and the top of the measuring machine body 2 are arranged in a coplanar manner, the third acrylic plate 29 is embedded at the top end of the image table 26, and the sealing tube 24 is arranged right above the third acrylic plate 29;
the bottom end of the second servo electric cylinder 30 is fixedly connected with the bottom of the inner cavity of the measuring machine body 2, the end part of the second connecting rod 31 is fixedly connected with the second magnetic ring 25, the second magnetic ring 25 is movably connected with the outer wall of the sealing pipe 24 through the second connecting rod 31 and the second servo electric cylinder 30, and a control console is arranged on one side of the measuring machine body 2.
The implementation mode is specifically as follows: the control console is used for sending a control command, controlling the electromagnet 19 to be powered off, taking the placing cover plate 20 out of the bottom of the measuring box 8, placing the taken blades on the second acrylic plate 23 in the placing groove 22, paving the blades, then controlling the electromagnet 19 to be powered on, inserting the magnetic rod 21 at the top of the placing cover plate 20 into the connecting hole 18 to enable the magnetic rod 21 to be adsorbed with the electromagnet 19, fixing the placing cover plate 20 at the bottom of the measuring box 8, setting the extension height of the output shaft of the second servo electric cylinder 30, fixing the height of the second magnetic ring 25, controlling the first servo electric cylinder 3 to work, setting the moving distance of the first servo electric cylinder 3 according to the distance between the first magnetic ring 12 and the second magnetic ring 25, driving the ejector rod 4 to move by the first servo electric cylinder 3, moving the measuring box 8 at the bottom end of the push rod 5 into the sealing pipe 24, and when the first magnetic ring 12 is adsorbed with the second magnetic ring 25, the first servo electric cylinder 3 stops working, then the first servo electric cylinder 3 is controlled to continuously push 1cm downwards, the push rod 5 is enabled to extrude the pressing block 13 in the limiting rod, the pressing block 13 moves downwards, the first acrylic plate 17 at the bottom of the pressing block is pressed on the surface of the blade on the second acrylic plate 23, the blade is enabled to be flattened in the placing groove 22, the plurality of red lamp beads 16 are controlled to work, the blade is irradiated by red light with long wavelength, the red light which is not shielded by the blade is emitted into the collecting cavity 27 through the second acrylic plate 23 and the third acrylic plate 29, at the moment, the plurality of camera units 28 are controlled to work, capturing and shooting the blade image, processing the image by a computer to obtain the area of the blade, and then controlling the first servo electric cylinder 3 to move upwards, removing the measuring box 8 from the sealing pipe 24, detaching the placing cover plate 20, and taking out the blade to measure other blades.
A leaf area measurement method based on computer vision specifically comprises the following steps;
step one, a control command is sent by using a control console, the electromagnet 19 is controlled to be powered off, the placing cover plate 20 is taken out from the bottom of the measuring box 8, the picked blade is placed on a second acrylic plate 23 in a placing groove 22 and is laid flat, then the electromagnet 19 is controlled to be powered on, a magnetic rod 21 at the top of the placing cover plate 20 is aligned with a connecting hole 18 to be inserted, the magnetic rod 21 and the electromagnet 19 are adsorbed, and the placing cover plate 20 is fixed at the bottom of the measuring box 8;
step two, setting the extension height of an output shaft of a second servo electric cylinder 30, fixing the height of a second magnetic ring 25, controlling the first servo electric cylinder 3 to work, setting the moving distance of the first servo electric cylinder 3 according to the distance between a first magnetic ring 12 and the second magnetic ring 25, driving a push rod 4 to move by the first servo electric cylinder 3, moving a measuring box 8 at the bottom end of a push rod 5 to the inside of a sealing pipe 24, stopping the first servo electric cylinder 3 when the first magnetic ring 12 and the second magnetic ring 25 are adsorbed, then controlling the first servo electric cylinder 3 to continue to push downwards for 1cm, enabling the push rod 5 to extrude a pressing block 13 in a limiting rod, enabling the pressing block 13 to move downwards, enabling a first acrylic plate 17 at the bottom of the pressing block to be pressed on the surface of a blade on a second acrylic plate 23, and enabling the blade to be flattened in a placing groove 22;
step three, controlling a plurality of red lamp beads 16 to work, irradiating the blades by using red light with long wavelength, emitting the red light which is not shielded by the blades into a collecting cavity 27 through a second acrylic plate 23 and a third acrylic plate 29, controlling a plurality of camera units 28 to work at the moment, and capturing and shooting the images of the blades;
inputting the red-light image pictures acquired by the plurality of camera units 28 into the computer, and distributing the red-light image pictures into a plurality of groups of images, so that the plurality of groups of images form a training set, each group of images is marked as a category, and the training set is used for external feature recognition;
step five, acquiring a training set by a convolutional neural network in an image processing unit, scanning and inputting the training set by using a selective search algorithm in an R-CNN (R-CNN), searching possible objects in the training set, producing a plurality of regional suggestions, operating a convolutional neural network on the regional suggestions, finally outputting each convolutional neural network to an SVM (support vector machine), tightening a bounding box of the object by using a linear regression, and reading the bounding box of the object contained in the training set;
and step six, calculating the area of each object frame by using a Helen formula according to the acquired length and width data of the object frames, and then obtaining the area of the measured blade after superposition.
The working principle of the invention is as follows:
referring to the accompanying drawings 1-6 of the specification, by arranging the sealing tube 24 and the measuring box 8, the measured blade is placed in the placing groove 22 in the measuring box 8 and is pressed by the pressing block 13, when an image is obtained, the measuring box 8 is placed in the sealing tube 24, the blade is irradiated by red light with long wavelength, the image information of the blade is obtained by the plurality of camera units 28, the error between the obtained image information and the real contour of the blade is minimized, the obtained image information is more real, and when the area of the blade is calculated by using a convolutional neural network, high-precision blade area information is obtained.
The points to be finally explained are: first, in the description of the present application, it should be noted that, unless otherwise specified and limited, the terms "mounted," "connected," and "connected" should be understood broadly, and may be a mechanical connection or an electrical connection, or a communication between two elements, and may be a direct connection, and "upper," "lower," "left," and "right" are only used to indicate a relative positional relationship, and when the absolute position of the object to be described is changed, the relative positional relationship may be changed;
secondly, the method comprises the following steps: in the drawings of the disclosed embodiments of the invention, only the structures related to the disclosed embodiments are referred to, other structures can refer to common designs, and the same embodiment and different embodiments of the invention can be combined with each other without conflict;
and finally: the above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that are within the spirit and principle of the present invention are intended to be included in the scope of the present invention.
Claims (6)
1. The blade area measuring device based on computer vision is characterized by comprising a workbench (1), wherein a measuring machine body (2) is arranged at the top of the workbench (1), first servo electric cylinders (3) are arranged at the top of the measuring machine body (2), the number of the first servo electric cylinders (3) is two, ejector rods (4) are connected to the top ends of output shafts of the two first servo electric cylinders (3), push rods (5) are arranged at the bottoms of the ejector rods (4), first connecting rods (6) are arranged on two sides of each push rod (5), connecting pipes (7) are arranged at the bottom ends of the first connecting rods (6), a measuring box (8) is arranged at the bottom end of each push rod (5), and an image collecting device (9) is arranged inside the measuring machine body (2);
the measuring box (8) comprises a box body (10), a limiting pipe (11) is arranged at the top of the box body (10), a first magnetic ring (12) is embedded on the outer wall of the box body (10), a pressing block (13) is arranged in the inner cavity of the box body (10), a connecting spring (14) is arranged at the top of the pressing block (13), a lamp light chamber (15) is arranged at the bottom end of the pressing block (13), a plurality of red lamp beads (16) are arranged in the lamp light chamber (15), a first acrylic plate (17) is arranged at the bottom of the lamp light chamber (15), a plurality of connecting holes (18) are formed in the bottom wall of the box body (10), an electromagnet (19) is arranged at the top of the inner cavity of the connecting holes (18), a placing cover plate (20) is arranged at the bottom of the box body (10), a plurality of magnetic rods (21) are arranged at the edge of the surface of the placing cover plate (20), a, a second acrylic plate (23) is arranged at the bottom of the placing groove (22);
the image collecting device (9) comprises a sealing pipe (24), a second magnetic ring (25) is sleeved outside the sealing pipe (24), an image table (26) is arranged at the bottom end of the sealing pipe (24), a collecting cavity (27) is formed in the top of the image table (26), a plurality of camera units (28) are arranged at the bottom of the inner cavity of the collecting cavity (27), a third acrylic plate (29) is arranged at the top of each camera unit (28), second servo electric cylinders (30) are arranged on two sides of the image table (26), and second connecting rods (31) are arranged at the top ends of output shafts of the second servo electric cylinders (30);
the bottom end of the second servo electric cylinder (30) is fixedly connected with the bottom of the inner cavity of the measuring machine body (2), the end part of the second connecting rod (31) is fixedly connected with the second magnetic ring (25), the second magnetic ring (25) is movably connected with the outer wall of the sealing pipe (24) through the second connecting rod (31) and the second servo electric cylinder (30), and a control console is arranged on one side of the measuring machine body (2).
2. A computer vision based blade area measuring device according to claim 1, wherein: the bottom of the workbench (1) is provided with a support rod, and the bottom end of the support rod is provided with a movable wheel.
3. A computer vision based blade area measuring device according to claim 1, wherein: the measuring box (8) is movably connected with the measuring machine body (2) through a push rod (5), an ejector rod (4) and a first servo electric cylinder (3), the top end of a first connecting rod (6) is fixedly connected with the ejector rod (4), the bottom end of a connecting pipe (7) is fixedly arranged at the top of the measuring box (8), and the first connecting rod (6) is movably connected with the connecting pipe (7).
4. A computer vision based blade area measuring device according to claim 1, wherein: the limiting pipe (11) penetrates through the top of the box body (10), the bottom end of the push rod (5) penetrates through the limiting pipe (11) and is fixedly connected with the top of the pressing block (13), the top end of the connecting spring (14) is fixedly connected with the top of the inner cavity of the box body (10), the outer wall of the first magnetic ring (12) and the outer wall of the box body (10) are arranged in a coplanar mode, and the pressing block (13) is movably connected with the box body (10) through the connecting spring (14).
5. A computer vision based blade area measuring device according to claim 1, wherein: sealing tube (24) inlay and locate the top of measuring organism (2), sealing tube (24) top sets up with measurement organism (2) top coplane, third ya keli board (29) inlay the top of locating image platform (26), sealing tube (24) set up directly over third ya keli board (29).
6. A method for measuring blade area based on computer vision by using the blade area measuring device of any one of claims 1 to 5, which comprises the following steps:
step one, a control console is used for sending a control command, controlling an electromagnet (19) to be powered off, taking a placing cover plate (20) out of the bottom of a measuring box (8), placing the taken blade on a second acrylic plate (23) in a placing groove (22), paving the blade, then controlling the electromagnet (19) to be powered on, aligning a magnetic rod (21) at the top of the placing cover plate (20) to a connecting hole (18) for insertion, enabling the magnetic rod (21) to be adsorbed with the electromagnet (19), and fixing the placing cover plate (20) at the bottom of the measuring box (8);
step two, setting the extension height of an output shaft of a second servo electric cylinder (30), fixing the height of a second magnetic ring (25), controlling the first servo electric cylinder (3) to work, setting the moving distance of the first servo electric cylinder (3) according to the distance between a first magnetic ring (12) and the second magnetic ring (25), driving a mandril (4) to move by the first servo electric cylinder (3), moving a measuring box (8) at the bottom end of a push rod (5) into a sealing pipe (24), stopping the first servo electric cylinder (3) when the first magnetic ring (12) and the second magnetic ring (25) are adsorbed, then controlling the first servo electric cylinder (3) to continuously push downwards for 1cm, enabling the push rod (5) to extrude a pressing block (13) in a limiting rod, and enabling the pressing block (13) to move downwards to press a first acrylic plate (17) at the bottom of the pressing block to be pressed on the surface of a blade on a second acrylic plate (23), flattening the leaves in the placement groove (22);
step three, controlling a plurality of red lamp beads (16) to work, irradiating the blades by using red light with long wavelength, emitting the red light which is not shielded by the blades into a collecting cavity (27) through a second acrylic plate (23) and a third acrylic plate (29), controlling a plurality of camera units (28) to work at the moment, and capturing and shooting the images of the blades;
inputting red-light image pictures acquired by a plurality of camera units (28) into a computer, distributing the red-light image pictures into a plurality of groups of images, enabling the plurality of groups of images to form a training set, marking each group of images as a category, and using the training set to perform external feature recognition;
step five, acquiring a training set by a convolutional neural network in an image processing unit, scanning and inputting the training set by using a selective search algorithm in an R-CNN (R-CNN), searching possible objects in the training set, producing a plurality of regional suggestions, operating a convolutional neural network on the regional suggestions, finally outputting each convolutional neural network to an SVM (support vector machine), tightening a bounding box of the object by using a linear regression, and reading the bounding box of the object contained in the training set;
and step six, calculating the area of each object frame by using a Helen formula according to the acquired length and width data of the object frames, and then obtaining the area of the measured blade after superposition.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910522389.9A CN110274557B (en) | 2019-06-17 | 2019-06-17 | Blade area measuring device and method based on computer vision |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910522389.9A CN110274557B (en) | 2019-06-17 | 2019-06-17 | Blade area measuring device and method based on computer vision |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110274557A CN110274557A (en) | 2019-09-24 |
CN110274557B true CN110274557B (en) | 2021-02-09 |
Family
ID=67960911
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910522389.9A Active CN110274557B (en) | 2019-06-17 | 2019-06-17 | Blade area measuring device and method based on computer vision |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110274557B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114241033B (en) * | 2021-12-17 | 2022-07-22 | 贵州省山地资源研究所 | Ectopic acquisition method for plant leaf area vertical structure |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1952599A (en) * | 2005-10-20 | 2007-04-25 | 中国科学院自动化研究所 | A flat leaf area nondestructive measuring system and method thereof |
CN101358842A (en) * | 2008-09-19 | 2009-02-04 | 广东省农业科学院水稻研究所 | Plant blade area measuring apparatus |
JP2011133451A (en) * | 2009-11-27 | 2011-07-07 | Kyushu Univ | Optical vegetation index sensor |
CN105004284A (en) * | 2015-07-14 | 2015-10-28 | 河南科技大学 | Photoelectric tobacco leaf area measuring device |
CN205120060U (en) * | 2015-11-09 | 2016-03-30 | 平顶山学院 | Leaf area appearance |
-
2019
- 2019-06-17 CN CN201910522389.9A patent/CN110274557B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1952599A (en) * | 2005-10-20 | 2007-04-25 | 中国科学院自动化研究所 | A flat leaf area nondestructive measuring system and method thereof |
CN101358842A (en) * | 2008-09-19 | 2009-02-04 | 广东省农业科学院水稻研究所 | Plant blade area measuring apparatus |
JP2011133451A (en) * | 2009-11-27 | 2011-07-07 | Kyushu Univ | Optical vegetation index sensor |
CN105004284A (en) * | 2015-07-14 | 2015-10-28 | 河南科技大学 | Photoelectric tobacco leaf area measuring device |
CN205120060U (en) * | 2015-11-09 | 2016-03-30 | 平顶山学院 | Leaf area appearance |
Non-Patent Citations (1)
Title |
---|
设施黄瓜叶片病害检测仪设计;卫婷婷;《中国优秀硕士学位论文全文数据库 农业科技辑》;20190115(第01期);D046-277 * |
Also Published As
Publication number | Publication date |
---|---|
CN110274557A (en) | 2019-09-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105717115B (en) | High-throughput Plant phenotypic analysis device and method based on optical image technology | |
Chen et al. | Three-dimensional perception of orchard banana central stock enhanced by adaptive multi-vision technology | |
CN104482860B (en) | Fish morphological parameters self-operated measuring unit and method | |
CN110118528B (en) | Line structure light calibration method based on chessboard target | |
CN105547152A (en) | Crop phenotype field high-flux active measuring apparatus and method | |
CN107860366B (en) | Mobile greenhouse crop information measurement and diagnosis system | |
CN110152938A (en) | A kind of component dispensing track extraction method and automatically control machine people system | |
CN207081412U (en) | A kind of animal body chi ultrasound noncontacting measurement device | |
CN109405755B (en) | Large-size pipe plate aperture and pitch measuring device and measuring method | |
CN112529948A (en) | Mature pomegranate positioning method based on Mask R-CNN and 3-dimensional sphere fitting | |
CN110293559A (en) | A kind of installation method of automatic identification positioning alignment | |
CN110274557B (en) | Blade area measuring device and method based on computer vision | |
CN113920106A (en) | Corn growth three-dimensional reconstruction and stem thickness measurement method based on RGB-D camera | |
CN109584293A (en) | A kind of chamber crop three-dimensional configuration panorama measuring system based on RGB-D vision technique | |
CN113607659A (en) | Conveyor belt type crop phenotype acquisition method, system and device | |
CN113554691B (en) | Plant height measuring method | |
CN102663727B (en) | Method for calibrating parameters by dividing regions in a camera based on CMM moving target | |
Jin et al. | Low-damage transplanting method for leafy vegetable seedlings based on machine vision | |
CN117029673A (en) | Fish body surface multi-size measurement method based on artificial intelligence | |
CN103075982A (en) | Three-dimensional reconstruction and measurement device and method of greenhouse strawberry canopy | |
CN110741790A (en) | Multi-claw transplanting-sorting processing method for plug seedlings based on depth camera | |
CN109636856A (en) | Object 6 DOF degree posture information union measuring method based on HOG Fusion Features operator | |
CN110064601B (en) | Seedling detection and classification system and classification method for vegetable grafting | |
CN111390911A (en) | Manipulator position calibration system and calibration method | |
CN113822896B (en) | Plant population three-dimensional phenotype data acquisition device and method |
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 | ||
GR01 | Patent grant | ||
TR01 | Transfer of patent right | ||
TR01 | Transfer of patent right |
Effective date of registration: 20240712 Address after: Room G237, D18, Zone D, No. 6 Huangcun North Road, Tianhe District, Guangzhou City, Guangdong Province 510000 Patentee after: Guangzhou Nongwei Technology Co.,Ltd. Country or region after: China Address before: No.1 Xiangwang South Road, Gulou District, Xuzhou City, Jiangsu Province Patentee before: XUZHOU College OF INDUSTRIAL TECHNOLOGY Country or region before: China |