CN214122027U - Rice phenotype monitoring system based on machine vision technology - Google Patents
Rice phenotype monitoring system based on machine vision technology Download PDFInfo
- Publication number
- CN214122027U CN214122027U CN202120153243.4U CN202120153243U CN214122027U CN 214122027 U CN214122027 U CN 214122027U CN 202120153243 U CN202120153243 U CN 202120153243U CN 214122027 U CN214122027 U CN 214122027U
- Authority
- CN
- China
- Prior art keywords
- rice
- light shield
- machine vision
- image acquisition
- monitoring system
- 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
- 235000007164 Oryza sativa Nutrition 0.000 title claims abstract description 119
- 235000009566 rice Nutrition 0.000 title claims abstract description 119
- 238000012544 monitoring process Methods 0.000 title claims abstract description 21
- 238000005516 engineering process Methods 0.000 title claims description 15
- 240000007594 Oryza sativa Species 0.000 title 1
- 241000209094 Oryza Species 0.000 claims abstract description 118
- 239000000463 material Substances 0.000 claims abstract description 34
- 235000013339 cereals Nutrition 0.000 claims abstract description 29
- 238000003860 storage Methods 0.000 claims abstract description 18
- 239000011324 bead Substances 0.000 claims description 3
- 238000000034 method Methods 0.000 abstract description 25
- 238000004458 analytical method Methods 0.000 abstract description 5
- 230000002349 favourable effect Effects 0.000 abstract description 3
- 238000001514 detection method Methods 0.000 description 14
- 238000012216 screening Methods 0.000 description 5
- 238000004519 manufacturing process Methods 0.000 description 4
- 238000011176 pooling Methods 0.000 description 4
- 230000015572 biosynthetic process Effects 0.000 description 3
- 238000013527 convolutional neural network Methods 0.000 description 3
- 238000005520 cutting process Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000000877 morphologic effect Effects 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000003709 image segmentation Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000011084 recovery Methods 0.000 description 2
- 238000000701 chemical imaging Methods 0.000 description 1
- 238000013145 classification model Methods 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 239000012535 impurity Substances 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000007477 logistic regression Methods 0.000 description 1
- 239000002932 luster Substances 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000003333 near-infrared imaging Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000012372 quality testing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
Images
Landscapes
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The utility model discloses a rice phenotype monitoring system based on machine vision technique, the system includes: the rice grain detecting device comprises a light shield, a material storage device, an image acquisition device and a PC (personal computer), wherein the light shield is arranged above the material storage device, the material storage device is used for placing a rice sample, the image acquisition device is arranged in the light shield, the image acquisition device is used for acquiring a phenotype image of the rice sample, and a light shield cover plate is arranged at the top of the light shield; the image acquisition device is electrically connected with the PC through a data line. The utility model provides a rice phenotype monitoring system based on machine vision technique carries out comprehensive objective data acquisition and analysis to rice phenotype characteristic through machine vision technique, is favorable to grading rice high-speed, accurate quality, improves rice source quality to make the enterprise practice thrift the cost by a wide margin.
Description
Technical Field
The utility model relates to an agricultural grain quality testing technique field especially relates to a rice phenotype monitoring system based on machine vision technique.
Background
For quality detection of rice appearance, grading is mainly performed by a color selector in rice production at present, but the color selector in the market has a single technical algorithm and low accuracy. Before each batch of rice is screened by using the color selector, the equipment is usually debugged repeatedly by experienced users, and because the color selector only screens the rice color, the shape, size, texture and other phenotypic characters of the rice cannot be comprehensively judged, and ideal optimal results cannot be obtained by repeated screening for many times. In the quality detection link of the national grain depot, the appearance characters (namely the roughness) of the rice are still evaluated by a manual inspection method, so that the time and the labor are wasted, the detection result depends on the subjectivity of inspectors to a great extent, the level of the inspectors is not uniform, and the regional difference of judgment standards ensures that the accurate judgment of the rice quality detection cannot be realized at present. Especially for rice processing factories, it is difficult to perform all-round quality detection on purchased rice, and only the rice yield and the moisture are often detected.
In the huge agricultural companies abroad, the application of the image technology to grain crops and seeds is leading, and the image technology is popularized in quality detection and production processes, but the quality detection of rice by utilizing the spectral imaging and machine vision technology is still in the research stage at present, and the commercialized application is not realized.
SUMMERY OF THE UTILITY MODEL
The utility model aims at providing a rice phenotype monitoring system based on machine vision technique carries out comprehensive objective data acquisition and analysis to rice phenotype characteristic through machine vision technique, is favorable to the high-speed, accurate quality classification to rice, improves rice source quality to make the enterprise practice thrift the cost by a wide margin.
In order to achieve the above object, the utility model provides a following scheme:
a rice phenotype monitoring system based on machine vision technology, the system comprising: the rice grain detecting device comprises a light shield, a material storage device, an image acquisition device and a PC (personal computer), wherein the light shield is arranged above the material storage device, the material storage device is used for placing a rice sample, the image acquisition device is arranged in the light shield, the image acquisition device is used for acquiring a phenotype image of the rice sample, and a light shield cover plate is arranged at the top of the light shield; the image acquisition device is electrically connected with the PC through a data line.
Furthermore, the material storage device comprises a material drawer and drawer guide rails, the two parallel drawer guide rails are fixedly arranged on two opposite sides of the bottom of the light shield, the material drawer is in sliding connection with the drawer guide rails through sliding blocks, and the material drawer is used for placing rice samples.
Further, a fixed support is fixedly connected to the inner side wall of the light shield, the image acquisition device is arranged on the fixed support and comprises a multispectral camera and an LED annular light source, and the multispectral camera and the LED annular light source are respectively fixed above and below the fixed support; the multispectral camera is electrically connected with the PC through a data line.
Further, the LED annular light source is provided with LED lamp beads with different wave bands.
Further, the system also comprises an LED backlight plate which is arranged below the material storage device.
According to the utility model provides a concrete embodiment, the utility model discloses a following technological effect: the utility model provides an among the rice phenotype monitoring system based on machine vision technique, the lens hood can shelter from external disturbance, improve image acquisition quality, gather the image of rice sample through image acquisition device, specifically adopt the PC to link to each other with multispectral camera, the inspector tiles rice to the material drawer and can accomplish the accurate formation of image of rice, supporting LED annular light source in the lens hood has arranged the LED lamp pearl of different wave bands, adopt the effect of LED backlight panel to make the outline of rice image clearer, easily image segmentation algorithm obtains single grain rice; after the image acquisition is finished, the material drawer with the recovery groove is pulled out along the drawer guide rail and is inclined to collect the rice, the detector pours new rice into the material drawer and then puts the new rice back to perform new detection, and the detection efficiency is improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without inventive labor.
FIG. 1 is a schematic diagram of a rice phenotype monitoring system based on machine vision technology according to an embodiment of the present invention;
FIG. 2 is a flow chart of a rice phenotype monitoring method based on machine vision technology according to an embodiment of the present invention;
fig. 3 is a schematic diagram of a convolutional neural network VGG16 model according to an embodiment of the present invention;
description of the drawings: 1. a drawer rail; 2. a light shield cover plate; 3. a multispectral camera; 4. an LED annular light source; 5. fixing a bracket; 6. a light shield; 7. a material drawer; 8. an LED backlight plate; 9. a data line; 10. a PC machine.
Detailed Description
The technical solutions in the embodiments of the present invention will be described clearly and completely with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments in the present invention, all other embodiments obtained by a person skilled in the art without creative work belong to the protection scope of the present invention.
The utility model aims at providing a rice phenotype monitoring system based on machine vision technique carries out comprehensive objective data acquisition and analysis to rice phenotype characteristic through machine vision technique, is favorable to the high-speed, accurate quality classification to rice, improves rice source quality to make the enterprise practice thrift the cost by a wide margin.
In order to make the above objects, features and advantages of the present invention more comprehensible, the present invention is described in detail with reference to the accompanying drawings and the detailed description.
As shown in fig. 1, the utility model provides a rice phenotype monitoring system based on machine vision technique, include: the rice grain harvester comprises a light shield 2, a material storage device, an image acquisition device and a PC (personal computer) 10, wherein the light shield 2 is covered above the material storage device, the material storage device is used for placing a rice sample, the image acquisition device is arranged in the light shield 2 and is used for acquiring a phenotype image of the rice sample, and a light shield cover plate 6 is arranged at the top of the light shield 2; the image acquisition device is electrically connected with the PC 10 through a data line 9.
The material storage device comprises a material drawer 7 and drawer guide rails 1, the two parallel drawer guide rails 1 are fixedly arranged on two opposite sides of the bottom of the light shield 2, the material drawer 7 is in sliding connection with the drawer guide rails 1 through sliding blocks, and the material drawer 7 is used for placing rice samples.
The inner side wall of the light shield 2 is fixedly connected with a fixed support 5, the image acquisition device is arranged on the fixed support 5 and comprises a multispectral camera 3 and an LED annular light source 4, and the multispectral camera 3 and the LED annular light source 4 are respectively fixed above and below the fixed support 5; the multispectral camera 3 is electrically connected with the PC 10 through a data line 9.
The LED annular light source 4 is provided with LED lamp beads with different wave bands; the system further comprises an LED backlight plate 8, and the LED backlight plate 8 is arranged below the material storage device.
The utility model provides an among the rice phenotype monitoring system based on machine vision technique, the lens hood can shelter from external disturbance, improve image acquisition quality, gather the image of rice sample through image acquisition device, specifically adopt the PC to link to each other with multispectral camera, the inspector tiles rice to the material drawer and can accomplish the accurate formation of image of rice, supporting LED annular light source in the lens hood has arranged the LED lamp pearl of different wave bands, adopt the effect of LED backlight panel to make the outline of rice image clearer, easily image segmentation algorithm obtains single grain rice; after the image acquisition is finished, the material drawer with the recovery groove is pulled out along the drawer guide rail and is inclined to collect the rice, the detector pours new rice into the material drawer and then puts the new rice back to perform new detection, and the detection efficiency is improved.
As shown in fig. 2, the utility model also provides a rice phenotype monitoring method based on machine vision technology, which is applied to the rice phenotype monitoring system based on machine vision technology, and comprises the following steps:
s1, acquiring an image of the rice sample by using the multispectral camera and transmitting the image to a PC (personal computer);
s2, segmenting single rice from the images collected by the multispectral camera by using a watershed algorithm;
s3, obtaining color characteristics and texture characteristics of the single grain rice from the segmented image of the single grain rice, carrying out binarization operation on the obtained single grain rice, and obtaining morphological characteristics of the rice through the communication areas of the pixel points.
The utility model discloses an algorithm is cut apart to the image is watershed algorithm, because the rice that tiles probably adheres to each other, consequently need cut apart single grain rice from the image that the multispectral camera gathered. In step S2, the watershed algorithm is used to segment single grain rice from the image collected by the multispectral camera, and the method specifically includes:
firstly, changing a colored original image into a gray image, then obtaining a binary image of the rice approximate outline by adopting an OSTU algorithm, wherein small black and white noises possibly exist in the binary image and can be removed by using a form opening operation; the method comprises the steps of performing expansion operation on an obtained image to enable the image of real rice to be a subset of the expanded image, then obtaining a central area of each grain of rice by using a distanceTransform algorithm, enabling the image of the central area to be the subset of the image of the real rice, subtracting the central area from the expanded image to be an uncertain rice edge area, then creating a mask image by using a connected components algorithm to enable the image of the central area of each grain of rice to have a mark of the image, representing the number of each grain of rice by using different colors for distinguishing, finally obtaining the exact position of the edge of each grain of rice by using a watershed algorithm, reserving only one grain of rice in an original image before each cutting in order to prevent interference with an adjacent rice image in the cutting process, and finally cutting the single grain of rice from the image through the obtained position.
In step S3, obtaining color features and texture features of the single grain rice from the segmented image of the single grain rice, performing binarization operation on the obtained single grain rice, and obtaining morphological features of the rice through a connected region of pixel points, specifically including:
the characteristic extraction of the method is to carry out binarization operation on the obtained single rice grains according to color characteristics (RGB, HSI and NIR) and texture characteristics (some gray scale change rules) obtained from the divided single rice grains, and obtain morphological characteristics (area, length, width, perimeter, length-to-axis ratio and eccentricity) of the rice through a connected region of pixel points. Since the rice is randomly placed in the positions and the directions, the Features of each picture are extracted by using a speedup Robust Features (SURF) algorithm, and the algorithm has the characteristics of scale invariance and rotation invariance. The algorithm firstly uses an integral graph to calculate convolution, then uses Hessian response to measure whether a certain point is a characteristic point, creates a descriptor to describe the characteristic, and puts the obtained characteristic into classifiers such as logistic regression, support vector machine, K nearest neighbor, ensemble learning and the like to classify the quality of rice by a machine learning method.
The method establishes a rice screening model according to a deep learning algorithm of a convolutional neural network, and uses a large number of 4-channel single-grain rice images (RGB + NIR) acquired by a multispectral camera as a data set. The machine learning method specifically comprises the following steps:
as shown in fig. 3, the convolutional neural network VGG16 model is used as a basic model, the model has 16 layers, 13 convolutional layers and 2 fully-connected layers, after the first convolution with 64 convolutional kernels, a pooling layer is used, after the second convolution with 128 convolutional kernels, the pooling layer is used again, 256 convolutional kernels are used for three times, the pooling layer is used again, after the third convolution with 512 convolutional kernels is repeated twice again, the pooling layer is used again, then the fully-connected layer with 256 sizes and one Dropout layer are used twice, and finally the classification is performed by softmax. The model reduces the parameters of the model and improves the rice recognition precision and efficiency of the model by changing the last three full-connection layers of the original model. And (3) training the rice data set in the model to obtain a final classification model, and finally accurately classifying the rice with different qualities.
In the actual seed production process, rice of the same variety but different batches or different ages may have different appearance traits, and in addition, the change of the environment (such as illumination, vibration, dust and the like) of the system can cause the drift of the algorithm model. Therefore, to ensure the stability of the system, the upgrade and maintenance steps of the hierarchical model are as follows:
firstly, establishing a self-adaptive calibration program of a system, and automatically eliminating the influence caused by the environment by comparing current data with historical data;
secondly, a simple and easy-to-use rice screening model training process is established, so that a user can conveniently and quickly calibrate and maintain the model, and the optimization of an output result is ensured;
thirdly, a marked rice image big database is established, a cross-variety generalized hierarchical model is developed by applying a deep learning algorithm, and rice can still be intelligently optimized and graded when a user does not have the capability of accurately calibrating a system and developing a rice screening training model;
and fourthly, finally establishing a grain resource database and a regional grain big data platform to form a regional grain data center.
The utility model provides a rice phenotype monitoring method based on machine vision technique, treat that multispectral camera accomplishes the formation of image back of corresponding seed and carries out the communication with the PC and accomplish image acquisition, adopt machine vision and degree of depth learning algorithm to the appearance property of rice, carry out comprehensive analysis and judgement including color and luster, shape, size, texture etc. and realize rice accurate, preferably and hierarchical intelligently, compare with the general look selection machine on the market, the utility model discloses strong speciality, the screening degree of accuracy is high, reduces false positive rate and false negative rate simultaneously to help rice manufacturing enterprise and national grain depot to improve the purchase product quality and practice thrift the detection cost by a wide margin.
The utility model provides a rice quality detection system carries out comprehensive objective comprehensive analysis and judgement to imperfect rate, defect, the impurity rate etc. of rice through machine vision technique, adopts near infrared imaging technique to measure indexes such as moisture, mildening and rot in addition. The product can make up the blank of rice quality detection, realize high-speed and accurate quality grading of rice, improve the source quality of rice and greatly save the cost of enterprises.
The principle and the implementation of the present invention are explained herein by using specific examples, and the above description of the embodiments is only used to help understand the method and the core idea of the present invention; meanwhile, for the general technical personnel in the field, according to the idea of the present invention, there are changes in the concrete implementation and the application scope. In summary, the content of the present specification should not be construed as a limitation of the present invention.
Claims (5)
1. A rice phenotype monitoring system based on machine vision technology, comprising: the rice grain detecting device comprises a light shield, a material storage device, an image acquisition device and a PC (personal computer), wherein the light shield is arranged above the material storage device, the material storage device is used for placing a rice sample, the image acquisition device is arranged in the light shield, the image acquisition device is used for acquiring a phenotype image of the rice sample, and a light shield cover plate is arranged at the top of the light shield; the image acquisition device is electrically connected with the PC through a data line.
2. The rice phenotype monitoring system based on machine vision technology of claim 1, wherein the material storage device comprises a material drawer and a drawer guide rail, two parallel drawer guide rails are fixedly arranged at two opposite sides of the bottom of the light shield, the material drawer is slidably connected with the drawer guide rails through sliding blocks, and the material drawer is used for placing rice samples.
3. The rice phenotype monitoring system based on machine vision technology of claim 1, wherein the inner side wall of the light shield is fixedly connected with a fixed support, the image acquisition device is arranged on the fixed support, the image acquisition device comprises a multispectral camera and an LED annular light source, and the multispectral camera and the LED annular light source are respectively fixed above and below the fixed support; the multispectral camera is electrically connected with the PC through a data line.
4. A rice phenotype monitoring system based on machine vision technology as claimed in claim 3, wherein the LED annular light source is arranged with LED lamp beads of different wave bands.
5. A rice phenotype monitoring system based on machine vision technology as claimed in claim 3, wherein the system further comprises an LED backlight panel disposed below the material storage device.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202120153243.4U CN214122027U (en) | 2021-01-20 | 2021-01-20 | Rice phenotype monitoring system based on machine vision technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202120153243.4U CN214122027U (en) | 2021-01-20 | 2021-01-20 | Rice phenotype monitoring system based on machine vision technology |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN214122027U true CN214122027U (en) | 2021-09-03 |
Family
ID=77495693
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202120153243.4U Active CN214122027U (en) | 2021-01-20 | 2021-01-20 | Rice phenotype monitoring system based on machine vision technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN214122027U (en) |
-
2021
- 2021-01-20 CN CN202120153243.4U patent/CN214122027U/en active Active
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111862064B (en) | Silver wire surface flaw identification method based on deep learning | |
| CN109269951B (en) | Image-based flotation tailing ash content, concentration and coarse particle content detection method | |
| Li et al. | Computer vision based system for apple surface defect detection | |
| CN110389127B (en) | System and method for identifying metal ceramic parts and detecting surface defects | |
| Omid et al. | An expert egg grading system based on machine vision and artificial intelligence techniques | |
| Dehrouyeh et al. | Grading and quality inspection of defected eggs using machine vision | |
| CN108181316B (en) | Bamboo strip defect detection method based on machine vision | |
| CN107437243B (en) | Tire impurity detection method and device based on X-ray image | |
| CN110276386A (en) | A kind of apple grading method and system based on machine vision | |
| CN111815564B (en) | Method and device for detecting silk ingots and silk ingot sorting system | |
| CN111047655B (en) | High-definition camera cloth defect detection method based on convolutional neural network | |
| CN110415181A (en) | Flue-cured tobacco RGB image intelligent recognition and grade determination method under a kind of open environment | |
| CN113505865B (en) | Sheet surface defect image recognition processing method based on convolutional neural network | |
| CN111667455A (en) | AI detection method for various defects of brush | |
| CN110705655A (en) | Tobacco leaf classification method based on coupling of spectrum and machine vision | |
| CN110403232A (en) | A kind of cigarette quality detection method based on second level algorithm | |
| CN111160451A (en) | Flexible material detection method and storage medium thereof | |
| Najeeb et al. | Dates maturity status and classification using image processing | |
| CN110736709A (en) | blueberry maturity nondestructive testing method based on deep convolutional neural network | |
| CN113221881B (en) | Multi-level smart phone screen defect detection method | |
| CN115731228B (en) | Gold-plated chip defect detection system and method | |
| CN112858311A (en) | Rice phenotype monitoring system and method based on machine vision technology | |
| Pandey et al. | Non-destructive quality grading of mango (Mangifera Indica L) based on CIELab colour model and size | |
| Chen et al. | A surface defect detection system for golden diamond pineapple based on CycleGAN and YOLOv4 | |
| Sola-Guirado et al. | A smart system for the automatic evaluation of green olives visual quality in the field |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| GR01 | Patent grant | ||
| GR01 | Patent grant |