NL2026437A - Deep convolutional neural network based high-throughput method for determining alkali spreading value of rice grain - Google Patents
Deep convolutional neural network based high-throughput method for determining alkali spreading value of rice grain Download PDFInfo
- Publication number
- NL2026437A NL2026437A NL2026437A NL2026437A NL2026437A NL 2026437 A NL2026437 A NL 2026437A NL 2026437 A NL2026437 A NL 2026437A NL 2026437 A NL2026437 A NL 2026437A NL 2026437 A NL2026437 A NL 2026437A
- Authority
- NL
- Netherlands
- Prior art keywords
- image
- rice
- neural network
- convolutional neural
- training
- Prior art date
Links
- 235000007164 Oryza sativa Nutrition 0.000 title claims abstract description 74
- 235000009566 rice Nutrition 0.000 title claims abstract description 74
- 235000013339 cereals Nutrition 0.000 title claims abstract description 59
- 239000003513 alkali Substances 0.000 title claims abstract description 35
- 238000000034 method Methods 0.000 title claims abstract description 32
- 238000013527 convolutional neural network Methods 0.000 title claims abstract description 29
- 230000007480 spreading Effects 0.000 title abstract description 37
- 240000007594 Oryza sativa Species 0.000 title 1
- 241000209094 Oryza Species 0.000 claims abstract description 73
- 238000006243 chemical reaction Methods 0.000 claims abstract description 51
- 238000012549 training Methods 0.000 claims abstract description 32
- 238000012360 testing method Methods 0.000 claims abstract description 25
- 238000013135 deep learning Methods 0.000 claims abstract description 6
- 238000000605 extraction Methods 0.000 claims abstract description 6
- 238000005259 measurement Methods 0.000 claims abstract description 6
- 230000004888 barrier function Effects 0.000 claims description 15
- 230000006870 function Effects 0.000 claims description 7
- 238000011160 research Methods 0.000 claims description 6
- 238000010200 validation analysis Methods 0.000 claims description 6
- 238000004458 analytical method Methods 0.000 claims description 5
- 238000003709 image segmentation Methods 0.000 claims description 4
- 238000012216 screening Methods 0.000 claims description 4
- 230000004069 differentiation Effects 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 239000006185 dispersion Substances 0.000 claims 2
- 239000003518 caustics Substances 0.000 claims 1
- 230000003247 decreasing effect Effects 0.000 claims 1
- 238000009413 insulation Methods 0.000 claims 1
- 238000001514 detection method Methods 0.000 abstract description 7
- 230000008569 process Effects 0.000 abstract description 5
- 238000012545 processing Methods 0.000 abstract 1
- 238000002955 isolation Methods 0.000 description 14
- 238000007789 sealing Methods 0.000 description 9
- HEMHJVSKTPXQMS-UHFFFAOYSA-M sodium hydroxide Substances [OH-].[Na+] HEMHJVSKTPXQMS-UHFFFAOYSA-M 0.000 description 7
- 238000010586 diagram Methods 0.000 description 5
- KWYUFKZDYYNOTN-UHFFFAOYSA-M Potassium hydroxide Chemical compound [OH-].[K+] KWYUFKZDYYNOTN-UHFFFAOYSA-M 0.000 description 4
- 238000011156 evaluation Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000002834 transmittance Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013524 data verification Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000001747 exhibiting effect Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000010561 standard procedure Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
-
- 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
-
- 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
- G06N3/08—Learning methods
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N2021/8466—Investigation of vegetal material, e.g. leaves, plants, fruits
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30128—Food products
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
Abstract
The present invention relates to a deep convolutional neural network based high-throughput method for determining an alkali spreading value of a rice grain, including performing an alkaline reaction of rice grains through a single-grain-single-grid, multi-split reaction plate, performing high-throughput collection; after image processing, performing feature extraction and classification using a CNN-based convolutional neural network image classifier, and carrying out training under specific conditions; based on model parameters obtained by deep learning of the data training set, performing machine recognition on the images of the test samples to obtain a level of alkali spreading value. Through the determination method of the present invention, detection error caused by manual measurement is reduced, and the specific reaction plate is used for testing, which can ensure that the rice grains may not drift during the test process, thereby improving the clarity of later observations, and increasing the accuracy of detection. Moreover, the assessment result is no longer directly related to the operator's personal understanding, work experience, personal status and the like, which reduces the difficulty of detection, and the test results are more accurate and representative.
Description
1 001635P-NL DEEP CONVOLUTIONAL NEURAL NETWORK BASED HIGH-THROUGHPUT
[0001] The invention relates to the field of rice quality inspection, in particular to a deep convolutional neural network based high-throughput method for determining the alkali spreading value of rice grains.
[0002] In evaluation of rice quality, alkali spreading value is the main index. According to the standard determination method of alkali spreading value, 6 full and complete rice grains are placed in the same special box, to which potassium hydroxide is added to react for 23 hours, and then the rice grains are observed for spreading degree, rated as levels 1-7. For example, at level 1, there is no change in the rice grains; at level 2, the rice grains swell; and at level 4, the rice grains swell and the ring is complete and wide. But in the actual operation process, the test efficiency and test accuracy of the standard determination method have certain defects. First of all, the test should be carried out cautiously and slowly, because the rice grains float in the reaction solution, and a slight vibration can easily cause grain movement, making the rice grains shield each other, which makes the later observation difficult. Secondly, the level of alkali spreading value is related to the professional level of the tester. Although there are specific text descriptions, in actual operation, the evaluation results are related to the operator's personal understanding, work experience, personal status, and environmental conditions at the time.
[0003] In order to solve the above-mentioned technical problems, the present invention adopts the following solutions:
[0004] A deep convolutional neural network based high-throughput method for
2 001635P-NL determining an alkali spreading value of a rice grain, comprising the following steps:
[0005] 1) putting a full and complete grain of rice into each grid in a reaction box, adding lye slowly along an outer wall of the reaction box until the lye covers all of a reaction plate, and then sealing the reaction box with a sealing cover and placing the sealed reaction box into an oven to react for 23 hours;
[0006] 2) placing the reaction plate after reaction into an image collection device for high-throughput collection, and setting a collection pixel based on a model training result; after obtaining images, rating, by a professional alkali spreading value tester, a spreading degree of the rice gain in each grid to construct an original data set;
[0007] 3) screening rice samples of grades 1-7 of different rice varieties from the annual measurement and analysis sample library for rice quality research of the China National Rice Research Institute, testing and sampling as described in the step 1) to the step 2) to obtain big data samples for subsequent image recognition;
[0008] 4) performing image segmentation to convert the image into a grayscale image, performing a Gaussian blur operation for noise removal, and then performing binarization; next, finding out a contour of the image through OpenCV built-in function, and then finding out an enclosing rectangle of each contour using OpenCV built-in function and screening a rectangle conforming with the segmented image in size;
[0009] 5) performing, by an image classifier based on convolutional neural network (CNN), feature extraction and classification using ResNet18, an 18-layer convolutional network, or using ResNet50, a 50-layer convolutional network; inputting an image that is a 3-channel (i.e. RGB) image with a size of 224*224; and using Pytorch to transform an image if an original size of the image is not 224*224;
[0010] 6) setting training hyper-parameters, wherein a learning cycle comprises 90 epochs, and each epoch is one cycle going through a full training set;
[0011] setting a learning rate (LR) to 0.1 initially, and then reducing the learning rate to 1/10 of the original value every 30 epochs;
[0012] setting batch size to 8;
[0013] setting regularization weight to 10% to prevent overfitting where the
3 001635P-NL training set performs well but the test set performs poorly;
[0014] 7) recording data as training sets and validation sets respectively during training, wherein the number of training sets is 4-10 times that of validation sets, and the two cannot overlap; obtaining files of image mean and standard deviation of the data sets, wherein the image mean and the standard deviation are required for normalization of image preprocessing during training; and using the ResNet18 or ResNet50 structure model for testing and differentiation after training;
[0015] 8) based on model parameters obtained by deep learning of the data training set, performing machine recognition on the images of the test samples to obtain the level of the alkali spreading value.
[0016] The deep convolutional neural network based high-throughput method for determining the alkali spreading value of the rice grain is characterized in that in the step 1) the rice grain is submerged in the dye, and a height of the lye does not exceed a height of the grid divided by isolation barriers in the reaction box.
[0017] The deep convolutional neural network based high-throughput method for determining the alkali spreading value of the rice grain is characterized in that a collector of the collection device in the step 2) comprises a scanner, an industrial vertical camera, a preliminarily fixed mobile phone camera; and the collection pixel comprises a positive film or negative film mode.
[0018] The deep convolutional neural network based high-throughput method for determining the alkali spreading value of the rice grain is characterized in that, 1000-5000 samples are used in the step 2).
[0019] The deep convolutional neural network based high-throughput method for determining the alkali spreading value of the rice grain is characterized in that the reaction box includes a reaction plate, an isolation barrier and a sealing cover arranged above the isolation barrier and fit with the reaction plate.
[0020] The deep convolutional neural network based high-throughput method for determining the alkali spreading value of the rice grain is characterized in that a plurality of small grids are arranged in the reaction plate through the isolation barrier, andthe small grid is 20-25 mm long, 20-25 mm wide, and 3 mm high.
4 001635P-NL
[0021] The deep convolutional neural network based high-throughput method for determining the alkali spreading value of the rice grain is characterized in that the height of the isolation barrier is 1-2 cm lower than the height of the reaction plate.
[0022] The deep convolutional neural network based high-throughput method for determining the alkali spreading value of the rice grain is characterized in that the number of grids divided by the isolation barrier per plate in the reaction plate is generally 50-200, and each grid is marked with a serial number.
[0023] The deep convolutional neural network based high-throughput method for determining the alkali spreading value of the rice grain is characterized in that the materials used for the reaction plate, the isolation barrier and the sealing cover include PE and PP, and the bottom light transmittance is greater than 80%.
[0024] The deep convolutional neural network based high-throughput method for determining the alkali spreading value of the rice grain is characterized in that the shape of the reaction plate includes a circle, a square, and an ellipse.
[0025] The deep convolutional neural network based high-throughput method for determining the alkali spreading value of the rice grain has the following beneficial effects:
[0026] Through the determination method of the present invention, detection error caused by manual measurement is reduced, and the specific reaction plate is used for testing, which can ensure that the rice grains may not drift during the test process, thereby improving the clarity of later observations, and increasing the accuracy of detection. Moreover, the assessment result is no longer directly related to the operator's personal understanding, work experience, personal status and the like, which reduces the difficulty of detection, and the test results are more accurate and representative.
[0027] FIG. 1 shows a novel reaction box used in the present invention;
[0028] FIG. 2 shows rice grain alkali spreading value collection images;
[0029] FIG. 3 is a schematic diagram of binarization of rice grain alkali
001635P-NL spreading value images;
[0030] FIG. 4 is a schematic diagram of contour segmentation of rice grain alkali spreading value images;
[0031] FIG. 5 is a structure diagram of ResNet18 and ResNet50; 5 [0032] FIG. 6 shows images for level-4 alkali spreading value of rice grains;
[0033] FIG. 7 shows images for level-5 alkali spreading value of rice grains;
[0034] In the drawings, 1 represents reaction plate; 2 represents isolation barrier; and 3 represents sealing cover.
[0035] The specific embodiments of the present invention will be described in further detail below in conjunction with the accompanying drawings to make the technical solution of the present invention easier to understand and grasp.
[0036] With reference to FIG. 1, in order to meet the needs of high-throughput sampling, the existing test system where 6 rice grains are placed in one reaction box is improved into a new type of reaction box, which includes a reaction plate 1, isolation barriers 2 disposed in the reaction plate 1, and a sealing cover 3 disposed on the top of the reaction plate 1. The isolation barriers 2 are sealed in the reaction plate 1 by the sealing cover 3.
[0037] Among them, the materials used for the reaction plate 1, the isolation barrier 2 and the sealing cover 3 include PE, PP and the like, and the bottom light transmittance is greater than 90%. The isolation barriers 2 are interlaced to result in a number of small grids to form the reaction plate. The small grids are 20-25 mm long, 20-25 mm wide, and 3 mm high. The height of the isolation barrier 2 is 1-2 cm lower than the height of the reaction plate 1. The reaction plate 1 has a shape of circle, square or the like, and can be designed to have various shapes according to different requirements. The size of the reaction plate 1 varies according to different collection methods. The number of grids divided by the isolation barriers 2 per plate is generally 50-200 in the reaction plate 1, and each grid is marked with a serial number.
[0038] A deep convolutional neural network based high-throughput method for
6 001635P-NL determining an alkali spreading value of a rice grain, comprising the following steps:
[0039] According to the standard method, a full and complete grain of rice is placed in each grid of the reaction plate 1 used in the present invention; lye (i.e. the reaction solution) is slowly added along the outer wall so that the lye gradually fills the entire reaction plate through the bottom, wherein the lye is a 0.304 mol/L potassium hydroxide solution. The height of the lye solution is required to submerge the rice grains, but not higher than the height of a single grid, so as to ensure that the grains fully contact the reaction solution without drifting or moving during the movement of the reaction box. After adding the solution, the reaction box is covered and sealed with the sealing cover, and then placed in an oven at 30°C to react for 23 hours.
[0040] After 23 hours of reaction, the reaction plate is placed in the image collection device for high-throughput collection. The collector of the collection device can be a scanner, an industrial vertical camera, a mobile phone camera which is initially fixed. The collection pixels are set according to the training results of the model, and can be in a positive film or negative film mode, with the main purpose of highlighting the color difference. The actual effect is shown in FIG. 2. After obtaining the images, professional alkali spreading value testers from the rice quality research center of China National Rice Research Institute are invited to rate the spreading degree of the rice grain in each single grid to construct the original data set.
[0041] Totally 1000-5000 rice samples of grades 1-7 of different varieties are screened from the annual measurement and analysis sample library for rice quality research of the China National Rice Research Institute. According to the above process, the samples are measured and sampled to obtain big data samples for subsequent image recognition.
[0042] Image segmentation: First, the image is converted into a grayscale image, a Gaussian blur operation is then performed to remove noise, and then binarization is performed. The result is shown in FIG. 3. Then, the contour of the image is found out through OpenCV built-in function, thereafter an enclosing rectangle of each contour is found out using OpenCV built-in function, and a rectangle conforming with the segmented image in size is screened. All rectangles and the final
7 001635P-NL result are shown in FIG. 4.
[0043] Deep learning: The present invention uses an image classifier based on Convolutional Neural Network (CNN). The architecture used is ResNet (Residual Network) which has the strongest feature extraction capability at present. This residual learning network can be as deep as more than 1000 layers while still exhibiting excellent feature extraction capability, far superior over the previous excellent deep learning structures including VGG. At present, ResNet18, an 18-layer convolutional layer network, is temporarily used for feature extraction and classification. The structure diagrams of ResNet18 and ResNet50 are shown in FIG. 4 below. The input image is a 3-channel (RGB) image with a size of 224*224, and Pytorch is used to transform an image if an original size of the image is not 224*224. Pytorch is a scientific computing tool based on python. In this method, Pytorch built-in functions are called to process images. The structure diagrams of ResNet18 and ResNet50 are shown in FIG. 5.
[0044] Training hyper-parameter settings:
[0045] Learning cycle (epoch): 90 epochs, with each epoch going through a full training set.
[0046] Learning rate (LR): It is initially set to 0.1, and then reduced to 1/10 of the original value every 30 epochs.
[0047] Batch Size: There are 7 categories in total, so the batch size is set to 8.
[0048] Weight decay: It is set to 10 to prevent overfitting where the training set performs well but the test set performs poorly.
[0049] The data are recorded as training sets and validations set during training respectively. The number of training sets is 4-10 times that of validation sets, and the two cannot overlap. Files of image mean and standard deviation of the data sets are obtained. The image mean and the standard deviation are required for normalization of image preprocessing during training. After training, the model is used for testing and differentiation.
[0050] Result judgment. Based on the model parameters obtained by deep learning of the data training set, machine recognition is performed on the images of
8 001635P-NL the test samples to obtain the level of the alkali elimination value. The alkali spreading value classification standard is shown in Table 1.
Table 1 Alkali spreading value classification standard Spreading degree Clarity No change in rice grain White rice grain core 9 Swollen rice grain White rice grain core, with powdery ring 3 Swollen rice grain, with | White rice grain core, with incomplete or narrow ring | flocculent or nebulous ring 4 Swollen rice grain, with | Cotton-white rice grain complete and wide ring core, with nebulous ring Cracked rice grain, with | Cotton-white rice grain complete and wide ring core, with clear ring Partially dispersed and | Cloud-white rice grain core, dissolved rice grain, | with ring disappeared merged with the ring 7 Completely dispersed rice | Both rice grain core and grain ring disappear 5 [0051] Test 1:
[0052] Using samples with an alkali spreading value of levels 4-7 as test samples, the recognition accuracy of the method of the present invention was verified, and the results are as follows:
[0053] (1) Obtaining images through the scanner; (2) Image segmentation to obtain single-grid images; (3) Determining the alkali spreading value of each rice grain using the CNN-based ResNet model; (4) Determining the alkali spreading value by professionals for data verification.
[0054] Final results: Recognition accuracy
91.67%
92.31% 6 |100% 100% Overall recognition rate 96.15%
[0055] Result analysis: The overall recognition rate reached 96.15%, and the recognition rate in the case of level 4 and level 5 was only about 92%. Specific analysis found that an image that
9 001635P-NL was artificially judged to be level 4 was recognized as level 5 by the machine.
The two images are shown in FIG. 6 and FIG. 7 respectively.
Since level 4 and level 5 are relatively close in performance, even professional technicians think that level 4 and level 5 are quite difficult to distinguish and can be interchanged.
Therefore, the results this time are relatively excellent, reflecting the effectiveness of the ResNet deep learning model.
The evaluation results no longer have a direct relationship with the operator's personal understanding, work experience, personal status and the like, making the test results more accurate and reducing detection errors caused by artificial measurement.
Claims (4)
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010129034.6A CN111289512B (en) | 2020-02-28 | 2020-02-28 | Rice grain alkali elimination value high-throughput determination method based on deep convolutional neural network |
| NL2026437A NL2026437B1 (en) | 2020-02-28 | 2020-09-10 | Deep convolutional neural network based high-throughput method for determining alkali spreading value of rice grain |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010129034.6A CN111289512B (en) | 2020-02-28 | 2020-02-28 | Rice grain alkali elimination value high-throughput determination method based on deep convolutional neural network |
| NL2026437A NL2026437B1 (en) | 2020-02-28 | 2020-09-10 | Deep convolutional neural network based high-throughput method for determining alkali spreading value of rice grain |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| NL2026437A true NL2026437A (en) | 2020-10-08 |
| NL2026437B1 NL2026437B1 (en) | 2021-10-04 |
Family
ID=78595201
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| NL2026437A NL2026437B1 (en) | 2020-02-28 | 2020-09-10 | Deep convolutional neural network based high-throughput method for determining alkali spreading value of rice grain |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN111289512B (en) |
| NL (1) | NL2026437B1 (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116994032A (en) * | 2023-06-28 | 2023-11-03 | 河北大学 | Rectal polyp multi-classification method based on deep learning |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111289512B (en) * | 2020-02-28 | 2021-04-13 | 中国水稻研究所 | Rice grain alkali elimination value high-throughput determination method based on deep convolutional neural network |
| CN112369322B (en) * | 2020-11-12 | 2023-10-03 | 安徽荃银高科种业股份有限公司 | Method for identifying high-quality rice variety in high temperature period by using alkali extinction value |
| CN112369319B (en) * | 2020-11-12 | 2023-10-03 | 安徽荃银高科种业股份有限公司 | Method for breeding high-quality rice varieties in high temperature period based on alkali extinction value |
| CN116149280B (en) * | 2023-04-04 | 2023-07-07 | 福建德尔科技股份有限公司 | Intelligent production system of electronic-grade potassium hydroxide |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111289512A (en) * | 2020-02-28 | 2020-06-16 | 中国水稻研究所 | Rice grain alkali elimination value high-throughput determination method based on deep convolutional neural network |
| WO2020172558A1 (en) * | 2019-02-21 | 2020-08-27 | The Trustees Of Dartmouth College | System and method for automatic detection of vertebral fractures on imaging scans using deep networks |
Family Cites Families (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7964536B2 (en) * | 2001-12-04 | 2011-06-21 | Wayne State University | Neoepitope detection of disease using protein arrays |
| CN102759527A (en) * | 2011-04-28 | 2012-10-31 | 河南工业大学 | Method for detecting alkali digestion degree of rice |
| CN102788752B (en) * | 2012-08-08 | 2015-02-04 | 江苏大学 | Non-destructive detection device and method of internal information of crops based on spectrum technology |
| CN105067784B (en) * | 2015-07-22 | 2016-08-24 | 中国水稻研究所 | Utilize the physicochemical character index of rice quality to the method judging rice eating-quality |
| CN104992068B (en) * | 2015-08-13 | 2018-10-26 | 四川农业大学 | A kind of prediction technique of topsoil nitrogen distribution |
| CN206788149U (en) * | 2017-06-08 | 2017-12-22 | 中国水稻研究所 | A kind of overall caustic SCC determination box |
| CN107748144A (en) * | 2017-11-13 | 2018-03-02 | 中国科学院昆明植物研究所 | The middle infrared spectrum detecting system of quick measure SOIL CARBON AND NITROGEN and its stable isotope |
| CN108959661B (en) * | 2018-08-20 | 2021-08-06 | 安徽大学 | 3S technology-based soil nutrient grade classification map generation method and precision evaluation method thereof |
| CN109934297B (en) * | 2019-03-19 | 2023-04-18 | 广东省农业科学院农业生物基因研究中心 | Rice seed test method based on deep learning convolutional neural network |
| CN109949323B (en) * | 2019-03-19 | 2022-12-20 | 广东省农业科学院农业生物基因研究中心 | Crop seed cleanliness judgment method based on deep learning convolutional neural network |
| CN110517311A (en) * | 2019-08-30 | 2019-11-29 | 北京麦飞科技有限公司 | Pest and disease monitoring method based on leaf spot lesion area |
-
2020
- 2020-02-28 CN CN202010129034.6A patent/CN111289512B/en active Active
- 2020-09-10 NL NL2026437A patent/NL2026437B1/en not_active IP Right Cessation
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020172558A1 (en) * | 2019-02-21 | 2020-08-27 | The Trustees Of Dartmouth College | System and method for automatic detection of vertebral fractures on imaging scans using deep networks |
| CN111289512A (en) * | 2020-02-28 | 2020-06-16 | 中国水稻研究所 | Rice grain alkali elimination value high-throughput determination method based on deep convolutional neural network |
Non-Patent Citations (2)
| Title |
|---|
| MARIOTTI M ET AL: "Alkali spreading value and Image Analysis", JOURNAL OF CEREAL SCIENCE, ACADEMIC PRESS LTD, GB, vol. 52, no. 2, 1 September 2010 (2010-09-01), pages 227 - 235, XP027273897, ISSN: 0733-5210, [retrieved on 20100619] * |
| TUAÑO ARVIN PAUL P ET AL: "Estimation of Alkali Spreading Value and Gelatinization Temperature of Some Philippine Rice Varieties Using Digital Photometry", 31 December 2018 (2018-12-31), XP055799111, Retrieved from the Internet <URL:https://www.researchgate.net/profile/Arvin-Paul-Tuano/publication/330193474_Estimation_of_Alkali_spreading_value_and_gelatinization_temperature_of_some_Philippine_rice_varieties_using_digital_photometry/links/5c3c6c9592851c22a373730c/Estimation-of-Alkali-spreading-value-and-gelatinization-temperatur> [retrieved on 20210427] * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116994032A (en) * | 2023-06-28 | 2023-11-03 | 河北大学 | Rectal polyp multi-classification method based on deep learning |
| CN116994032B (en) * | 2023-06-28 | 2024-02-27 | 河北大学 | Rectal polyp multi-classification method based on deep learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111289512A (en) | 2020-06-16 |
| CN111289512B (en) | 2021-04-13 |
| NL2026437B1 (en) | 2021-10-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| NL2026437B1 (en) | Deep convolutional neural network based high-throughput method for determining alkali spreading value of rice grain | |
| US11436493B2 (en) | Chromosome recognition method based on deep learning | |
| CN109523520B (en) | Chromosome automatic counting method based on deep learning | |
| CN104730091B (en) | The extraction of gas turbine blades defect and analysis method based on region segmentation detection | |
| CN110376198A (en) | A kind of cervical liquid-based cells chipping qualities detection system | |
| CN111784633A (en) | Insulator defect automatic detection algorithm for power inspection video | |
| CN112580748B (en) | Method for counting classified cells of stain image | |
| CN109872335A (en) | A kind of automatic read tablet method and its system for PD-L1 antibody stained slice | |
| CN113505865A (en) | Sheet surface defect image recognition processing method based on convolutional neural network | |
| CN111598849B (en) | Pathological image cell counting method, equipment and medium based on target detection | |
| WO2021109011A1 (en) | Intelligent capacitor internal defect detection method based on ultrasound image | |
| CN108960413A (en) | A kind of depth convolutional neural networks method applied to screw surface defects detection | |
| Haoran et al. | Egg crack detection based on support vector machine | |
| Vaishak et al. | Currency and Fake Currency Detection using Machine Learning and Image Processing–An Application for Blind People using Android Studio | |
| CN112613505A (en) | Cell micronucleus identification, positioning and counting method based on deep learning | |
| WO2022267799A1 (en) | Water quality testing method and water quality testing apparatus | |
| Rodriguez et al. | Segmentation and error detection of PV modules | |
| Guan et al. | Ceramic ring defect detection based on improved YOLOv5 | |
| CN115937195A (en) | Surface defect detection method and device and computer equipment | |
| CN115546099A (en) | Forging flaw detection defect detection method and device based on convolutional neural network | |
| CN115690001A (en) | Method for detecting defects in steel pipe welding digital radiographic image | |
| CN112581424B (en) | Classification extraction method for surface and subsurface defects of optical element | |
| CN112508946B (en) | Cable tunnel anomaly detection method based on antagonistic neural network | |
| CN113808079B (en) | Industrial product surface defect self-adaptive detection method based on deep learning model AGLNet | |
| CN117576093B (en) | EL image-based solar cell piece hidden crack detection method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| MM | Lapsed because of non-payment of the annual fee |
Effective date: 20231001 |