WO2020206963A1 - 一种杂交秋葵种子分类鉴别装置和方法 - Google Patents

一种杂交秋葵种子分类鉴别装置和方法 Download PDF

Info

Publication number
WO2020206963A1
WO2020206963A1 PCT/CN2019/110493 CN2019110493W WO2020206963A1 WO 2020206963 A1 WO2020206963 A1 WO 2020206963A1 CN 2019110493 W CN2019110493 W CN 2019110493W WO 2020206963 A1 WO2020206963 A1 WO 2020206963A1
Authority
WO
WIPO (PCT)
Prior art keywords
hybrid
okra
seeds
seed
movable claw
Prior art date
Application number
PCT/CN2019/110493
Other languages
English (en)
French (fr)
Inventor
何勇
张金诺
冯旭萍
聂鹏程
Original Assignee
浙江大学
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 浙江大学 filed Critical 浙江大学
Priority to CH000718/2020A priority Critical patent/CH716240B9/de
Publication of WO2020206963A1 publication Critical patent/WO2020206963A1/zh

Links

Images

Classifications

    • 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
    • G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
    • G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
    • G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
    • 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
    • G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
    • G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • 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
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2201/00—Features of devices classified in G01N21/00
    • G01N2201/12—Circuits of general importance; Signal processing
    • G01N2201/129—Using chemometrical methods
    • G01N2201/1296—Using chemometrical methods using neural networks

Definitions

  • the invention relates to the technical field of agricultural hybrid breeding, in particular to a device and method for classifying and identifying hybrid okra seeds.
  • Cross breeding is an effective way to obtain new crop varieties by selecting and breeding two or more varieties with good traits.
  • hybrid rice plays a pivotal role in food supply, accounting for about 60% of the total rice planting area.
  • Cross breeding recombines the genes of the parents to form different varieties, providing a wealth of material classification.
  • the obtained hybrid seeds have genetic diversity, which can improve the yield and quality of hybrid plants.
  • Okra (Abelmoschusesculentus (L.) Moench) has rich fiber content and medicinal value, and has always been regarded as a very nutritious vegetable crop. Okra seeds are also rich in antioxidant compounds that are beneficial to health, and have anti-fatigue and anti-aging effects.
  • Cross breeding helps to improve the yield and quality of the above two crops. However, the offspring that produce mutations during the cross breeding process need to be further cultivated into plants and display certain traits before they can be accurately identified. By changing the leaf width, plant height or color of hybrid plants, breeding experts can select plants that meet the requirements and perform genetic operations such as selfing to produce hybrid seeds. This process is often time-consuming and labor-intensive.
  • hyperspectral imaging technology has been widely used in the fields of seed identification, seed quality inspection and food inspection.
  • the molecular overtone transitions and combined vibrations of the test sample are the basis of near-infrared spectroscopy.
  • the use of hyperspectral imaging technology can obtain detailed information on the composition and characteristics of the sample at the molecular level.
  • Near-infrared hyperspectral imaging technology is a non-destructive and rapid method to obtain the space and spectrum information of the sample in the range of 780-2500nm.
  • the current seed identification devices using hyperspectral imaging technology are all large-scale laboratory spectroscopy equipment, and the spectrum identification of seeds needs to be performed in the laboratory, which greatly reduces the portability.
  • the purpose of the present invention is to provide a hybrid okra seed classification and identification device, which can measure the hybrid seeds in the field without damage and on-site measurement, which greatly meets the needs of hybrid breeding and reduces the labor burden in the hybrid breeding process.
  • Another object of the present invention is to provide a hybrid okra seed classification and identification method, which is based on the above hybrid okra seed classification and identification device, solves the complex selection process of hybrid breeding, and quickly and conveniently meets the needs of identifying different varieties of hybrid autumn Demand for sunflower seeds.
  • the hybrid okra seed classification and identification device includes a device main body and a clamping mechanism arranged on the device main body;
  • the device main body includes a fixed claw fixed on the workbench and a hinge mechanism and a fixed claw Articulated movable claw;
  • the clamping mechanism includes a base connected to the fixed claw and an installation press table connected to the movable claw.
  • the installation press table is provided with a light source for irradiating hybrid okra seeds and a device for collecting image information
  • a memory and a processor for processing image information are also arranged on the main body of the device, and the image information and a hybrid okra seed identification model are stored in the memory.
  • the hybrid okra seeds to be tested are clamped in the clamping mechanism, the light source is used to illuminate the area to be tested of the hybrid seeds, and the CCD camera and imaging spectrometer are used to obtain the spatial information and spectral information of the hybrid seeds.
  • the spectral image information is processed by the processor to process the spectral data of the area to be measured.
  • the okra seed identification model uses the spectral data to identify the type of hybrid okra seed and output and display it.
  • the portable measuring device can measure the hybrid okra seeds in the field or in the laboratory without damage and on-site measurement, which greatly meets the needs of hybrid breeding and reduces the labor burden in the hybrid breeding process.
  • a driving button for controlling the opening of the movable claw is provided on the main body of the device, and a first return spring is provided at the bottom of the driving button.
  • the hinge mechanism includes a rotating shaft and a disk fixed on the rotating shaft, and also includes a connecting rod hinged between the disk and the button, and the disk is also hinged with the end of the movable claw.
  • the connecting rod drives the disc to rotate, so that the movable claw moves upward, and the installation press table is opened.
  • hybrid okra seeds can be placed on the base.
  • the button is released, under the action of the first return spring, the button is reset, and the mounting and pressing table moves down and presses and fixes the hybrid okra seeds.
  • a second return spring is provided between the movable claw and the fixed claw.
  • the top of the base is provided with a groove for placing hybrid okra seeds, and the bottom is provided with a buffer spring.
  • the depth of the groove should not be greater than the particle size of hybrid okra seeds.
  • the buffer spring is used to avoid crushing the seeds.
  • the clamping mechanism is provided with a switch that triggers the operation of the light source, the camera and the spectrometer, and a synapse for pressing the switch is fixed above the switch.
  • a switch that triggers the operation of the light source, the camera and the spectrometer
  • a synapse for pressing the switch is fixed above the switch.
  • the spectrometer is an instrument that acquires image information in a multi-channel, continuous and high spectral resolution manner. It can obtain dozens of simultaneous acquisitions of the same area by organically combining traditional spatial imaging technology with ground object spectroscopy technology. Reflectance spectrum images of ground objects in several to several hundred bands. It can also quickly measure and analyze the spectral composition of multiple objects simultaneously within the field of view of the instrument. These characteristics are helpful to be used in the classification and identification device of portable hybrid okra seeds.
  • the spectrometer collects spectral information of 200 bands in the range of 874 to 1734 nm.
  • the hybrid okra seed identification model is obtained by the following method:
  • the average spectral data of the area to be tested is calculated through the spectral data of the pixel points of the hybrid seed's area to be tested. Then the average spectral data is sent to the trained convolutional neural network model, and the type of seed to be tested is judged by the model data and displayed on the LCD screen.
  • the convolutional neural network includes two convolutional layers, two pooling layers and four fully connected layers; the filters of the two convolutional layers are set to 32 and 64 respectively; used to process the average of hybrid okra seeds
  • the convolution kernel of the spectral data is a 1x3 one-dimensional convolution kernel, and the compensation and filling of the convolution kernel is set to 1; the kernel size of the maximum pooling layer is 1x3, and the stride is 1; the number of neurons in the four-layer fully connected layer They are 512, 256, 128 and 64 respectively.
  • step b the method of small batch training is adopted, the batch number is selected as 25, and the number of training times is 1500.
  • a batch normalization and discarding method are added between each fully connected layer to prevent overfitting, and the discarding probability of the discarding method is 0.5.
  • the hybrid okra seed classification and identification method includes the following steps:
  • the imaging spectrometer used to obtain the spectral data of hybrid okra seeds collects spectral information of 200 bands in the range of 874 to 1734 nm.
  • step 3 the sample hybrid seed is placed in the area to be tested for collection, and the optical system is used to collect the seed spectral information, and the number of spectral information of the pixel points of the hybrid seed is obtained as L and the sum of spectral data is Z.
  • the sum of the spectral data of the pixels collected by the imaging spectrometer is Z, the number of pixels is L, and the uncorrected average spectrum P is:
  • the device for classifying and identifying different types of hybrid okra seeds and the identification method based on the device of the present invention can measure the hybrid okra seeds without damage and quickly in the field or in the laboratory, which greatly meets the needs of hybrid breeding and breeding seeds. It solves the time-consuming and labor-consuming problem in the process of hybrid breeding, and provides a very important means for speeding up the experiment process of hybrid breeding.
  • Figure 1 is a schematic structural diagram of a hybrid okra seed classification and identification device in an embodiment of the present invention
  • FIG. 2 is a flowchart of a method for classifying and identifying hybrid okra seeds in an embodiment of the present invention
  • Fig. 3 is an average spectrum diagram of hybrid okra seeds obtained by using the hybrid okra seed classification and identification method in the embodiment of the present invention
  • Fig. 4 is a before and after comparison diagram of the classification effect diagram obtained by the hybrid okra seed classification and identification method in the embodiment of the present invention.
  • the hybrid okra seed classification and identification device of this embodiment includes a device main body and a clamping mechanism 17 arranged on the device main body.
  • the main body of the device includes a casing 14 and a controller 10 arranged in the casing 14.
  • a battery 11 is also provided in the casing 14.
  • the controller 10 is connected with a data memory 12 and is connected with a liquid crystal display for receiving and displaying measurement results. 13.
  • the device body also includes a fixed claw 15, a movable claw 4, an excitation light assembly and a photoelectric conversion assembly. Between the fixed pawl 15 and the movable pawl 4, a clamping mechanism 17 matched with the hybrid okra is provided, and a push switch 6 is provided.
  • the fixed claw 15 is connected to the housing 14 as a whole.
  • the fixed claw 15 is provided with a buffer spring 16 to keep the clamping mechanism 17 always in the working position and avoid crushing the seeds.
  • the movable claw 4 is provided with a sliding groove on the inner side, and the fixed claw 15 can slide up and down in the sliding groove, which makes the clamping jaws more flexible. Even if hybrid okra seeds are not put in, the whole device will not be disturbed by the outside. When the hybrid okra seeds are clamped, the push switch 6 is triggered to control the operation of the excitation light assembly and the photoelectric conversion assembly.
  • the clamping mechanism 17 includes a base connected to the fixed claw 15 and a mounting press table connected to the movable claw 4.
  • a button 7 is provided on one side of the housing 14.
  • the button 7 controls the movement of the movable claw through a hinge mechanism.
  • the hinge mechanism includes a rotating shaft and a disk 9 fixed on the rotating shaft, and also includes a connecting rod hinged between the disk 9 and the button 7, the disk 9 is also hinged with the end of the movable claw 4.
  • Two first return springs 8 are provided under the button 7. After the button is pressed, the clamping jaw is controlled to open by the hinge mechanism. After the button is released, the original position is restored under the action of the first return spring 8 and the inserted hybrid okra seeds are clamped.
  • the middle of the movable claw 4 is provided with a second return spring 5, which makes the entire clamping mechanism more stable.
  • the excitation light assembly of this embodiment is a light source 1 arranged on a mounting platform of the movable claw 4.
  • the photoelectric conversion component includes a CCD camera 2 and an imaging spectrometer 3 arranged on the mounting platform.
  • the light generated by the light source 1 illuminates the surface of the hybrid okra seed to be tested, and then the spatial information and spectral information reflected by the surface enter the CCD camera 2 and the imaging spectrometer 3 and transmit the collected spectral information to the controller 10.
  • the controller 10 processes the spectrum information to obtain the pixel point spectrum information of the hybrid seed, and calculates the average spectrum information of the hybrid seed through the pixel point spectrum information of the hybrid seed.
  • the data memory 12 stores the average spectral information of the hybrid seeds.
  • the data transmission element transmits the average spectral data measured by the measuring device to the controller and combines the output result of the deep learning discriminant model.
  • the controller 10 sends the detection result to the liquid crystal display 13 and then the liquid crystal The display shows.
  • the battery 11, the excitation light assembly, the photoelectric conversion assembly, and the push switch 6 are respectively connected to the controller 10 in the body through wires.
  • the excitation light source illuminates the surface of the hybrid okra seeds to be tested, and the reflection spectrum information and image information of the hybrid seeds are obtained by using a CCD camera and an imaging spectrometer;
  • the hybrid okra seed classification and identification device of this embodiment includes a device main body and a clamping mechanism 17 arranged on the device main body.
  • the main body of the device includes a casing 14 and a controller 10 arranged in the casing 14.
  • a battery 11 is also provided in the casing 14.
  • the controller 10 is connected with a data memory 12 and is connected with a liquid crystal display for receiving and displaying measurement results. 13. It can be understood that, as shown in FIG. 2, the battery 11 is directly or indirectly connected to the power-consuming units in the main body of the device, such as the data memory 12, the liquid crystal display 13, the controller 10, etc., so as to supply these power-consuming units. Electrical energy.
  • the device body also includes a fixed claw 15, a movable claw 4, an excitation light assembly and a photoelectric conversion assembly.
  • a clamping mechanism 17 matched with hybrid okra seeds is arranged between the fixed claw and the movable claw 4, and a push switch 6 is provided.
  • the fixed claw 15 is integrated with the housing.
  • the fixed claw 15 is provided with a buffer spring 16, one end of the buffer spring 16 is fixed to the fixed claw 15 and the other end is connected with the clamping mechanism 17 to keep the clamping mechanism always in the working position.
  • the elasticity of the buffer spring 16 enables the clamping mechanism 17 to maintain elastic contact with the upper and lower clamping jaws when clamping hybrid okra seeds, so as to avoid crushing the okra seeds to be tested in the clamping mechanism 17.
  • the movable claw 4 is provided with a sliding groove on the inner side, and the fixed claw 15 can slide up and down in the sliding groove, which makes the clamping jaws more flexible. Even if hybrid okra seeds are not put in, the whole device will not be disturbed by the outside. When the hybrid okra seeds are clamped, the push switch 6 is triggered to control the operation of the excitation light assembly and the photoelectric conversion assembly.
  • the clamping mechanism 17 includes a base connected to the fixed claw 15 and a mounting press table connected to the movable claw 4.
  • a button 7 is provided on one side of the housing 14, and the button 7 controls the movement of the movable pawl 4 through a hinge mechanism.
  • the hinge mechanism includes a rotating shaft and a disk 9 fixed on the rotating shaft, and also includes a connecting rod hinged between the disk 9 and the button 7, the disk 9 is also hinged with the end of the movable claw 4.
  • Two first return springs 8 are provided under the button 7. After the button 7 is pressed, the pressed button 7 activates the disc 9 through a connecting rod.
  • the disk 9 fixed to the rotating shaft thus rotates along with the rotating shaft.
  • the rotating disc 9 drives the movement of the movable pawl 4, thereby controlling the opening of the clamping pawl through a hinge mechanism.
  • a first return spring 8 is provided under the button 7. One end of the first return spring 8 is connected to the lower part of the button 7, and the other end is fixedly connected to the housing 14, so that after the button 7 is released, the button 7 can return to its original position under the action of the first return spring 8.
  • the returning button 7 resets the movable pawl 4 through the hinge structure, so that the clamping pawl is closed again, and then the inserted hybrid okra seeds are clamped.
  • a second return spring 5 can be provided in the middle of the movable claw 4 to make the entire clamping mechanism more stable.
  • the device body includes a fixed claw 15 fixed on the workbench/housing 14 and a movable claw 4 movably hinged to the fixed claw through a hinge mechanism; the movable claw 4 can move closer to or away from the fixed claw
  • the claw 15 is used to clamp or loosen the okra seed to be tested placed between the movable claw 4 and the fixed claw 15
  • the clamping mechanism 17 includes a base connected to the fixed claw and a connection
  • a mounting platform on the movable claw, the mounting platform is provided with a light source 1 for illuminating the okra seeds to be tested, a camera 3 for collecting image information, and a Spectrometer 3 for the spectral information of sunflower seeds
  • the main body of the device is also provided with a memory 12 and a processor 10 for processing image information.
  • the memory 12 stores contrast image information and hybrid okra seed identification models.
  • the described hybrid okra seed identification model is obtained by the following method:
  • the convolutional neural network includes two convolutional layers, two pooling layers and four fully connected layers; the filters of the two convolutional layers are set to 32 and 64 respectively; used to process hybrid okra seeds
  • the convolution kernel of the average spectral data is a 1x3 one-dimensional convolution kernel, and the compensation and filling of the convolution kernel is set to 1; the kernel size of the maximum pooling layer is 1x3, and the stride is 1; the nerve of the four-layer fully connected layer
  • the number of yuan is 512, 256, 128 and 64 respectively.
  • the light source 1 is activated to illuminate the surface of the okra seed to be tested, and the camera 2 And the spectrometer 3 to obtain the image information and spectral information of the okra seed to be tested; the processor 10 according to the image information and spectral information of the okra seed to be tested and the comparison image stored in the memory 12 Information and hybrid okra seed identification model to identify hybrid okra seeds.
  • the excitation light assembly of this embodiment is a light source 1 arranged on a mounting platform of the movable claw 4.
  • the photoelectric conversion component includes a CCD camera 2 and an imaging spectrometer 3 arranged on the mounting platform.
  • the light generated by the light source 1 illuminates the surface of the hybrid okra seed to be tested, and then the spatial information and spectral information reflected by the surface enter the CCD camera 2 and the imaging spectrometer 3 and transmit the collected spectral information to the controller 10.
  • the controller 10 processes the spectrum information to obtain the pixel point spectrum information of the hybrid seed, and calculates the average spectrum information of the hybrid seed through the pixel point spectrum information of the hybrid seed.
  • the data memory 12 stores the average spectral information of the hybrid seeds.
  • the data transmission element transmits the average spectral data measured by the measuring device to the controller and combines the output result of the deep learning discriminant model.
  • the controller 10 sends the detection result to the liquid crystal display 13 and then the liquid crystal The display shows.
  • the battery 11, the excitation light assembly, the photoelectric conversion assembly, and the push switch 6 are respectively connected to the controller 10 in the body through wires.
  • Excitation light source 1 illuminates the surface of hybrid okra seeds to be tested, and uses CCD camera 3 and imaging spectrometer 4 to obtain reflection spectrum information and image information of hybrid seeds;

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Spectroscopy & Molecular Physics (AREA)
  • Biochemistry (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • General Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Software Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Medical Informatics (AREA)
  • Data Mining & Analysis (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Investigating Or Analysing Materials By Optical Means (AREA)
  • Investigating Or Analysing Biological Materials (AREA)
  • Investigating, Analyzing Materials By Fluorescence Or Luminescence (AREA)

Abstract

一种杂交秋葵种子分类鉴别装置和方法，属于农业杂交育种的技术领域，装置包括装置主体和设置在装置主体上的装夹机构（17）；装置主体包括固定在工作台（14）上的固定爪（15）和通过铰链机构与固定爪（15）铰接的可动爪（4）；装夹机构（17）包括连接在固定爪（15）上的底座和连接在可动爪（4）上的安装压台，安装压台上设有用于照射杂交秋葵种子的光源（1）以及用于采集图像信息的相机（2）和用于获取杂交秋葵种子的光谱信息的光谱仪（3）；装置主体上还设有存储器（12）和用于处理图像信息的处理器（10），存储器（12）内存储有图像信息和杂交秋葵种子鉴别模型。该装置可以在田间或实验室对杂交秋葵种子无损、实地测量，满足了杂交育种的需要，减轻了杂交育种过程中的劳动负担。

Description

一种杂交秋葵种子分类鉴别装置和方法 技术领域
本发明涉及农业杂交育种的技术领域，具体地说，涉及一种杂交秋葵种子分类鉴别装置和方法。
背景技术
杂交育种是通过选配和培育两种或两种以上具有良好性状的品种以获得新的作物品种的有效方法。在中国，杂交水稻在粮食供应中起着举足轻重的作用，大概占据了百分之六十的水稻总体种植面积。杂交育种将亲本的基因重组形成不同的品种，提供了丰富的材料分类。获得的杂交种子具有遗传多样性，可以提高杂交植株的产量和质量。
秋葵(Abelmoschusesculentus(L.)Moench)具有丰富的纤维含量和药用价值，一直以来被人们认为是非常有营养的蔬菜作物。秋葵种子中也富含对健康有益的抗氧化化合物，具有抗疲劳和抗衰老的功效。杂交育种有助于提高以上两种作物的产量和品质，然而杂交育种过程中产生变异的后代需要进一步培育成植株并表现出一定的性状后才能够被准确识别。通过杂交植物的叶宽、株高或者颜色变化，育种专家可以挑选出符合要求植株，并进行自交等遗传学操作产生杂交种子，这个过程往往是十分耗时且耗力的。
现有技术中，高光谱成像技术已经被广泛应用在了种子识别、种子质量检测和食品检测等领域。测试样品的分子泛音跃迁和组合振动是近红外光谱分析的基础，利用高光谱成像技术可以获得样品在分子水平上的组成和特征的详细信息。近红外高光谱成像技术是一种无损快速获取780～2500nm范围内检测样品空间和光谱信息的方法。通过对以超立方体的形式存在的样品的高光谱数据进行分析，可以提取测试样品的内部物质信息和外部纹理信息。
然而，目前利用高光谱成像技术进行种子识别的装置都是大型的实验室光谱设备，对种子进行光谱识别时，需要在实验室进行，便携性大大降低。
技术解决方案
本发明的目的为提供一种杂交秋葵种子分类鉴别装置，该装置可以在田间对杂交种子无损、实地测量，大大满足了杂交育种的需要，减轻了杂交育种过程中劳动负担。
本发明的另一目的为提供一种杂交秋葵种子分类鉴别方法，该方法基于上述杂交秋葵种子分类鉴别装置实现，解决杂交育种的复杂选种过程，快速且方便地满足鉴别不同品种杂交秋葵种子的需求。
为了实现上述目的，本发明提供的杂交秋葵种子分类鉴别装置包括装置主体和设置在所述装置主体上的装夹机构；装置主体包括固定在工作台上的固定爪和通过铰链机构与固定爪铰接的可动爪；装夹机构包括连接在固定爪上的底座和连接在可动爪上的安装压台，安装压台上设有用于照射杂交秋葵种子的光源以及用于采集图像信息的相机和用于获取杂交秋葵种子的光谱信息的光谱仪；装置主体上还设有存储器和用于处理图像信息的处理器，存储器内存储有图像信息和杂交秋葵种子鉴别模型。
上述技术方案中，将待测的杂交秋葵种子夹持在装夹机构中，利用光源照射 杂交种子的待测区域的同时，使用CCD相机和成像光谱仪获得杂交种子的空间信息和光谱信息。同时通过处理器对光谱图像信息进行处理待测区域的光谱数据，秋葵种子鉴别模型利用光谱数据鉴别出杂交秋葵种子的类型并输出显示。该便携式测量装置可以在田间或实验室对杂交秋葵种子无损、实地测量，大大满足了杂交育种的需要，减轻了杂交育种过程中劳动负担。
为了方便操作可动爪工作，作为优选，装置主体上设有用于控制可动爪打开的驱动按钮，驱动按钮的底部设有第一复位弹簧。
作为优选，铰链机构包括一旋转轴和固定在旋转轴上的圆盘，还包括铰接在圆盘与按钮之间的连杆，圆盘还与可动爪的端部相铰接。当按压按钮时，连杆带动圆盘旋转，使可动爪向上移动，打开安装压台，此时可将杂交秋葵种子放置到底座上。当松开按钮时，在第一复位弹簧的作用下，按钮复位，安装压台向下移动并压紧固定杂交秋葵种子。
作为优选，可动爪与固定爪之间设有第二复位弹簧。
作为优选，底座的顶部设有用于放置杂交秋葵种子的凹槽，底部设有缓冲弹簧。凹槽的深度不应大于杂交秋葵种子的粒径。缓冲弹簧用于避免将种子压坏。
作为优选，装夹机构上设有触发光源、相机和光谱仪工作的开关，在位于开关的上方固定有用于按压开关的突触。杂交秋葵种子顶住安装压台后使得安装压台和底座见存在缝隙，从而使突触按压开关，从而触发光源、相机和光谱仪工作。
光谱仪是一种以多路、连续并具有高光谱分辨率方式获取图像信息的仪器，它通过将传统的空间成像技术与地物光谱技术有机地结合在一起，可以实现对同一地区同时获取几十个到几百个波段的地物反射光谱图像。它还可以在仪器的视场范围内同时快速测量和分析多个物体的光谱构成。这些特性均有助于应用于便携式杂交秋葵种子的分类鉴别装置中。作为优选，光谱仪采集的是874～1734nm范围内200个波段的光谱信息。
作为优选，杂交秋葵种子鉴别模型通过以下方法得到：
a.获取秋葵种子的光谱信息，并计算平均光谱，以杂交秋葵种子的类型作为标签，形成训练数据；
b.将平均光谱输入卷积进神经网络，输出种子预测类型，利用杂交秋葵种子的平均光谱和对应的标签对卷积神经网络进行训练，得到杂交秋葵种子鉴别模型。
通过杂交种子的待测区域的像素点光谱数据，计算得到待测区域的平均光谱数据。然后将平均光谱数据送入训练好的卷积神经网络模型，通过判别模型数据待测种子的种类并显示在液晶显示屏上。
作为优选，卷积神经网络包括两层卷积层、两层池化层和四层全连接层；两层卷积层的过滤器分别设置为32和64；用来处理杂交秋葵种子的平均光谱数据的卷积核为1x3的一维卷积核，卷积核的补偿和填充设置为1；最大池化层的内核尺寸为1x3，步幅为1；四层全连接层的神经元数量分别为512、256、128和64。步骤b中采用小批量训练的方法，批量数选择的25个，训练次数为1500次。在每一个全连接层间加入了批量归一化和丢弃法防止过拟合出现，丢弃法丢弃概率为0.5。
为了实现上述另一目的，本发明提供的杂交秋葵种子分类鉴别方法包括以下步骤：
1)收集未放置杂交秋葵种子的底座的黑白背景，白背景的光谱数据为W，黑背景的光谱数据为D；
2)将未知品种的杂交秋葵种子放入底座的检测位，当相机检测到种子的存在 后，控制装夹机构对杂交秋葵种子进行夹紧，并获取杂交秋葵种子的光谱数据；
3)提取待测区域内像素点的光谱数据信息，将待测区域内的图像像素点进行平均处理，获得种子的平均光谱为P，实际平均光谱数据为S为：
Figure PCTCN2019110493-appb-000001
4)实际平均光谱数据S输入训练好的杂交秋葵种子鉴别模型，输出杂交秋葵种子类型并进行显示。
具体的，获取杂交秋葵种子的光谱数据使用的成像光谱仪采集的是874～1734nm范围内200个波段的光谱信息。
步骤3)中将样本杂交种子放入待测区域采集，利用光学系统采集种子光谱信息，获得杂交种子的像素点的光谱信息数量为L和光谱数据总和为Z。通过成像光谱仪采集的像素点的光谱数据总和为Z，像素点的个数为L，则未经校正的平均光谱P为：
Figure PCTCN2019110493-appb-000002
与现有技术相比，本发明的有益效果为：
本发明的对不同种类杂交秋葵种子进行分类鉴别的装置和基于该装置的鉴别方法可以在田间或实验室内对杂交秋葵种子无损、快速测量，大大满足了杂交育种选育种子的需要。解决了杂交育种过程中耗时耗力的问题，为加快杂交育种实验过程提供了十分重要的手段。
附图说明
图1为本发明实施例中杂交秋葵种子分类鉴别装置的结构示意图；
图2为本发明实施例中杂交秋葵种子分类鉴别方法的流程图；
图3为本发明实施例中利用杂交秋葵种子分类鉴别方法获得的杂交秋葵种子的平均光谱图；
图4为本发明实施例中利用杂交秋葵种子分类鉴别方法获得的分类效果图的前后对比图。
本发明的实施方式
为使本发明的目的、技术方案和优点更加清楚，以下结合实施例及其附图对本发明作进一步说明。
实施例1
参见图1和图2，本实施例的杂交秋葵种子分类鉴别装置包括装置主体和设置在所述装置主体上的装夹机构17。
其中，装置主体包括外壳14和设置在外壳14内的控制器10，壳体14内还设有电池11，控制器10连接有数据存储器12，并连接有用于接收和显示测量结果的液晶显示屏13。
装置主体还包括固定爪15、可动爪4、激发光组件和光电转换组件。在固定爪15和可动爪4之间设置与杂交秋葵配合的装夹机构17，并设有按压式开关6。
固定爪15与外壳14连成一体。固定爪15设置有缓冲弹簧16，保持装夹机构17始终处于工作位置，并避免将种子压坏。可动爪4的内侧设置滑槽，固定爪15可在滑槽内进行上下滑动，使夹爪更加灵活，即使不放入杂交秋葵种子，整个装置也不会受到外界干扰。当杂交秋葵种子被夹持后会触发按压式开关6， 控制激发光组件和光电转换组件工作。
装夹机构17包括连接在固定爪15上的底座和连接在可动爪4上的安装压台。外壳14一侧设有按钮7，按钮7通过铰链机构控制可动爪的活动。铰链机构包括一旋转轴和固定在旋转轴上的圆盘9，还包括铰接在圆盘9与按钮7之间的连杆，圆盘9还与可动爪4的端部相铰接。按钮7的下方设置有两个第一复位弹簧8。按压按钮后通过铰链机构控制夹持爪打开，松开按钮后在第一复位弹簧8的作用下恢复原位并对放入的杂交秋葵种子进行夹紧。可动爪4的中部设有第二复位弹簧5，使整个夹持机构更加稳定。
本实施例的激发光组件为设置在可动爪4的安装压台上的光源1。光电转换组件包括设置在安装压台上的CCD相机2和成像光谱仪3。光源1产生的光照射待测杂交秋葵种子的表面，随后表面反射的空间信息和光谱信息进入CCD相机2和成像光谱仪3内并将采集到的光谱信息传输至控制器10。通过控制器10对光谱信息进行处理获得杂交种子的像素点光谱信息，通过杂交种子的像素点光谱信息，计算得到杂交种子的平均光谱信息。数据存储器12存储杂交种子的平均光谱信息，数据传输元件将测量装置测量到的平均光谱数据传输给控制器结合深度学习判别模型输出结果，控制器10将检测结果发送给液晶显示屏13后经液晶显示屏进行显示。
本实施例中电池11、激发光组件、光电转换组件、按压式开关6分别通过电线与机身内的控制器10相连。
利用上述杂交秋葵种子分类鉴别装置对待测杂交秋葵种子进行鉴别的方法步骤如下：
1)将黑色背景片和白色背景片置于可动爪4的检测位置，进行黑白校正的光谱数据采集，白色背景的光谱数据为W，黑色背景的光谱数据为D；
2)将待测的杂交秋葵种子放入可动爪4的检测位置，控制两个夹持爪闭合进行杂交秋葵种子的光谱数据采集，参见图3；
3)激发光源照射待测杂交秋葵种子表面，利用CCD相机和成像光谱仪获得杂交种子的反射光谱信息和图像信息；
4)对获取的图像处理，提取待测区域内像素点的光谱数据信息；
5)通过将获得杂交种子的图像像素点信息进行平均处理，获得种子的平均光谱为P，实际应用的平均光谱数据为S为：
Figure PCTCN2019110493-appb-000003
6)将获得的杂交秋葵种子的平均光谱数据与深度学习算法建立的判别分析模型进行结合，判断杂交种子的种类并最终输出在液晶显示屏上，分类效果图参见图4。
实施例2
参见图1和图2，本实施例的杂交秋葵种子分类鉴别装置包括装置主体和设置在所述装置主体上的装夹机构17。
其中，装置主体包括外壳14和设置在外壳14内的控制器10，壳体14内还设有电池11，控制器10连接有数据存储器12，并连接有用于接收和显示测量结果的液晶显示屏13。可以理解，如图2所示，电池11与装置主体内的用电单元，例如数据存储器12,、液晶显示屏13、控制器10等均有直接或间接的连接，以便为这些用电单元供应电能。
装置主体还包括固定爪15、可动爪4、激发光组件和光电转换组件。在固定爪和可动爪4之间设置与杂交秋葵种子配合的装夹机构17，并设有按压式开关6。
固定爪15与外壳连成一体。固定爪15上设置有缓冲弹簧16，该缓冲弹簧16一端固定于固定爪15，另一端与装夹机构17连接，进而保持装夹机构始终处于工作位置。此外，缓冲弹簧16的弹性，使得装夹机构17在夹持杂交秋葵种子时，上下两夹持爪保持弹性接触，避免将装夹机构17中的待测秋葵种子压坏。可动爪4的内侧设置滑槽，固定爪15可在滑槽内进行上下滑动，使夹爪更加灵活，即使不放入杂交秋葵种子，整个装置也不会受到外界干扰。当杂交秋葵种子被夹持后会触发按压式开关6，控制激发光组件和光电转换组件工作。
装夹机构17包括连接在固定爪15上的底座和连接在可动爪4上的安装压台。外壳14一侧设有按钮7，按钮7通过铰链机构控制可动爪4的活动。铰链机构包括一旋转轴和固定在旋转轴上的圆盘9，还包括铰接在圆盘9与按钮7之间的连杆，圆盘9还与可动爪4的端部相铰接。按钮7的下方设置有两个第一复位弹簧8。按压按钮7后，下压的按钮7通过连杆致动所述圆盘9。固定于旋转轴的圆盘9以此随着旋转轴一起转动。转动的圆盘9带动可动爪4的活动，以此通过铰链机构控制夹持爪打开。按钮7的下方设置有第一复位弹簧8。所述第一复位弹簧8的一端与按钮7的下部连接，另一端与外壳14固定连接，使得在按钮7被松开后，按钮7能够在第一复位弹簧8的作用下恢复原位。恢复原位的按钮7通过铰链结构，复位可动爪4，使得夹持爪重新闭合，进而对放入的杂交秋葵种子进行夹紧。优选地，还可以在可动爪4的中部设有第二复位弹簧5，使整个夹持机构更加稳定。
所述装置主体包括固定在工作台/外壳14上的固定爪15和通过铰链机构与所述固定爪可活动铰接的可动爪4；所述可动爪4可活动地接近或远离所述固定爪15，以夹持或松开置于所述可动爪4与所述固定爪15之间的待测秋葵种子；所述装夹机构17包括连接在所述固定爪上的底座和连接在所述可动爪上的安装压台，所述安装压台上设有用于照射所述待测秋葵种子的光源1以及用于采集图像信息的相机3和用于获取所述待测秋葵种子的光谱信息的光谱仪3；所述装置主体上还设有存储器12和用于处理图像信息的处理器10，所述存储器12内存储有对比图像信息和杂交秋葵种子鉴别模型。
所述的杂交秋葵种子鉴别模型通过以下方法得到：
a.获取秋葵种子的光谱信息，并计算平均光谱，以杂交秋葵种子的类型作为标签，形成训练数据；
b.将平均光谱输入卷积进神经网络，输出种子预测类型，利用杂交秋葵种子的平均光谱和对应的标签对卷积神经网络进行训练，得到杂交秋葵种子鉴别模型。
其中，所述的卷积神经网络包括两层卷积层、两层池化层和四层全连接层；两层卷积层的过滤器分别设置为32和64；用来处理杂交秋葵种子的平均光谱数据的卷积核为1x3的一维卷积核，卷积核的补偿和填充设置为1；最大池化层的内核尺寸为1x3，步幅为1；四层全连接层的神经元数量分别为512、256、128和64。
当所述待测秋葵种子被夹持于所述可动爪4与所述固定爪15之间时，所述光源1被启动以照射所述待测秋葵种子表面，同时所述相机2和所述光谱仪3获取所述待测秋葵种子的图像信息和光谱信息；所述处理器10根据所述待测秋葵种子的图像信息和光谱信息以及存储在所述存储器12内的对比图像信息和杂交秋葵种子鉴别模型，鉴别出杂交秋葵种子。
本实施例的激发光组件为设置在可动爪4的安装压台上的光源1。光电转换组件包括设置在安装压台上的CCD相机2和成像光谱仪3。光源1产生的光照射待测杂交秋葵种子的表面，随后表面反射的空间信息和光谱信息进入CCD相机2和成像光谱仪3内并将采集到的光谱信息传输至控制器10。通过控制器10对光谱信息进行处理获得杂交种子的像素点光谱信息，通过杂交种子的像素点光谱信息，计算得到杂交种子的平均光谱信息。数据存储器12存储杂交种子的平均光谱信息，数据传输元件将测量装置测量到的平均光谱数据传输给控制器结合深度学习判别模型输出结果，控制器10将检测结果发送给液晶显示屏13后经液晶显示屏进行显示。
本实施例中电池11、激发光组件、光电转换组件、按压式开关6分别通过电线与机身内的控制器10相连。
利用上述杂交秋葵种子分类鉴别装置对待测杂交秋葵种子进行鉴别的方法步骤如下：
1)将黑色背景片和白色背景片置于可动爪4的检测位置，进行黑白校正的光谱数据采集；其中，白色背景的光谱数据为W，黑色背景的光谱数据为D；
2)将待测的杂交秋葵种子放入可动爪4的检测位置，控制两个夹持爪闭合进行杂交秋葵种子的光谱数据采集；杂交秋葵种子的光谱数据采集可参见图3；
3)激发光源1照射待测杂交秋葵种子表面，利用CCD相机3和成像光谱仪4获得杂交种子的反射光谱信息和图像信息；
4)对获取的所述图像信息进行处理，提取待测区域内像素点的图像像素点光谱数据信息；
5)通过将获得杂交种子的图像像素点光谱数据信息进行平均处理，获得种子的平均光谱为P，实际应用的平均光谱数据为S为：
Figure PCTCN2019110493-appb-000004
6)将获得的杂交秋葵种子的平均光谱数据S与深度学习算法建立的判别分析模型进行结合，判断杂交种子的种类并最终输出在液晶显示屏上，分类效果图参见图4。

Claims (20)

  1. 一种杂交秋葵种子分类鉴别装置，其特征在于，包括装置主体和设置在所述装置主体上的装夹机构；
    所述装置主体包括固定在工作台上的固定爪和通过铰链机构与所述固定爪铰接的可动爪；
    所述装夹机构包括连接在所述固定爪上的底座和连接在所述可动爪上的安装压台，所述安装压台上设有用于照射杂交秋葵种子的光源以及用于采集图像信息的相机和用于获取杂交秋葵种子的光谱信息的光谱仪；
    所述装置主体上还设有存储器和用于处理图像信息的处理器，所述存储器内存储有图像信息和杂交秋葵种子鉴别模型。
  2. 根据权利要求1所述的杂交秋葵种子分类鉴别装置，其特征在于，所述的装置主体上设有用于控制所述可动爪打开的驱动按钮，所述驱动按钮的底部设有第一复位弹簧。
  3. 根据权利要求2所述的杂交秋葵种子分类鉴别装置，其特征在于，所述的铰链机构包括一旋转轴和固定在所述旋转轴上的圆盘，还包括铰接在所述圆盘与所述按钮之间的连杆，所述圆盘还与所述可动爪的端部相铰接。
  4. 根据权利要求3所述的杂交秋葵种子分类鉴别装置，其特征在于，所述的可动爪与所述固定爪之间设有第二复位弹簧。
  5. 根据权利要求1所述的杂交秋葵种子分类鉴别装置，其特征在于，所述的底座的顶部设有用于放置杂交秋葵种子的凹槽，底部设有缓冲弹簧。
  6. 根据权利要求1所述的杂交秋葵种子分类鉴别装置，其特征在于，所述的装夹机构上设有触发所述光源、相机和光谱仪工作的开关，在位于所述开关的上方固定有用于按压所述开关的突触。
  7. 根据权利要求1所述的杂交秋葵种子分类鉴别装置，其特征在于，所述的光谱仪采集的是874～1734nm范围内200个波段的光谱信息。
  8. 根据权利要求1所述的杂交秋葵种子分类鉴别装置，其特征在于，所述的杂交秋葵种子鉴别模型通过以下方法得到：
    a.获取秋葵种子的光谱信息，并计算平均光谱，以杂交秋葵种子的类型作为标签，形成训练数据；
    b.将平均光谱输入卷积进神经网络，输出种子预测类型，利用杂交秋葵种子的平均光谱和对应的标签对卷积神经网络进行训练，得到杂交秋葵种子鉴别模型。
  9. 根据权利要求8所述的杂交秋葵种子分类鉴别装置，其特征在于，所述的卷积神经网络包括两层卷积层、两层池化层和四层全连接层；两层卷积层的过滤器分别设置为32和64；用来处理杂交秋葵种子的平均光谱数据的卷积核为1x3的一维卷积核，卷积核的补偿和填充设置为1；最大池化层的内核尺寸为1x3，步幅为1；四层全连接层的神经元数量分别为512、256、128和64。
  10. 一种杂交秋葵种子分类鉴别方法，其特征在于，基于权利要求1～9任意权利要求所述的杂交秋葵种子分类鉴别装置实现，该方法包括以下步骤：
    1)收集未放置杂交秋葵种子的底座的黑白背景，白背景的光谱数据为W，黑背景的光谱数据为D；
    2)将未知品种的杂交秋葵种子放入底座的检测位，当相机检测到种子的存在后，控制装夹机构对杂交秋葵种子进行夹紧，并获取杂交秋葵种子的光谱数据；
    3)提取待测区域内像素点的光谱数据信息，将待测区域内的图像像素点进行平均处理，获得种子的平均光谱为P，实际平均光谱数据为S为：
    Figure PCTCN2019110493-appb-100001
    4)实际平均光谱数据S输入训练好的杂交秋葵种子鉴别模型，输出杂交秋葵种子类型并进行显示。
  11. 一种杂交秋葵种子分类鉴别装置，其特征在于，包括装置主体和设置在所述装置主体上的装夹机构；
    所述装置主体包括固定在工作台上的固定爪和通过铰链机构与所述固定爪可活动铰接的可动爪；所述可动爪可活动地接近或远离所述固定爪，以夹持或松开置于所述可动爪与所述固定爪之间的待测秋葵种子；
    所述装夹机构包括连接在所述固定爪上的底座和连接在所述可动爪上的安装压台，所述安装压台上设有用于照射所述待测秋葵种子的光源以及用于采集图像信息的相机和用于获取所述待测秋葵种子的光谱信息的光谱仪；
    所述装置主体上还设有存储器和用于处理图像信息的处理器，所述存储器内存储有对比图像信息和杂交秋葵种子鉴别模型；
    当所述待测秋葵种子被夹持于所述可动爪与所述固定爪之间时，所述光源被启动以照射所述待测秋葵种子表面，同时所述相机和所述光谱获取所述待测秋葵种子的图像信息和光谱信息；所述处理器根据所述待测秋葵种子的图像信息和光谱信息以及存储在所述存储器内的对比图像信息和杂交秋葵种子鉴别模型，鉴别出杂交秋葵种子。
  12. 根据权利要求11所述的杂交秋葵种子分类鉴别装置，其特征在于，所述的装置主体上设有用于控制所述可动爪打开的驱动按钮；所述驱动按钮通过铰链机构与所述可动爪连接，使得所述驱动按钮被按压时能够致动所述可动爪，以控制所述可动爪的打开；其中，所述驱动按钮的底部设有第一复位弹簧，所述第一复位弹簧固定于所述工作台，使得所述驱动按钮在失去按压力后能够在所述第一复位弹簧的弹力下复位。
  13. 根据权利要求12所述的杂交秋葵种子分类鉴别装置，其特征在于，所述的铰链机构包括一旋转轴和固定在所述旋转轴上的圆盘，还包括铰接在所述圆盘与所述驱动按钮之间的连杆，所述圆盘还与所述可动爪的端部相铰接；所述驱动在被按压时，通过所述连杆带动所述圆盘转动；转动的所述圆盘进一步驱动所述可动爪的端部，以控制所述可动爪的打开。
  14. 根据权利要求13所述的杂交秋葵种子分类鉴别装置，其特征在于，所述可动爪与所述固定爪之间设有第二复位弹簧。
  15. 根据权利要求11所述的杂交秋葵种子分类鉴别装置，其特征在于，所述底座的顶部设有用于容置杂交秋葵种子的凹槽，底部设有缓冲弹簧；所述缓冲弹簧一端连接于所述底座的 底部，另一端固定连接于所述工作台；所述缓冲弹簧允许所述底座在受到一定压力时向下活动，且保持所述底座与所述安装压台之间的夹紧。
  16. 根据权利要求11所述的杂交秋葵种子分类鉴别装置，其特征在于，所述的装夹机构上设有触发所述光源、相机和光谱仪工作的第一开关，所述第一开关能够在所述待测秋葵种子被夹持后被触发，控制所述光源、相机和光谱仪工作。
  17. 根据权利要求11所述的杂交秋葵种子分类鉴别装置，其特征在于，所述的光谱仪采集的是874～1734nm范围内200个波段的光谱信息。
  18. 根据权利要求11所述的杂交秋葵种子分类鉴别装置，其特征在于，所述的杂交秋葵种子鉴别模型通过以下方法得到：
    a.获取秋葵种子的光谱信息，并计算平均光谱，以杂交秋葵种子的类型作为标签，形成训练数据；
    b.将平均光谱输入卷积进神经网络，输出种子预测类型，利用杂交秋葵种子的平均光谱和对应的标签对卷积神经网络进行训练，得到所述杂交秋葵种子鉴别模型。
  19. 根据权利要求18所述的杂交秋葵种子分类鉴别装置，其特征在于，所述的卷积神经网络包括两层卷积层、两层池化层和四层全连接层；两层卷积层的过滤器分别设置为32和64；用来处理杂交秋葵种子的平均光谱数据的卷积核为1x3的一维卷积核，卷积核的补偿和填充设置为1；最大池化层的内核尺寸为1x3，步幅为1；四层全连接层的神经元数量分别为512、256、128和64。
  20. 一种基于权利要求11～19任意一个权利要求所述的杂交秋葵种子分类鉴别装置实现杂交秋葵种子分类鉴别方法，其特征在于，该方法包括以下步骤：
    收集未放置秋葵种子的底座的黑背景白背景，其中，所述白背景的光谱数据为W，所述黑背景的光谱数据为D；
    将待测秋葵种子放入底座的检测位后，装夹机构对所述待测秋葵种子进行夹紧，相机和光谱仪并获取所述待测秋葵种子的光谱数据和图像信息；
    根据所述图像信息提取待测区域内像素点的图像像素点光谱数据信息，将待测区域内的图像像素点进行平均处理，获得种子的平均光谱为P，实际平均光谱数据为S为：
    Figure PCTCN2019110493-appb-100002
    将所述实际平均光谱数据S输入训练好的所述杂交秋葵种子鉴别模型，输出杂交秋葵种子类型并进行显示。
PCT/CN2019/110493 2019-04-12 2019-10-11 一种杂交秋葵种子分类鉴别装置和方法 WO2020206963A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CH000718/2020A CH716240B9 (de) 2019-04-12 2019-10-11 Vorrichtung und Verfahren zur Klassifizierung und Identifizierung von Saatgut von hybridem Abelmoschus esculentus.

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910292970.6 2019-04-12
CN201910292970.6A CN110174357A (zh) 2019-04-12 2019-04-12 一种杂交秋葵种子分类鉴别装置和方法

Publications (1)

Publication Number Publication Date
WO2020206963A1 true WO2020206963A1 (zh) 2020-10-15

Family

ID=67689840

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/110493 WO2020206963A1 (zh) 2019-04-12 2019-10-11 一种杂交秋葵种子分类鉴别装置和方法

Country Status (3)

Country Link
CN (1) CN110174357A (zh)
CH (1) CH716240B9 (zh)
WO (1) WO2020206963A1 (zh)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110174357A (zh) * 2019-04-12 2019-08-27 浙江大学 一种杂交秋葵种子分类鉴别装置和方法
CN110892878A (zh) * 2019-12-05 2020-03-20 刘宝祥 三元杂交培育方法与训练方法及设备

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1831515A (zh) * 2006-04-03 2006-09-13 浙江大学 用可见光和近红外光谱技术无损鉴别作物种子品种的方法
US20130294656A1 (en) * 2011-07-19 2013-11-07 Ball Horticultural Company Seed holding device and seed classification system with seed holding device
CN104990890A (zh) * 2015-06-24 2015-10-21 中国农业大学 固体单粒无损检测与自动化分选系统及固体单粒分选方法
CN106383088A (zh) * 2016-08-19 2017-02-08 合肥工业大学 一种基于多光谱成像技术的种子纯度快速无损检测方法
CN106546541A (zh) * 2016-10-31 2017-03-29 浙江大学 一种基于高光谱转基因玉米籽粒的识别装置与方法
CN208109642U (zh) * 2018-04-20 2018-11-16 浙江大学 一种便携式航空喷施作业的雾滴沉积效果测量装置
CN110174357A (zh) * 2019-04-12 2019-08-27 浙江大学 一种杂交秋葵种子分类鉴别装置和方法

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2013133171A1 (ja) * 2012-03-05 2013-09-12 住友電気工業株式会社 種子選別方法及び種子選別装置
CN104850836B (zh) * 2015-05-15 2018-04-10 浙江大学 基于深度卷积神经网络的害虫图像自动识别方法
CN206804510U (zh) * 2017-06-19 2017-12-26 东北林业大学 用于检测林区树叶的便携装置
CN107576618B (zh) * 2017-07-20 2020-04-28 华南理工大学 基于深度卷积神经网络的水稻穗瘟检测方法及系统
CN207280952U (zh) * 2017-10-13 2018-04-27 西北农林科技大学 一种夹持式苹果霉心病检测设备
CN108500990A (zh) * 2018-04-07 2018-09-07 李正梅 一种夹具及其工作方法
CN108985237A (zh) * 2018-07-20 2018-12-11 安徽农业大学 一种基于深度混合的小麦赤霉病的检测方法及其系统
CN109352618A (zh) * 2018-10-27 2019-02-19 邵继民 一种用于汽车停车场的入口处自动取卡辅助装置

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1831515A (zh) * 2006-04-03 2006-09-13 浙江大学 用可见光和近红外光谱技术无损鉴别作物种子品种的方法
US20130294656A1 (en) * 2011-07-19 2013-11-07 Ball Horticultural Company Seed holding device and seed classification system with seed holding device
CN104990890A (zh) * 2015-06-24 2015-10-21 中国农业大学 固体单粒无损检测与自动化分选系统及固体单粒分选方法
CN106383088A (zh) * 2016-08-19 2017-02-08 合肥工业大学 一种基于多光谱成像技术的种子纯度快速无损检测方法
CN106546541A (zh) * 2016-10-31 2017-03-29 浙江大学 一种基于高光谱转基因玉米籽粒的识别装置与方法
CN208109642U (zh) * 2018-04-20 2018-11-16 浙江大学 一种便携式航空喷施作业的雾滴沉积效果测量装置
CN110174357A (zh) * 2019-04-12 2019-08-27 浙江大学 一种杂交秋葵种子分类鉴别装置和方法

Also Published As

Publication number Publication date
CH716240B8 (de) 2023-08-31
CH716240B1 (de) 2023-07-14
CN110174357A (zh) 2019-08-27
CH716240B9 (de) 2023-12-29

Similar Documents

Publication Publication Date Title
CN103063585B (zh) 瓜果成熟度快速无损检测装置及检测系统建立方法
CN105021617B (zh) 基于高光谱成像的整株水稻叶绿素含量测量装置及方法
Mahajan et al. Image acquisition techniques for assessment of legume quality
Shahin et al. A machine vision system for grading lentils
Yu et al. Identification of crack features in fresh jujube using Vis/NIR hyperspectral imaging combined with image processing
Chevallier et al. Application of PLS‐DA in multivariate image analysis
Barbedo et al. Detection of sprout damage in wheat kernels using NIR hyperspectral imaging
Belin et al. Thermography versus chlorophyll fluorescence imaging for detection and quantification of apple scab
KR101341815B1 (ko) 초분광 영상처리를 이용한 종자 선별 장치
CN110763698B (zh) 一种基于特征波长的高光谱柑橘叶片病害识别方法
CN108444928A (zh) 一种利用种胚谱图特征波段识别谷物种子冻伤情况的方法
WO2020206963A1 (zh) 一种杂交秋葵种子分类鉴别装置和方法
CN210465246U (zh) 小麦感染赤霉病等级及病粒率高光谱无损检测装置的光谱采集装置
Zhang et al. Non-destructive identification of slightly sprouted wheat kernels using hyperspectral data on both sides of wheat kernels
Wang et al. A multimodal machine vision system for quality inspection of onions
US10690592B2 (en) Haploid seed classification using single seed near-infrared spectroscopy
Lancelot et al. Near-infrared hyperspectral imaging for following imbibition of single wheat kernel sections
CN109470639A (zh) 基于可见光和近红外高光谱成像的桑螟幼虫及其对桑叶损害的快速识别方法
CN114419311B (zh) 一种基于多源信息的百香果成熟度无损检测方法及装置
CN116297236A (zh) 一种基于高光谱的单粒玉米种子活力鉴别方法及装置
CN114136920A (zh) 一种基于高光谱的单粒杂交水稻种子种类鉴定方法
CN106940292A (zh) 基于多光谱成像技术的虫蛀巴旦木原料快速无损鉴别方法
CN101911877B (zh) 基于激光漫反射图像技术的种子生命力鉴别装置和方法
Varela et al. A novel high-throughput hyperspectral scanner and analytical methods for predicting maize kernel composition and physical traits
CN108369192A (zh) 用于测量花序、种子和/或种子产量表型的方法和设备

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19923874

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19923874

Country of ref document: EP

Kind code of ref document: A1