CN117647492A

CN117647492A - Online detection device and method for quality of corn seeds by cooperative air suction of mechanical arm

Info

Publication number: CN117647492A
Application number: CN202311407358.1A
Authority: CN
Inventors: 杨然兵; 许�鹏; 符丽霞; 杨松梅; 查显涛; 青苡任; 王涛; 孙文斌; 徐康; 陈栋泉; 潘永菲; 钟涛; 郑刚
Original assignee: Hainan University
Current assignee: Hainan University
Priority date: 2023-10-27
Filing date: 2023-10-27
Publication date: 2024-03-05
Anticipated expiration: 2043-10-27
Also published as: CN117647492B

Abstract

The mechanical arm cooperated air suction type corn seed quality online detection device and method comprises a frame, a feeding device fixedly arranged on the frame, a conveying device, a detection device, a screening device and a control system; the multi-mechanical-arm cooperative air-aspiration type corn seed quality online detection device and method can be used for nondestructive and rapid detection of corn seed quality in fields or laboratories, and greatly meets the requirements of seed production and breeding. The method solves the problems of time consumption and labor consumption in the seed production and seed reproduction process, provides a reference for accelerating the seed production and seed reproduction process, adopts spectrum imaging to rapidly acquire the spectrum image of the corn seeds, can acquire a large amount of image information and spectrum information related to the seeds, and combines a deep learning technology to construct a corn seed quality detection model, thereby effectively improving the detection precision and meeting the requirements of online corn seed quality detection in the seed production and seed reproduction process.

Description

Online detection device and method for quality of corn seeds by cooperative air suction of mechanical arm

Technical Field

The invention relates to the field of connecting structures, in particular to a device and a method for online detection of the quality of corn seeds by cooperative air suction of a mechanical arm.

Background

Corn seeds are important preconditions for ensuring corn quality and yield, while important quality indicators of corn seeds include volume weight, imperfect grain (worm eroded grain, disease spot grain, broken grain, sprouted grain, mildewed grain, and heat damaged grain), impurities, color, odor, moisture content, and mildewed grain. Besides the external forms such as impurities, color and luster, there are important internal forms such as imperfect grains, moisture content, mildew grains and the like. These problems mainly arise from harvesting, drying, transportation and storage of corn production. For the internal form problem of corn seeds, the seed coats are intact, the appearance is not abnormal, the corn seeds are difficult to perceive and draw attention in a normal state, the germination and tidy emergence degree of the seeds, the disease resistance of the finally grown plants and the like are seriously influenced.

At present, the corn seed quality is detected mainly by adopting morphological, physiological and biochemical and biotechnology (protein electrophoresis, DNA molecular marking and the like), and the methods have the problems of high subjectivity and randomness, need to consume special reagents, have destructive property on samples, have limited detection quantity, are expensive and have high requirements on operators; in recent years, scholars at home and abroad have made a great deal of researches on the quality detection of corn seeds, and the characteristic indexes inside and outside the quality of the seeds are detected mainly by adopting a spectrum imaging technology and combining a machine learning algorithm. However, these studies or research-developed devices are all based on static conditions, and it is difficult to meet the online rapid detection of mass corn seed quality.

The patent number is: the CN201611000759.5 patent proposes a seed spectrum detection method and system, the system emits monochromatic light to a seed to be detected, so as to excite the seed to generate reflected light and fluorescence, collect the reflected light and the fluorescence, obtain reflected light spectrum data and fluorescence spectrum data, and obtain a detection result of the seed to be detected according to the reflected light spectrum data and the fluorescence spectrum data.

Disclosure of Invention

The invention mainly aims to provide a mechanical arm cooperative air suction type corn seed quality online detection device and method.

In order to achieve the above purpose, the technical scheme adopted by the invention is as follows:

the mechanical arm cooperated air suction type corn seed quality online detection device comprises a frame, a feeding device fixedly arranged on the frame, a conveying device, a detection device, a screening device and a control system;

the conveying device comprises a conveying belt and a driving motor, and the driving motor drives the conveying belt to operate;

the feeding device is arranged above the transmission device and is connected with the conveyor belt in a sliding manner, and comprises a feeding chamber and a baffle brush;

the detection device comprises a hyperspectral imaging system and a connecting wire, and the hyperspectral imaging system is arranged right above the conveyor belt;

the screening device comprises one or more groups of screening devices, wherein each group of screening device comprises a mechanical arm, a low-quality chamber, a high-quality chamber, a fan and a ventilation pipeline, and the high-quality chamber is arranged right below the tail end of the conveying direction of the conveying belt;

the control system comprises a connecting wire and a computer, and is connected with the detecting device and the screening device through the connecting wire.

Further, the outer surface of the conveyor belt is fixedly provided with a plurality of sample loading grooves with the same distance.

Further, the pan feeding mouth has been seted up to the feed chamber upper end, and the discharge gate has been seted up to this feed chamber lower extreme, and fixed mounting has the baffle brush on this feed chamber's the lateral wall, and this discharge gate is arranged in directly over the conveyer belt, has the first clearance of ejection of compact between this discharge gate and the conveyer belt, and this baffle brush also is arranged in directly over the conveyer belt, is equipped with the second clearance of ejection of compact between this baffle brush and the conveyer belt.

Further, the hyperspectral imaging system comprises a camera, a corresponding halogen lamp light source, an optical module and a sensor module, wherein the camera is arranged right above the conveyor belt; the halogen lamp light source is arranged around the camera and forms an included angle of 45 degrees with the upper end face of the conveyor belt, the optical module is electrically connected with the camera and used for collecting spectra and images in real time, and the sensing module is also electrically connected with the camera and used for detecting operation parameters of the camera and performing position fine adjustment.

Further, the camera comprises one of a hyperspectral camera, a multispectral camera, a fluorescence spectrum camera and a near infrared spectrum camera.

Further, the mechanical arm is fixedly arranged on one side or two sides of the frame, the lower end of the mechanical arm is fixedly provided with a rotating base, the rotating base is fixedly arranged on the frame, the tail end of the mechanical arm is fixedly provided with an adsorption seat, the adsorption seat is fixedly provided with an adsorption hole, the adsorption seat is fixedly connected with a ventilation pipeline, and the other end of the ventilation pipeline is fixedly connected with a fan.

Further, a filter screen is fixedly arranged at the adsorption hole of the adsorption seat.

Furthermore, a low-quality chamber is fixedly arranged on one side of the mechanical arm arranged on the frame.

Further, the height of the first gap is larger than the height of the second gap, the height of the first gap is not higher than the height of the normal seeds, and the height of the second gap is not higher than one third of the height of the normal seeds.

The detection method comprises the following steps:

s1, feeding, namely pouring seeds to be detected into a feeding device in batches, and naturally dropping the seeds onto a conveyor belt by the gravity of the seeds;

s2: secondly, feeding, namely controlling the quantity of seeds falling into the conveyor belt through a sample loading groove on the conveyor belt, namely automatically falling into the sample loading chamber when the conveyor belt with the sample loading groove passes through the position right below a feeding chamber of a feeding device, taking the seeds out of the feeding chamber through the sample loading chamber, and conveying the seeds to a monitoring position through the conveyor belt;

s3: seed limiting, namely blocking seeds on the conveyor belt except for the sample loading groove to be transferred to a detection area through a baffle brush;

s4: monitoring, namely illuminating and photographing a conveyor belt and seeds entering a monitoring device through a light source on a camera;

s5: sampling, namely establishing a comparison file library through a data acquisition device, comparing photo information newly acquired by a camera with data in the comparison file library, and transmitting the comparison information to a screening device;

s6: and (3) screening, namely adsorbing abnormal seeds from the conveyor belt through a screening device, placing the abnormal seeds into a low-quality chamber, moving the normal seeds to the upper part of the high-quality chamber through the conveyor belt, and enabling the normal seeds to fall into the high-quality chamber to finish monitoring.

Compared with the prior art, the invention has the beneficial effects that:

the multi-mechanical-arm cooperative air-aspiration type corn seed quality online detection device and method can be used for nondestructive and rapid detection of corn seed quality in fields or laboratories, and greatly meets the requirements of seed production and breeding. Solves the problems of time consumption and labor consumption in the seed production and seed reproduction process, and provides a reference for accelerating the seed production and seed reproduction process;

according to the invention, the spectrum image of the corn seeds is rapidly acquired by adopting the spectrum imaging, a large amount of image information and spectrum information related to the seeds can be obtained, and a corn seed quality detection model is constructed by combining a deep learning technology, so that the detection precision can be effectively improved, and the online detection requirement of the corn seed quality in the seed production and propagation processes can be met;

according to the invention, on the basis of automatically detecting the quality of the corn seeds, the corn seeds which do not meet the quality requirements can be simultaneously processed in a large scale by combining aerodynamics and mechanical kinematics, so that the time and the labor are saved, the efficiency is high, the practicability is high, and technical support can be provided for a seed company to improve the commodity quality of the seeds.

Drawings

FIG. 1 is a front view block diagram of the present invention;

FIG. 2 is a partial top view block diagram of the present invention;

FIG. 3 is a partial construction diagram of embodiment 1 of the present invention;

fig. 4 is a partial structural view of embodiment 2 of the present invention.

Reference numerals

1. A feed chamber; 2. a baffle brush; 3. a hyperspectral imaging system; 4. a mechanical arm; 5. a connecting wire; 6. a computer; 7. a conveyor belt; 8. a high quality chamber; 9. corn seeds; 10. a low-quality chamber; 11. a camera; 12. a frame; 13. a blower; 14. a ventilation duct; 16. an adsorption seat; 17. rotating the base; 18. a driving motor; 19. and (5) a sample loading groove.

Detailed Description

The preferred embodiments of the present invention will be described in detail below with reference to the attached drawings so that the advantages and features of the present invention can be more easily understood by those skilled in the art, thereby making clear and defining the scope of the present invention.

Example 1:

referring to fig. 1-3, an on-line detection device for the quality of corn seeds 9 by combining a mechanical arm 4 with air suction comprises a frame 12, a feeding device, a conveying device, a detection device, a screening device and a control system.

Wherein, the frame 12 is used for providing a bracket for installing and fixing a feeding device, a conveying device, a detecting device, a screening device and a control system, and is also a basic part of the whole online detecting device, so that the subsequent adjustment, use and function optimization can be facilitated.

The feeding device comprises a feeding chamber 1 and a baffle brush 2, the feeding chamber 1 is arranged on a frame 12 and is used for containing corn seeds 9 to be detected, when the whole online detection device provided by the scheme works, the corn seeds 9 fall into a conveying device from the feeding chamber 1 in an irregular mode under the action of gravity, when the conveying device conveys the corn seeds 9, the baffle brush 2 can block redundant corn seeds 9, each sample loading groove 19 only leaves one seed, baffles can be arranged on two sides of the baffle brush 2 and are used for preventing the corn seeds 9 from sliding out of a conveying belt 7 when being piled up.

A switch is arranged at the discharge hole of the feeding chamber 1 or inside the feeding chamber and is used for controlling the falling control of the corn seeds 9.

The conveyer includes conveyer belt 7 and driving motor 18, installs on frame 12, has offered a plurality of evenly distributed's dress appearance groove 19 on the conveyer belt 7, can make maize seed 9 fall into wherein completely, in fact, the conveyer belt 7 adopts flexible material processing to form, and the dress appearance groove 19 of seting up can be changed according to the size of waiting to detect maize seed 9 to satisfy the demand of detecting different varieties, in addition, when driving motor 18 during operation, whole conveyer belt 7 can rotate clockwise under the motor drive, thereby drives the seed and accomplish subsequent detection operation.

The detection device comprises a hyperspectral imaging system 3 and a connecting wire 5, and is arranged on a rack 12. The hyperspectral imaging system 3 comprises a spectral camera 11, which is arranged above the conveyor belt 7. The hyperspectral imaging system 3 is also provided with a corresponding halogen lamp light source, optical module and sensor module. The halogen lamp light source is arranged around the spectrum camera 11 and forms an included angle with the horizontal plane, and the light source illuminates the inside of the whole hyperspectral imaging system 3 and seeds in the sample loading groove 19 on the conveyor belt 7, so that a proper illumination environment is provided for the optical module. The optical module is used for collecting spectra and images in real time, the sensing module is used for detecting operation parameters of the spectrum camera 11 and performing position fine adjustment, and the distance between the sensing module and seeds to be detected, the focal length, the frame rate, the exposure time and the like can be adjusted according to different detection requirements and environments. In addition, the spectrum camera 11 may be replaced with a hyperspectral camera 11, a multispectral camera 11, a fluorescence spectrum camera 11, a near infrared spectrum camera 11, or the like according to actual demands.

When the whole online detection device works, a light source irradiates seeds to be detected on the conveyor belt 7, and reflected light of the seeds is captured by the spectrum camera 11 through a lens to obtain a spectrum image and spectrum information. With the continuous running of the seeds driven by the conveyor belt 7, continuous spectrum images and real-time spectrum information can be obtained, all data are recorded by a computer 6 in the control system, and a three-dimensional data cube containing spectrum and image information is obtained. By analyzing the spectrum data, the internal and external information such as seed size, color, texture, moisture, starch, defects, mildew and the like can be obtained, and the automatic screening of low-quality seeds is realized by the subsequent machine learning algorithm and the cooperative control of the mechanical arm 4.

The screening device comprises a mechanical arm 4, a low-quality chamber 10, a high-quality chamber 8, a fan 13 and a ventilation pipeline 14, and is arranged on a frame 12. The arm 4 includes a rotary joint, suction holes and a rotary base 17, and when the corn seeds 9 which do not meet the quality requirements pass through the area, the control system sends commands to the arm 4 and the fan 13. The fan 13 starts to operate so that the air pressure in the ventilation pipeline 14 is lower than the atmospheric pressure, and negative pressure is formed in the adsorption hole to generate the capability of adsorbing seeds. The adsorption holes adopt a mesh screen structure, and seeds are not sucked into the ventilation pipeline 14. Simultaneously, the mechanical arm 4 rotates to enable the adsorption hole to be opposite to the position of the sample loading groove 19 to be removed, and seeds are adsorbed. Then, the mechanical arm 4 is rotated again to the low-quality chamber 10, and the working state of the fan 13 is released, so that seeds naturally fall into the low-quality chamber 10 from the adsorption hole. The rotation joint is used for executing the movement of the mechanical arm 4 up and down (180 DEG in total), and the rotation base 17 is used for executing the movement of the mechanical arm 4 left and right (180 DEG in total). The plurality of air suction type mechanical arms 4 simultaneously perform collaborative operation, so that screening efficiency is improved. The other seeds meeting the quality requirements continue to rotate along with the conveyor belt 7 and naturally drop into the high-quality chamber 8, so that the screening process of the seed quality is completed.

The control system comprises a connecting wire 5 and a computer 6, and the detecting device and the screening device are connected through the connecting wire 5. The various microprocessors in the computer 6 may coordinate and process different tasks. The computer 6 analyzes and processes the hyperspectral image of the corn seeds 9 to be detected and establishes a quality identification model of the corn seeds 9. The computer 6 controls the operation of opening and closing the blower 13 and the operation of screening the plurality of robot arms 4 by the microprocessor. In the cooperative operation, the control system not only sends an instruction to the detection device to collect the spectrum image of the seed, but also positions the position coordinates (X, Y) of the seed in the sample loading slot 19 in the spectrum collection area, wherein X represents the number of rows and Y represents the number of columns. If a seed which does not meet the quality requirement is detected, the computer 6 records the position coordinate information of the seed and sends the position coordinate information of the seed to the screening device. When the conveyor belt 7 moves forward to the screening device area, the robot arm 4 will, according to the instructions of the control system, suck the seeds in the position of the specific loading slot 19 to transfer to the low-quality chamber 10 area.

The establishment of the quality identification model of the corn seeds 9 and the detection and screening process of the seeds which do not meet the quality requirements are as follows:

(1) The spectral information of the corn seeds 9 is acquired by means of a detection device. First, a batch of corn seeds 9 of known good quality is selected. Then, the characteristic parameters such as shape, color, texture, spectral reflectance and the like are counted by the computer 6, and a label is made for the characteristic parameters to form training data.

(2) And inputting the collected hyperspectral image and the statistical characteristic parameters into a deep convolutional neural network, and training the deep convolutional neural network by utilizing the data and the corresponding labels to obtain the corn seed 9 quality identification model.

(3) In cooperation, conveyor 7 transports corn seeds 9 to the area of hyperspectral imaging system 3, and their hyperspectral images and characteristic parameters are acquired using a spectral camera 11.

(4) And then sending the acquired spectral image and parameters into a trained deep convolutional neural network model, judging the quality of the seeds to be detected through the trained identification model, and displaying the result on a liquid crystal display screen of the computer 6.

According to the identification result, if seeds with unsatisfactory quality are found, a microprocessor in the computer 6 sends instructions to the screening device to control the fan 13 and the mechanical arms 4 to cooperatively work, so that seeds with quality problems are removed.

Wherein the feeding chamber 1 is used for feeding and the baffle brush 2 is used for limiting the movement of the corn seeds 9 on the conveyor belt 7, so that the seeds can only move into the detection area through the sample loading groove 19, and the redundant seeds are accumulated on one side of the baffle brush 2.

Example 2:

an online detection device for the quality of the corn seeds 9 by combining the mechanical arm 4 with the air suction type shown in the figures 1, 2 and 4 comprises a frame 12, a feeding device fixedly arranged on the frame 12, a conveying device, a detection device, a screening device and a control system;

the conveying device comprises a conveying belt 7 and a driving motor 18, and the driving motor 18 drives the conveying belt 7 to operate;

the feeding device is arranged above the conveying device and is in sliding connection with the conveying belt 7, and comprises a feeding chamber 1 and a baffle brush 2;

the detection device comprises a hyperspectral imaging system 3 and a connecting wire 5, wherein the hyperspectral imaging system 3 is arranged right above the conveyor belt 7;

the screening device comprises one or more groups of screening devices, wherein each group of screening device comprises a mechanical arm 4, a low-quality chamber 10, a high-quality chamber 8, a fan 13 and a ventilation pipeline 14, and the high-quality chamber 8 is arranged right below the tail end of the conveying direction of the conveying belt 7;

the control system comprises a connecting wire 5 and a computer 6, wherein the computer 6 is connected with the detection device and the screening device through the connecting wire 5.

The outer surface of the conveyor belt 7 is fixedly provided with a plurality of sample loading grooves 19 with the same distance.

The feeding chamber 1 upper end has seted up the pan feeding mouth, and the discharge gate has been seted up to this feeding chamber 1 lower extreme, and fixed mounting has baffle brush 2 on this feeding chamber 1's the lateral wall, and this discharge gate is arranged in directly over conveyer belt 7, has the first clearance of ejection of compact between this discharge gate and the conveyer belt 7, and this baffle brush 2 also is arranged in directly over the conveyer belt 7, is equipped with the second clearance of ejection of compact between this baffle brush 2 and the conveyer belt 7.

The hyperspectral imaging system 3 comprises a camera 11 which is arranged right above the conveyor belt 7, and the hyperspectral imaging system 3 is also provided with a corresponding halogen lamp light source, an optical module and a sensor module; the halogen lamp light source is arranged around the camera 11, forms an included angle of 45 degrees with the upper end face of the conveyor belt 7, is electrically connected with the camera 11 and is used for collecting spectra and images in real time, and the sensing module is also electrically connected with the camera 11 and is used for detecting operation parameters of the camera 11 and performing position fine adjustment.

The camera 11 includes one of a hyperspectral camera 11, a multispectral camera 11, a fluorescence spectrum camera 11, and a near infrared spectrum camera 11.

The mechanical arm 4 is fixedly arranged on one side or two sides of the frame 12, a rotating base 17 is fixedly arranged at the lower end of the mechanical arm 4, the rotating base 17 is fixedly arranged on the frame 12, an adsorption seat 16 is fixedly arranged at the tail end of the mechanical arm 4, an adsorption hole is fixedly formed in the adsorption seat 16, the adsorption seat 16 is fixedly connected with a ventilation pipeline 14, and the other end of the ventilation pipeline 14 is fixedly connected with the fan 13.

A filter screen is fixedly arranged at the adsorption hole of the adsorption seat 16.

The frame 12 is fixedly provided with a low-quality chamber 10 on one side of the mechanical arm 4.

The height of the first gap is greater than the height of the second gap, which is not greater than the height of the normal seed (taking the average of one hundred or more sub-thicknesses), and the height of the second gap is not greater than one third of the height of the normal seed.

Wherein the baffle brush 2 is conical, the conical tip faces the direction of the feeding chamber 1, a first gap is used for the first arrangement of corn seeds 9, namely, the width of a lower discharging hole of the feeding chamber is the same as that of a conveying belt, the seeds are slidably connected with the conveying belt along the lower discharging hole of the feeding chamber, a plurality of sample loading grooves 19 are formed in the conveying belt, when the seeds pass through the lower part of the feeding chamber, the seeds are carried away from the feeding chamber through the first gap by the sample loading grooves 19, the height of the sample loading grooves is the same as or slightly lower than the thickness of normal seeds, the main purpose of the first gap is to prevent excessive discharge, so that the seeds fall on the conveying belt in a large amount, meanwhile, the seeds can be prevented from being blocked by the lower end of the feeding chamber to damage the seeds, a second gap is used for screening out a plurality of seeds stacked in the sample loading grooves 19, the seeds are prevented from being stacked in the sample loading groove 19, meanwhile, the depth of the sample loading groove 19 is required to be 70% -90% of the normal seed thickness, through experiments, when the seed falls into the sample loading groove 19 and is lower than 50%, the seed is easy to sweep out of the sample loading groove 19, but is larger than 60%, and the larger the depth falling into the sample loading groove 19 is, the less easy to sweep out of the sample loading groove 19, so that when two seeds are stacked, even if the seed falling into the sample loading groove 19 is smaller in thickness, the seed above the seed is easy to sweep out of the sample loading groove 19, the detection accuracy of the rear camera 11 is improved, and the conical baffle brush 2 can block the brushed seed and move the seed on the conveyor belt 7 out of the conveyor belt 7 through the power of the conveyor belt 7.

The above description is merely of preferred embodiments of the present invention, and the scope of the present invention is not limited to the above embodiments, but all equivalent modifications or variations according to the present disclosure will be within the scope of the claims.

Claims

1. Mechanical arm (4) cooperate air-aspiration type corn seed (9) quality on-line measuring device, its characterized in that: comprises a frame (12), a feeding device fixedly arranged on the frame (12), a conveying device, a detecting device, a screening device and a control system;

the conveying device comprises a conveying belt (7) and a driving motor (18), and the driving motor (18) drives the conveying belt (7) to operate;

the feeding device is arranged above the conveying device and is in sliding connection with the conveying belt (7), and comprises a feeding chamber (1) and a baffle brush (2);

the detection device comprises a hyperspectral imaging system (3) and a connecting wire (5), wherein the hyperspectral imaging system (3) is arranged right above the conveyor belt (7);

the screening device comprises one or more groups of screening devices, wherein each group of screening device comprises a mechanical arm (4), a low-quality chamber (10), a high-quality chamber (8), a fan (13) and a ventilation pipeline (14), and the high-quality chamber (8) is arranged under the tail end of the conveying belt (7) in the conveying direction;

the control system comprises a connecting wire (5) and a computer (6), wherein the computer (6) is connected with the detection device and the screening device through the connecting wire (5).

2. The mechanical arm (4) cooperative air suction type corn seed (9) quality online detection device according to claim 1, wherein: the outer surface of the conveyor belt (7) is fixedly provided with a plurality of sample loading grooves (19) with the same distance.

3. The mechanical arm (4) cooperative air suction type corn seed (9) quality online detection device according to claim 1, wherein: the feeding chamber (1) upper end has seted up the pan feeding mouth, and the discharge gate has been seted up to this feeding chamber (1) lower extreme, and fixed mounting has baffle brush (2) on the lateral wall of this feeding chamber (1), and this discharge gate is arranged in directly over conveyer belt (7), has the first clearance of ejection of compact between this discharge gate and conveyer belt (7), and this baffle brush (2) also is arranged in directly over conveyer belt (7), is equipped with the second clearance of ejection of compact between this baffle brush (2) and conveyer belt (7).

4. The mechanical arm (4) cooperative air suction type corn seed (9) quality online detection device according to claim 1, wherein: the hyperspectral imaging system (3) comprises a camera (11) which is arranged right above the conveyor belt (7), and the hyperspectral imaging system (3) is also provided with a corresponding halogen lamp light source, an optical module and a sensor module; the halogen lamp light source is arranged around the camera (11), forms an included angle of 45 degrees with the upper end face of the conveyor belt (7), the optical module is electrically connected with the camera (11) for collecting spectra and images in real time, and the sensing module is electrically connected with the camera (11) for detecting operation parameters of the camera (11) and performing position fine adjustment.

5. The mechanical arm (4) cooperative air suction type corn seed (9) quality online detection device according to claim 4, wherein: the camera (11) comprises one of a hyperspectral camera (11), a multispectral camera (11), a fluorescence spectrum camera (11), and a near infrared spectrum camera (11).

6. The mechanical arm (4) cooperative air suction type corn seed (9) quality online detection device according to claim 1, wherein: the mechanical arm (4) is fixedly arranged on one side or two sides of the frame (12), a rotating base (17) is fixedly arranged at the lower end of the mechanical arm (4), the rotating base (17) is fixedly arranged on the frame (12), an adsorption seat (16) is fixedly arranged at the tail end of the mechanical arm (4), an adsorption hole is fixedly formed in the adsorption seat (16), a ventilation pipeline (14) is fixedly connected to the adsorption seat (16), and the other end of the ventilation pipeline (14) is fixedly connected with the fan (13).

7. The mechanical arm (4) cooperative air suction type corn seed (9) quality online detection device according to claim 6, wherein: and a filter screen is fixedly arranged at the adsorption hole of the adsorption seat (16).

8. The mechanical arm (4) cooperative air suction type corn seed (9) quality online detection device according to claim 1, wherein: a low-quality chamber (10) is fixedly arranged on one side of the mechanical arm (4) on the frame (12).

9. A mechanical arm (4) and air suction type corn seed (9) quality online detection device according to claim 3, characterized in that: the height of the first gap is larger than that of the second gap, the height of the first gap is not higher than that of the normal seeds, and the height of the second gap is not higher than one third of that of the normal seeds.

10. A detection method of a detection apparatus according to any one of claims 1 to 9, comprising the steps of:

s1, feeding, namely pouring seeds to be detected into a feeding device in batches, and naturally dropping the seeds onto a conveyor belt (7) by the gravity of the seeds;

s2: secondly, feeding, namely controlling the quantity of seeds falling onto the conveyor belt (7) through a sample loading groove (19) on the conveyor belt (7), namely automatically falling into the sample loading chamber when the conveyor belt (7) with the sample loading groove (19) passes through the position right below a feeding chamber (1) of a feeding device, taking the seeds out of the feeding chamber (1) through the sample loading chamber, and transferring the seeds to a monitoring position through the conveyor belt (7);

s3: the seeds are limited, and seeds except for the sample loading groove (19) on the conveyor belt (7) are blocked by the baffle brush (2) to be transferred to a detection area;

s4: monitoring, namely illuminating and photographing the conveyor belt (7) and seeds entering the monitoring device through a light source on a camera (11), wherein photographing comprises;

s5: sampling, establishing a comparison file library through a data acquisition device, comparing the photo information newly acquired by a camera (11) with data in the comparison file library, and transmitting the comparison information to a screening device;

s6: and (3) screening, wherein abnormal seeds are adsorbed out from the conveyor belt (7) through the screening device and placed into the low-quality chamber (10), the normal seeds move to the upper part of the high-quality chamber (8) through the conveyor belt (7), and the normal seeds fall into the high-quality chamber (8) to finish monitoring.