CN108465644B - Artificial intelligent wheat quality inspection robot and quality inspection method - Google Patents
Artificial intelligent wheat quality inspection robot and quality inspection method Download PDFInfo
- Publication number
- CN108465644B CN108465644B CN201810830624.4A CN201810830624A CN108465644B CN 108465644 B CN108465644 B CN 108465644B CN 201810830624 A CN201810830624 A CN 201810830624A CN 108465644 B CN108465644 B CN 108465644B
- Authority
- CN
- China
- Prior art keywords
- machine vision
- wheat
- feeding
- quality inspection
- sensor
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B07—SEPARATING SOLIDS FROM SOLIDS; SORTING
- B07C—POSTAL SORTING; SORTING INDIVIDUAL ARTICLES, OR BULK MATERIAL FIT TO BE SORTED PIECE-MEAL, e.g. BY PICKING
- B07C5/00—Sorting according to a characteristic or feature of the articles or material being sorted, e.g. by control effected by devices which detect or measure such characteristic or feature; Sorting by manually actuated devices, e.g. switches
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B07—SEPARATING SOLIDS FROM SOLIDS; SORTING
- B07C—POSTAL SORTING; SORTING INDIVIDUAL ARTICLES, OR BULK MATERIAL FIT TO BE SORTED PIECE-MEAL, e.g. BY PICKING
- B07C5/00—Sorting according to a characteristic or feature of the articles or material being sorted, e.g. by control effected by devices which detect or measure such characteristic or feature; Sorting by manually actuated devices, e.g. switches
- B07C5/02—Measures preceding sorting, e.g. arranging articles in a stream orientating
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B07—SEPARATING SOLIDS FROM SOLIDS; SORTING
- B07C—POSTAL SORTING; SORTING INDIVIDUAL ARTICLES, OR BULK MATERIAL FIT TO BE SORTED PIECE-MEAL, e.g. BY PICKING
- B07C5/00—Sorting according to a characteristic or feature of the articles or material being sorted, e.g. by control effected by devices which detect or measure such characteristic or feature; Sorting by manually actuated devices, e.g. switches
- B07C5/36—Sorting apparatus characterised by the means used for distribution
- B07C5/361—Processing or control devices therefor, e.g. escort memory
Landscapes
- Sorting Of Articles (AREA)
Abstract
An artificial intelligent wheat quality inspection robot and a quality inspection method, wherein the artificial intelligent wheat quality inspection robot comprises an artificial intelligent wheat quality inspection robot main body and a system server; the artificial intelligent wheat quality inspection robot main body comprises a box body, wherein a feeding port and a discharging port are formed in the box body, a feeding mechanism, a conveying mechanism, an upper double-sided machine vision mechanism, a lower double-sided machine vision mechanism and a sorting mechanism are arranged in the box body, the feeding port is communicated with the feeding mechanism, the feeding mechanism is communicated with the conveying mechanism, the conveying mechanism penetrates through the upper double-sided machine vision mechanism and the lower double-sided machine vision mechanism, the conveying mechanism is communicated with the discharging port, the sorting mechanism is located behind the upper double-sided machine vision mechanism and the lower double-sided machine vision mechanism, and the upper double-sided machine vision mechanism; the feeding mechanism adopts a sensor and intelligent control mode. The invention can complete the detection of imperfect wheat grains and impurities, and solves the worldwide problem of rapid detection of imperfect wheat grains.
Description
Technical Field
The invention relates to physical detection equipment and a method for wheat quality, in particular to an artificial intelligent wheat quality detection robot and a quality detection method.
Background
The imperfect wheat grain (insect, bud, disease, damage and mildew) detection has no modern means and equipment, the wheat detection is completely finished by artificial sensory identification, the artificial sensory identification has a plurality of uncertain factors, the same sample is identified by the same detector for a plurality of times, the data of each identification is different, the same sample is used by different detectors with the same standard, and the identification conclusion data are different; therefore, it is difficult to objectively, fairly and accurately distinguish the two types of information, and the information becomes a "technical means having a standard without performing the standard".
The existing artificial sensory recognition completes the detection of imperfect wheat grains, and has the following defects: 1) are susceptible to interference from environmental and personnel factors; 2) no unified metrology tool; 3) the efficiency is low.
CN207188252U discloses a system for detecting defective grains in grain in 2018, 4 and 6, which includes feeding by a screw mechanism, feeding by an optical glass disc, taking material images by upper and lower double vision relatively, identifying the detected object and the images in an image library in a mechanical comparison manner, sorting and weighing on the glass disc, and finally completing the detection of defective grains in grain. The following problems exist with this detection device and approach:
(1) the feeding mode of the screw mechanism is usually only suitable for standard parts, and when the material is a non-standard part of grains, the conditions of material blockage, material leakage, incomplete discharging and the like easily occur due to different sizes, shapes, lengths, specific gravities and integrity degrees of particles, so that the efficiency is seriously influenced by incomplete conveyed samples.
(2) Optical glass disc pay-off mode, the easy slippage of material when the disc rotates and reaches certain rotational speed, simultaneously, because the end that produces when the disc rotates jumps and makes the sample that sends into in the visual field lack stability, further make the measurement space distortion of the material of acquireing from this, the original information of being detected the thing takes place serious deviation.
(3) The relative up and down visual mode causes the mutual interference between the upper and lower visual information, and further causes the distortion and the larger deviation of the original information of the detected object.
(4) The original image of the detected object is mechanically compared with the image library of the standard sample, so that the method cannot adapt to the nonstandard characteristic identification of the grain particles, the classification of the imperfect grain particles is not in accordance with a linear relation, and the hard linear relation discrimination by using the mechanical comparison has extremely large error in the actual work.
(5) When the detected particles reach the sorting area from the identified area, the physical position often cannot ensure the sorting accuracy because of sliding and end jumping of the disc.
Disclosure of Invention
The invention aims to solve the technical problem of providing an artificial intelligent wheat quality inspection robot for detecting a plurality of subclasses and impurities of five types including imperfect grains (insects, buds, diseases, damages and mildews) of wheat aiming at the defects of the prior art.
The invention further aims to solve the technical problem of providing a method for performing quality inspection on wheat by using the artificial intelligent wheat quality inspection robot.
The technical scheme adopted by the invention for solving the technical problems is as follows: an artificial intelligence wheat quality inspection robot comprises an artificial intelligence wheat quality inspection robot main body and a system server; the artificial intelligent wheat quality inspection robot main body comprises a box body, wherein a feeding port and a discharging port are formed in the box body, a feeding mechanism, a conveying mechanism, an upper double-sided machine vision mechanism, a lower double-sided machine vision mechanism and a sorting mechanism are arranged in the box body, the feeding port is communicated with the feeding mechanism, the feeding mechanism is communicated with the conveying mechanism, the conveying mechanism penetrates through the upper double-sided machine vision mechanism and the lower double-sided machine vision mechanism, the conveying mechanism is communicated with the discharging port, the sorting mechanism is located behind the upper double-sided machine vision mechanism and the lower double-sided machine vision mechanism, and the upper double-sided machine vision mechanism; the feeding mechanism comprises a feeding box, a clapboard is arranged in the feeding box to divide the cavity of the feeding box into a front cavity and a rear cavity, a trough which penetrates through the front cavity and the rear cavity is arranged at the bottom of the feeding box, a sensor I for monitoring whether the rear cavity contains materials is arranged on the rear cavity close to the edge of the clapboard trough, the rear cavity is provided with a sensor II for monitoring the material level in the horizontal direction, the center of the trough of the front cavity close to the partition plate is provided with a sensor III for monitoring whether the rear cavity discharges materials or not, a sensor IV for monitoring whether the front cavity is blocked is arranged on the upper edge of the outlet of the front cavity trough, a mechanism for preventing the back cavity from being blocked is arranged on the feeding box, the feeding box is characterized in that a vibrator is arranged on the lower portion of the feeding box and connected with a voltage modulator, the sensor I, the sensor II, the sensor III, the sensor IV, the mechanism for preventing material blockage of the rear cavity and the voltage modulator are connected with a controller, and the controller is connected with a system server.
Furthermore, the sensor II for monitoring the material level in the horizontal direction is arranged at a position which is more than or equal to 1cm (preferably 2-3 cm) away from the bottom of the rear cavity in the horizontal direction.
Further, the mechanism for preventing back cavity putty includes the steering wheel, the steering wheel is established outside feeding box, the steering wheel even has the steering wheel plectrum through the steering wheel connecting rod, the steering wheel plectrum is located the back intracavity, and is close to baffle and silo.
Further, transport mechanism includes transparent synchronous conveyer belt and synchronous transfer gear, synchronous transfer gear arranges on the box, be equipped with synchronous tooth I on the synchronous transfer gear, be equipped with synchronous tooth II on the edge of transparent synchronous conveyer belt, the synchronous tooth II of transparent synchronous conveyer belt and the I assembly connection of synchronous tooth of synchronous transfer gear, the upper portion of transparent synchronous conveyer belt passes two-sided machine vision mechanism from top to bottom.
Further, the upper and lower double-sided machine vision mechanism comprises an upper machine vision mechanism, an upper shadowless light source, an upper vision bottom plate, a lower machine vision mechanism, a lower shadowless light source and a lower vision bottom plate, the upper machine vision mechanism and the upper shadowless light source are located behind the upper side of the transparent synchronous conveyor belt, the upper vision bottom plate is arranged on the lower side of the transparent synchronous conveyor belt and located under the upper shadowless light source, the lower machine vision mechanism and the lower shadowless light source are located in front of the lower side of the transparent conveyor belt, and the lower vision bottom plate is arranged on the upper side of the transparent synchronous conveyor belt and located over the lower shadowless light source.
Further, letter sorting mechanism includes letter sorting sensor, material receiving box, air pin and weighing sensor, the letter sorting sensor is installed on material receiving box upper portion, and is located transparent synchronous conveyer belt top, the air pin links to each other with the air pump through the pipeline that is equipped with the pneumatic valve, and just to waiting to sort the sample on the transparent synchronous conveyer belt, weighing sensor installs in material receiving box lower part, realizes waiting to sort weighing of sample.
Further, a material guiding strip is arranged and is positioned between the feeding mechanism and the conveying mechanism.
A method for performing quality inspection on wheat by using the artificial intelligence wheat quality inspection robot comprises the following steps:
the method comprises the steps that a sample to be detected of wheat reaches a feeding mechanism through a feeding port of an artificial intelligent wheat quality detection robot main body, meanwhile, the sensor is responsible for detecting the material feeding state of the sample to be detected in the feeding mechanism, if the material blocking condition occurs, the sensor transmits material blocking information to a system server through a controller, the system server controls a voltage modulator through the controller to further drive a vibrator of the feeding mechanism to vibrate, the feeding mechanism transmits the sample to be detected to a transmission mechanism, the transmission mechanism transmits the sample to be detected of the wheat to an upper double-sided machine vision mechanism and a lower double-sided machine vision mechanism, an original image of the detected wheat sample is obtained through the upper double-sided machine vision mechanism and the lower double-sided machine vision mechanism, algorithm processing is carried out on the original image of each detected wheat sample through an artificial neural network system which is constructed on the system server and supports graphic processing, and the color, converting the characteristics of the contour, the size and the like into characteristic information values, inputting the characteristic information values into an artificial neural network system, comparing the characteristic information values with corresponding characteristic information values stored on neurons to obtain the membership degrees of the corresponding characteristic information values, judging according to the membership degrees of the corresponding characteristic information values by using a judgment function, carrying out classification statistics, and sorting imperfect grains and side-by-side impurities by using a sorting mechanism; and finally, outputting the detected sample to a discharge port by a conveying mechanism, and automatically finishing the whole quality inspection process.
Further, the corresponding characteristic information value stored on the neuron is obtained by: the method comprises the steps of conveying wheat samples classified in advance to the lower part of an upper double-sided machine vision mechanism and the lower double-sided machine vision mechanism through a feeding mechanism and a conveying mechanism, obtaining an original image of each wheat sample by the upper double-sided machine vision mechanism and the lower double-sided machine vision mechanism, carrying out algorithm processing on the original image of each wheat sample through an artificial neural network system supporting graphic processing, converting the characteristics of color, texture, outline, size and the like of the wheat into characteristic information values, and storing the characteristic information values on neurons of the artificial neural network system, thereby forming a wheat characteristic information base.
In the invention, the feeding mechanism adopts a sensor and intelligent control mode, and through the information sensed by the sensor, the system can judge whether the material is blocked or not, whether the material is empty or not and whether the material flows smoothly and uniformly and can carry out corresponding processing according to the information so as to ensure continuous, stable and uniform feeding, and the conditions of material blocking, material leakage, incomplete discharging and the like can not occur.
In the invention, the transparent synchronous conveyor belt of the conveying mechanism is driven by the synchronous conveying wheel through the motor to finish the conveying of materials, and the transparent synchronous conveyor belt is a synchronous belt with the middle part positioned at the visual field position and the edge capable of being completely meshed with the synchronous conveying wheel. The visual field part is transparent, so that the upper and lower double-sided machine vision mechanisms can acquire sufficiently clear images, the transparent synchronous conveyor belt is driven by the motor under stable frequency at the synchronous conveying wheel to move at a constant speed, the clear images are ensured, and the original information of the detected object cannot deviate.
In the invention, the upper and lower double-sided machine vision mechanisms are completely staggered at corresponding positions, so that the accuracy and integrity of the original image acquisition are not influenced by mutual interference between the two vision lenses, and the reliability and stability of the vision environment are influenced by the mutual interference between the lamplight of the two vision lenses; therefore, the accurate, real and complete measurement space information of the detected object can be obtained.
In the invention, the sorting mechanism is provided with a sensor, so that the real-time accurate position of the sorted material close to the sorting opening can be accurately obtained, and the sorting accuracy, wrong sorting and no sorting leakage can be ensured. The detected imperfect grains fall into the material receiving box, and the weighing is finished by the weighing sensor.
Research practice shows that the detection time of each wheat sample is no more than 20 milliseconds from the acquisition of the information of the wheat sample to be detected to the completion of detection judgment by using the method, and the artificial intelligent detection robot can finish identification and data statistical analysis by detecting a 50-gram standard sample (about 1200 samples) in 3-4 minutes. The whole process does not need human intervention, the problem of detection index drift caused by human factors and external environment influence due to artificial sensory detection can be effectively solved, and meanwhile, the robot detects each sample one by one without detection blind spots and dead angles.
The actual detection accuracy and repeatability indexes of the invention are as follows:
imperfect wheat grain detection (taking 50 g of each sample in national standard as an example)
Imperfect granules, error less than or equal to 0.5 percent and repetition error less than or equal to 0.3 percent;
the worm is corroded by particles, the error is less than or equal to 0.5 percent, and the repetition error is less than or equal to 0.3 percent;
germinating grains, wherein the error is less than or equal to 0.5 percent, and the repetition error is less than or equal to 0.3 percent;
scab grains (gibberellic grains) with an error of less than or equal to 0.3 percent and a repetition error of less than or equal to 0.2 percent;
the error of the scab grains (black embryos) is less than or equal to 0.2 percent, and the repetition error is less than or equal to 0.1 percent;
broken grains (broken), the error is less than or equal to 0.2 percent, and the repeat error is less than or equal to 0.1 percent;
mildew grains with error less than or equal to 0.3 percent and repetition error less than or equal to 0.2 percent;
the error of each 1000 grains of the yellow rice is less than 1 grain, and the repetition error is zero.
The invention can complete the detection of imperfect wheat grains and impurities, and solves the worldwide problem of rapid detection of imperfect wheat grains.
Drawings
FIG. 1 is a block diagram of an embodiment of an artificial intelligence wheat quality inspection robot of the present invention;
FIG. 2 is a schematic external view of the main body of the artificial intelligence wheat quality inspection robot in the embodiment shown in FIG. 1;
FIG. 3 is an internal structure view of the main body of the artificial intelligence wheat quality inspection robot in the embodiment shown in FIG. 1;
FIG. 4 is a schematic structural diagram of a feeding mechanism of the artificial intelligence wheat quality inspection robot main body of the embodiment shown in FIG. 1;
FIG. 5 is a connection diagram of a robot controller for quality inspection of wheat according to the present invention;
FIG. 6 is a schematic structural diagram of a conveying mechanism of the main body of the artificial intelligence wheat quality inspection robot in the embodiment shown in FIG. 1;
FIG. 7 is a schematic structural diagram of an upper and lower double-sided machine vision mechanism of the main body of the artificial intelligence wheat quality inspection robot in the embodiment shown in FIG. 1;
fig. 8 is a schematic structural diagram of a sorting mechanism of the artificial intelligence wheat quality inspection robot main body of the embodiment shown in fig. 1;
FIG. 9 is a flow chart of the quality inspection process of the artificial intelligence wheat quality inspection robot.
Detailed Description
The invention is further explained with reference to the drawings and the embodiments.
Examples
Referring to fig. 1-5, an artificial intelligence wheat quality inspection robot comprises an artificial intelligence wheat quality inspection robot main body 1 and a system server 2; the artificial intelligent wheat quality inspection robot main body 1 comprises a box body 1-1, a feeding port 1-2 and a discharging port 1-3 are arranged on the box body 1-1, a feeding mechanism 1-4, a conveying mechanism 1-5, an upper and lower double-sided machine vision mechanism 1-6 and a sorting mechanism 1-7 are arranged in the box body 1-1, the feeding port 1-2 is communicated with the feeding mechanism 1-4, the feeding mechanism 1-4 is communicated with the conveying mechanism 1-5, the conveying mechanism 1-5 penetrates through the upper and lower double-sided machine vision mechanism 1-6, the conveying mechanism 1-5 is communicated with the discharging port 1-3, the sorting mechanism 1-7 is positioned behind the upper and lower double-sided machine vision mechanism 1-6, and the upper and lower double-sided machine vision mechanism 1-6, The sorting mechanisms 1-7 are connected with the system server 2;
the feeding mechanism 1-4 comprises a feeding box 1-4-1, a partition plate 1-4-2 is arranged in the feeding box 1-4-1, the cavity of the feeding box 1-4-1 is divided into a front cavity and a rear cavity, a linear trough 1-4-3 penetrating through the front cavity and the rear cavity is arranged at the bottom of the feeding box 1-4-1, queuing of irregular grain type wheat particles is facilitated, a sensor I1-4-4 for monitoring whether the rear cavity is filled with materials is arranged on the side, close to the partition plate trough, of the rear cavity, a sensor II 1-4-5 for monitoring the material level in the horizontal direction is arranged at a position 2cm away from the bottom of the rear cavity in the horizontal direction, a sensor III 1-4-6 for monitoring whether the rear cavity is filled with materials is arranged at the center, close to the partition plate, and a sensor IV 1-4-7 for monitoring whether the front cavity is filled with materials is arranged at the outlet of the front cavity A steering engine 1-4-8 is arranged outside the feeding box 1-4-1, the steering engine 1-4-8 is connected with a steering engine plectrum 1-4-9 through a steering engine connecting rod, the steering engine plectrum 1-4-9 is positioned in the rear cavity, and is close to the clapboard 1-4-2 and the linear trough 1-4-3, the lower part of the feeding box 1-4-1 is provided with a vibrator 1-4-10, the vibrator 1-4-10 is connected with a voltage modulator 4, the sensor I1-4-4, the sensor II 1-4-5, the sensor III 1-4-6, the sensor IV 1-4-7, the steering engine 1-4-8 and the voltage modulator 4 are connected with a controller 3, and the controller 3 is connected with a system server 2;
the working process is as follows:
1) a sensor II 1-4-5 for monitoring the material level in the horizontal direction is triggered, the controller 3 obtains information and sends the information to the system server 2, the system server 2 sends out a sampling stopping instruction, and the sensor is used for online detection (online automatic sampling);
2) the method comprises the following steps that a front cavity of a feeding mechanism is blocked, a sensor IV 1-4-7 for monitoring whether the front cavity is blocked is triggered, a controller 3 obtains information and sends the information to a system server 2, the system server 2 firstly sends a sampling stopping instruction, then the system server 2 sends an instruction to the controller 3, the controller 3 controls a voltage modulator 4, the vibration frequency of a vibrator is strengthened for 2 seconds, and then the vibration frequency of the vibrator is recovered; if the sensor IV 1-4-7 for monitoring whether the front cavity is blocked is not triggered to be released, the vibration frequency of the vibrator is enhanced again for 2 seconds, and the steps are repeated for 3 times until the front cavity blocking sensor is triggered to be released;
3) a sensor III 1-4-6 for monitoring whether the rear cavity discharges materials is triggered, a sensor I1-4-4 for monitoring whether the rear cavity discharges materials is not triggered, the situation that the rear cavity of the feeding mechanism is blocked is indicated, the controller obtains sensing information, a steering engine 1-4-8 is started, a steering engine shifting piece 1-4-9 shifts a rear cavity discharge hole sample, and the rear cavity blocking state is relieved; after the rear cavity samples are all conveyed through the feeding mechanism, the sensor I1-4-4 for monitoring whether the rear cavity is filled with materials is triggered, and the controller 3 sends information that the samples are all checked to the system server 2.
The input of the controller 3 is connected with each sensor (connected with a sensing signal-a blocking sensing signal and a full sensing signal), the output is connected with the voltage modulator 4, the voltage modulator 4 is connected with the vibrator 1-4-10 on the feeding mechanism, and the purpose of controlling the amplitude of the vibrator 1-4-10 is achieved by modulating the voltage of the vibrator 1-4-10 on the feeding mechanism.
The system server 2 is a command center of the wheat quality inspection robot and is the brain of the wheat quality inspection robot, the system server is connected with the controller through an RS232 communication port, the controller is connected with the sensors, the controller acquires the states of the sensors and sends the information of the sensors to the system server, and the system server sends instructions to the controller according to the acquired sensing information to complete material blockage treatment; the system server is connected with the upper and lower double-sided machine vision mechanisms through an RJ45 network interface to acquire picture information, then carries out de-noising processing on the pictures, carries out analysis, reasoning, judgment and judgment, and completes command coordination of all steps of the wheat quality inspection robot.
The steering engine moves and dredges the putty when the front cavity baffle plate and the rear cavity baffle plate are plugged. The feeding is controlled in an intelligent mode, and through information sensed by a sensor, the system can judge whether the material is blocked or not, whether the material is empty or not and whether the material flows smoothly and uniformly or not and can carry out corresponding processing according to the information so as to ensure continuous, stable and uniform feeding.
Referring to fig. 6, the conveying mechanism 1-5 comprises a transparent synchronous conveying belt 1-5-1 and a synchronous conveying wheel 1-5-2, the synchronous conveying wheel 1-5-2 is arranged on a box body 1-1 in a rectangular shape, a synchronous tooth I is arranged on the synchronous conveying wheel 1-5-2, a synchronous tooth II 1-5-3 is arranged on the edge of the transparent synchronous conveying belt 1-5-1, the synchronous tooth II 1-5-3 of the transparent synchronous conveying belt 1-5-1 is connected with the synchronous tooth I of the synchronous conveying wheel 1-5-2 in an assembling mode, and the upper portion of the transparent synchronous conveying belt 1-5-1 penetrates through an upper double-sided machine vision mechanism 1-6 and a lower double-sided machine vision mechanism 1-6.
Referring to FIG. 7, the upper and lower double-sided machine vision mechanism 1-6 comprises an upper machine vision mechanism 1-6-1, an upper shadowless light source 1-6-2, an upper vision bottom plate 1-6-3, a lower machine vision mechanism 1-6-4, a lower shadowless light source 1-6-5 and a lower vision bottom plate 1-6-6, the upper machine vision mechanism 1-6-1 and the upper shadowless light source 1-6-2 are located behind the upper side of the transparent synchronous conveyor belt 1-5-1, the upper vision bottom plate 1-6-3 is located on the lower side of the transparent synchronous conveyor belt 1-5-1 and is located under the upper shadowless light source 1-6-2, the lower machine vision mechanism 1-6-4 and the lower shadowless light source 1-6-5 are located in front of the transparent synchronous conveyor belt 1-5-1, the lower vision bottom plate 1-6-6 is arranged on the upper side of the transparent synchronous conveyor belt 1-5-1 and is positioned right above the lower shadowless light source 1-6-5.
Referring to fig. 8, the sorting mechanism 1-7 comprises a sorting sensor 1-7-1, a material receiving box 1-7-2, an air needle 1-7-3 and a weighing sensor 1-7-4, the sorting sensor 1-7-1 is installed on the upper portion of the material receiving box 1-7-2 and is located above a transparent synchronous conveyor belt 1-5-1, the air needle 1-7-3 is connected with an air pump through a pipeline provided with an air valve and is opposite to a sample 1-7-5 to be sorted on the transparent synchronous conveyor belt 1-5-1, and the weighing sensor 1-7-4 is installed on the lower portion of the material receiving box 1-7-2 to weigh the sample 1-7-5 to be sorted. When the automatic picking device works, the sorting sensor 1-7-1 senses a sample 1-7-5 to be picked, the controller controls the air needle 1-7-3 to be opened, the sample 1-7-5 to be picked is blown to the material receiving box 1-7-2, and the weighing sensor 1-7-4 is used for weighing.
In this embodiment, a material guiding strip (not shown) is further provided, and the material guiding strip is located between the feeding mechanism 1-4 and the conveying mechanism 1-5.
Referring to fig. 9, a method for performing quality inspection on wheat by using the artificial intelligence wheat quality inspection robot comprises the following steps:
the method comprises the steps that a sample to be detected of wheat reaches a feeding mechanism through a feeding port of an artificial intelligent wheat quality detection robot main body, meanwhile, the sensor is responsible for detecting the material feeding state of the sample to be detected in the feeding mechanism, if the material blocking condition occurs, the sensor transmits material blocking information to a system server through a controller, the system server controls a voltage modulator through the controller to further drive a vibrator of the feeding mechanism to vibrate, the feeding mechanism transmits the sample to be detected to a transmission mechanism, the transmission mechanism transmits the sample to be detected of the wheat to an upper double-sided machine vision mechanism and a lower double-sided machine vision mechanism, an original image of the detected wheat sample is obtained through the upper double-sided machine vision mechanism and the lower double-sided machine vision mechanism, algorithm processing is carried out on the original image of each detected wheat sample through an artificial neural network system which is constructed on the system server and supports graphic processing, and the color, converting the characteristics such as the contour, the size and the like into characteristic information values, inputting the characteristic information values into an artificial neural network system, comparing the characteristic information values with corresponding characteristic information values stored on neurons to obtain the membership degrees of the corresponding characteristic information values, judging according to the membership degrees of the corresponding characteristic information values by using an S-shaped judging function, carrying out classified statistics, and sorting imperfect grains and side-by-side impurities by using a sorting mechanism; and finally, outputting the detected sample to a discharge port by a conveying mechanism, and automatically finishing the whole quality inspection process.
The corresponding characteristic information value stored on the neuron is obtained by the following method: 20000 wheat samples (including insects, buds, diseases, damages and mildews) classified in advance are sent to the lower and upper double-sided machine vision mechanisms through the feeding mechanism and the conveying mechanism, the upper and lower double-sided machine vision mechanisms acquire an original image of each wheat sample, the original image of each wheat sample is subjected to algorithm processing through an artificial neural network system supporting graphic processing, the characteristics of the wheat, such as color, texture, contour, size and the like, are converted into characteristic information values, and the characteristic information values are stored on neurons of the artificial neural network system, so that a wheat characteristic information base is formed.
Research practice shows that the detection time of each wheat sample is no more than 20 milliseconds from the acquisition of the information of the wheat sample to be detected to the completion of detection judgment by using the method, and the artificial intelligent detection robot can finish identification and data statistical analysis by detecting a 50-gram standard sample in 3-4 minutes. The whole process does not need human intervention, the problem of detection index drift caused by human factors and external environment influence due to artificial sensory detection can be effectively solved, and meanwhile, the robot detects each sample one by one without detection blind spots and dead angles.
The actual detection accuracy and repeatability indexes of the invention are as follows:
imperfect wheat grain detection (taking 50 g of each sample in national standard as an example)
Imperfect granules, error less than or equal to 0.5 percent and repetition error less than or equal to 0.3 percent;
the worm is corroded by particles, the error is less than or equal to 0.5 percent, and the repetition error is less than or equal to 0.3 percent;
germinating grains, wherein the error is less than or equal to 0.5 percent, and the repetition error is less than or equal to 0.3 percent;
scab grains (gibberellic grains) with an error of less than or equal to 0.3 percent and a repetition error of less than or equal to 0.2 percent;
the error of the scab grains (black embryos) is less than or equal to 0.2 percent, and the repetition error is less than or equal to 0.1 percent;
broken grains (broken), the error is less than or equal to 0.2 percent, and the repeat error is less than or equal to 0.1 percent;
mildew grains with error less than or equal to 0.3 percent and repetition error less than or equal to 0.2 percent;
the error of each 1000 grains of the yellow rice is less than 1 grain, and the repetition error is zero.
Claims (7)
1. The artificial intelligent wheat quality inspection robot comprises an artificial intelligent wheat quality inspection robot main body and a system server; the artificial intelligent wheat quality inspection robot main body comprises a box body, wherein a feeding port and a discharging port are formed in the box body, a feeding mechanism, a conveying mechanism, an upper double-sided machine vision mechanism, a lower double-sided machine vision mechanism and a sorting mechanism are arranged in the box body, the feeding port is communicated with the feeding mechanism, the feeding mechanism is communicated with the conveying mechanism, the conveying mechanism penetrates through the upper double-sided machine vision mechanism and the lower double-sided machine vision mechanism, the conveying mechanism is communicated with the discharging port, the sorting mechanism is located behind the upper double-sided machine vision mechanism and the lower double-sided machine vision mechanism, and the upper double-sided machine vision mechanism; the method is characterized in that: the feeding mechanism comprises a feeding box, a clapboard is arranged in the feeding box to divide the cavity of the feeding box into a front cavity and a rear cavity, a trough which penetrates through the front cavity and the rear cavity is arranged at the bottom of the feeding box, a sensor I for monitoring whether the rear cavity contains materials is arranged on the rear cavity close to the edge of the clapboard trough, the rear cavity is provided with a sensor II for monitoring the material level in the horizontal direction, the center of the trough of the front cavity close to the partition plate is provided with a sensor III for monitoring whether the rear cavity discharges materials or not, a sensor IV for monitoring whether the front cavity is blocked is arranged on the upper edge of the outlet of the front cavity trough, a mechanism for preventing the back cavity from being blocked is arranged on the feeding box, the lower part of the feeding box is provided with a vibrator, the vibrator is connected with a voltage modulator, the sensor I, the sensor II, the sensor III, the sensor IV, the mechanism for preventing the rear cavity from being blocked and the voltage modulator are connected with a controller, and the controller is connected with a system server;
the conveying mechanism comprises a transparent synchronous conveying belt and a synchronous conveying wheel, the synchronous conveying wheel is arranged on the box body, a synchronous tooth I is arranged on the synchronous conveying wheel, a synchronous tooth II is arranged on the edge of the transparent synchronous conveying belt, the synchronous tooth II of the transparent synchronous conveying belt is assembled and connected with the synchronous tooth I of the synchronous conveying wheel, and the upper part of the transparent synchronous conveying belt penetrates through the upper and lower double-sided machine vision mechanisms;
the upper and lower double-sided machine vision mechanism comprises an upper machine vision mechanism, an upper shadowless light source, an upper vision bottom plate, a lower machine vision mechanism, a lower shadowless light source and a lower vision bottom plate, the upper machine vision mechanism and the upper shadowless light source are located at the rear of the upper side of the transparent synchronous conveyor belt, the upper vision bottom plate is arranged on the lower side of the transparent synchronous conveyor belt and located under the upper shadowless light source, the lower machine vision mechanism and the lower shadowless light source are located in front of the lower side of the transparent synchronous conveyor belt, and the lower vision bottom plate is arranged on the upper side of the transparent synchronous conveyor belt and located under the lower shadowless light source.
2. The artificial intelligence wheat quality inspection robot of claim 1, wherein: and the sensor II for monitoring the material level in the horizontal direction is arranged at the position, which is more than or equal to 1cm away from the bottom, of the rear cavity in the horizontal direction.
3. The artificial intelligence wheat quality inspection robot of claim 1 or 2, wherein: the mechanism for preventing back cavity putty includes the steering wheel, the steering wheel is established outside feeding box, the steering wheel even has the steering wheel plectrum through the steering wheel connecting rod, the steering wheel plectrum is located the back intracavity, and is close to baffle and silo.
4. The artificial intelligence wheat quality inspection robot of claim 1 or 2, wherein: the sorting mechanism comprises a sorting sensor, a material receiving box, an air needle and a weighing sensor, the sorting sensor is installed on the upper portion of the material receiving box and located above the transparent synchronous conveying belt, the air needle is connected with an air pump through a pipeline provided with an air valve and is just opposite to a sample to be sorted on the transparent synchronous conveying belt, the weighing sensor is installed on the lower portion of the material receiving box, and weighing of the sample to be sorted is achieved.
5. The artificial intelligence wheat quality inspection robot of claim 1 or 2, wherein: and the guide strip is positioned between the feeding mechanism and the conveying mechanism.
6. A method for performing quality inspection on wheat by using the artificial intelligence wheat quality inspection robot as claimed in claim 1, which is characterized by comprising the following steps:
the method comprises the steps that a wheat sample to be detected reaches a feeding mechanism through a feeding port of an artificial intelligent wheat quality inspection robot main body, the material feeding state of the sample to be detected in the feeding mechanism is detected through a sensor, if the material blocking condition occurs, the sensor transmits material blocking information to a system server through a controller, the system server controls a voltage modulator through the controller to further drive a vibrator of the feeding mechanism to vibrate, the feeding mechanism transmits the sample to be detected to a transmission mechanism one by one, the transmission mechanism transmits the sample to be detected to an upper double-sided machine vision mechanism and a lower double-sided machine vision mechanism, an original image of the detected wheat sample is obtained through the upper double-sided machine vision mechanism and the lower double-sided machine vision mechanism, the original image of each detected wheat sample is processed through an artificial neural network system which is constructed on the system server and supports graphic processing, and the characteristic of each detected wheat sample is converted into a characteristic, inputting an artificial neural network system, comparing with corresponding characteristic information values stored on neurons to obtain the membership degree of the corresponding characteristic information values, judging according to the membership degree of the corresponding characteristic information values by using a judgment function, carrying out classified statistics, and sorting imperfect grains and side-by-side impurities by using a sorting mechanism; and finally, outputting the detected sample to a discharge port by a conveying mechanism, and automatically finishing the whole quality inspection process.
7. The method for wheat quality inspection by the artificial intelligence wheat quality inspection robot as claimed in claim 6, wherein the corresponding characteristic information value stored on the neuron is obtained by:
the method comprises the steps that wheat samples classified in advance are sent to the lower part of an upper double-sided machine vision mechanism and a lower double-sided machine vision mechanism through a feeding mechanism and a conveying mechanism, the upper double-sided machine vision mechanism and the lower double-sided machine vision mechanism acquire original images of each wheat sample, algorithm processing is carried out on the original images of each wheat sample through an artificial neural network system supporting graph processing, characteristics of wheat are converted into characteristic information values, and the characteristic information values are stored on neurons of the artificial neural network system, so that a wheat characteristic information base is formed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810830624.4A CN108465644B (en) | 2018-07-26 | 2018-07-26 | Artificial intelligent wheat quality inspection robot and quality inspection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810830624.4A CN108465644B (en) | 2018-07-26 | 2018-07-26 | Artificial intelligent wheat quality inspection robot and quality inspection method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108465644A CN108465644A (en) | 2018-08-31 |
| CN108465644B true CN108465644B (en) | 2020-06-02 |
Family
ID=63259898
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810830624.4A Active CN108465644B (en) | 2018-07-26 | 2018-07-26 | Artificial intelligent wheat quality inspection robot and quality inspection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108465644B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111359907B (en) * | 2020-02-24 | 2021-08-20 | 南京农业大学 | Real-time online detection system and method for wheat scab infection particle rate based on embedded deep learning |
| CN111389747A (en) * | 2020-04-26 | 2020-07-10 | 成都复星联创科技有限公司 | Bamboo strip intelligent sorting system and method based on visual identification |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| SU971524A1 (en) * | 1981-04-09 | 1982-11-07 | Центральный научно-исследовательский институт оловянной промышленности | Self-oscillating separator |
| CN202070498U (en) * | 2011-04-15 | 2011-12-14 | 伟钧精密机械有限公司 | Object detector |
| JP5793942B2 (en) * | 2011-04-27 | 2015-10-14 | シンフォニアテクノロジー株式会社 | Goods separation device |
| CN104475362B (en) * | 2014-12-08 | 2016-11-30 | 浙江大学 | A kind of Fructus Jujubae based on machine vision technique detects and grading plant in real time |
| CN207507850U (en) * | 2017-02-08 | 2018-06-19 | 云南万视智能设备有限公司 | A kind of crawler type color sorting machine |
| CN106743191A (en) * | 2017-03-27 | 2017-05-31 | 重庆市永川区新窝建材有限公司 | A kind of oscillation feed gear |
-
2018
- 2018-07-26 CN CN201810830624.4A patent/CN108465644B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN108465644A (en) | 2018-08-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102829726B (en) | Machine vision automatic detection system for rivets | |
| CN109230351B (en) | Method for identifying abnormal operation of belt conveyor | |
| US9024223B2 (en) | Optical type granule sorting machine | |
| CN102217559B (en) | On-line automatic detection and sorting equipment and method for incubated egg quality | |
| CN103776841B (en) | Synthetic leather automatic defect detecting device and detection method | |
| CN102539443B (en) | Bottle body defect automatic detection method based on machine vision | |
| CN108465644B (en) | Artificial intelligent wheat quality inspection robot and quality inspection method | |
| CN112452801A (en) | High-efficiency online detection device for defects of glass bottles and working method thereof | |
| CN108465649A (en) | Artificial intelligence corn quality inspection robot and quality detecting method | |
| CN108469435A (en) | Artificial intelligence rice quality inspection robot and quality detecting method | |
| CN104634790A (en) | Capsule detection method and high-speed fully-automatic detection device | |
| CN113023293A (en) | Inspection method, device, equipment and system for belt conveyor | |
| CN108898179B (en) | Method and system for packaging ceramic tiles according to grades | |
| CN111545483B (en) | Chopstick sorting detection device and detection algorithm thereof | |
| CN109570051B (en) | Chinese chestnut wormhole detection device based on machine vision, laser and acoustics | |
| CN110125044A (en) | A kind of apple sorting equipment based on binocular vision | |
| US5973286A (en) | Grain sorting method and a device thereof | |
| CN202133655U (en) | Online automatic detecting sorting equipment for quality of incubated eggs | |
| CN109663747B (en) | Intelligent detection method for Chinese chestnut wormholes | |
| CN206966141U (en) | A kind of jujube separator of near infrared spectrum combination computer vision technique | |
| CN114653610A (en) | Fish identification and sorting implementation method | |
| CN202471610U (en) | Automatic bottle body defect detecting device based on machine vision | |
| US20220284699A1 (en) | System and method of object detection using ai deep learning models | |
| CN108875845B (en) | Medicine sorting device | |
| CN114318426B (en) | Exception handling method and system based on slot outlet information detection |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |