CN112001283B - Corn kernel breakage rate online detection device and detection method - Google Patents
Corn kernel breakage rate online detection device and detection method Download PDFInfo
- Publication number
- CN112001283B CN112001283B CN202010818460.0A CN202010818460A CN112001283B CN 112001283 B CN112001283 B CN 112001283B CN 202010818460 A CN202010818460 A CN 202010818460A CN 112001283 B CN112001283 B CN 112001283B
- Authority
- CN
- China
- Prior art keywords
- layer
- corn
- breakage rate
- corn kernel
- kernel
- 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
- 240000008042 Zea mays Species 0.000 title claims abstract description 138
- 235000002017 Zea mays subsp mays Nutrition 0.000 title claims abstract description 138
- 235000005824 Zea mays ssp. parviglumis Nutrition 0.000 title claims abstract description 136
- 235000005822 corn Nutrition 0.000 title claims abstract description 136
- 238000001514 detection method Methods 0.000 title claims abstract description 85
- 238000000034 method Methods 0.000 claims abstract description 21
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 13
- 238000012545 processing Methods 0.000 claims abstract description 10
- 239000010410 layer Substances 0.000 claims description 90
- 230000001360 synchronised effect Effects 0.000 claims description 34
- 239000002356 single layer Substances 0.000 claims description 33
- 238000011176 pooling Methods 0.000 claims description 18
- 238000012549 training Methods 0.000 claims description 17
- 230000006870 function Effects 0.000 claims description 9
- 230000008569 process Effects 0.000 claims description 6
- 230000001105 regulatory effect Effects 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 5
- 230000004913 activation Effects 0.000 claims description 3
- 238000007599 discharging Methods 0.000 claims description 3
- 210000002569 neuron Anatomy 0.000 claims description 3
- 238000012795 verification Methods 0.000 claims description 3
- 238000013528 artificial neural network Methods 0.000 claims description 2
- 238000003672 processing method Methods 0.000 abstract 1
- 235000013339 cereals Nutrition 0.000 description 11
- 238000010586 diagram Methods 0.000 description 9
- 238000005516 engineering process Methods 0.000 description 4
- 239000012535 impurity Substances 0.000 description 4
- 238000003306 harvesting Methods 0.000 description 3
- 238000005286 illumination Methods 0.000 description 3
- 235000016383 Zea mays subsp huehuetenangensis Nutrition 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000010191 image analysis Methods 0.000 description 2
- 235000009973 maize Nutrition 0.000 description 2
- 230000002035 prolonged effect Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000007873 sieving Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000000903 blocking effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000011217 control strategy Methods 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000011897 real-time detection Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
-
- A—HUMAN NECESSITIES
- A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
- A01D—HARVESTING; MOWING
- A01D41/00—Combines, i.e. harvesters or mowers combined with threshing devices
- A01D41/12—Details of combines
- A01D41/127—Control or measuring arrangements specially adapted for combines
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- 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
- G01N2021/8411—Application to online plant, process monitoring
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N2021/8466—Investigation of vegetal material, e.g. leaves, plants, fruits
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/06—Recognition of objects for industrial automation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Software Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Analytical Chemistry (AREA)
- Chemical & Material Sciences (AREA)
- Multimedia (AREA)
- Biochemistry (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Environmental Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention discloses an online detection device and method for corn kernel breakage rate. The device comprises a detection box, a feeding system, a layering device, a power system, an image acquisition system and an image processing system. The method comprises the steps of obtaining color images of complete corn and broken corn through a color industrial camera, obtaining single corn kernel images through an image processing method, and establishing a corn kernel breakage rate detection model through a convolutional neural network. The invention has better working reliability and longer continuous working time under a complex working environment, and the recognition accuracy of the established corn kernel breakage rate recognition model reaches more than 93 percent. Compared with manual detection and detection methods based on grain geometric features, the method can realize rapid and nondestructive detection of the corn grains, has high detection accuracy, can effectively identify the corn grains with different crushing degrees, and has good application prospect and market value.
Description
Technical Field
The invention relates to the technical field of detection, in particular to an online detection device and method for corn kernel breakage rate.
Background
Corn kernel breakage rate is one of important parameters for evaluating corn kernel harvesting quality, and a series of researches are carried out on detection and control of corn breakage rate by domestic and foreign scholars, scientific research institutions and enterprises to obtain parameters affecting the corn kernel breakage rate, wherein the parameters mainly comprise the travelling speed of a harvester, the rotating speed of a separation roller, the structure of threshing elements, concave plate gaps and the like. The existing corn combine harvester generally depends on the experience of a driver to control the speed of the corn combine harvester, so that the feeding quantity is kept in a reasonable interval, and the phenomenon of grain breakage caused by blockage is avoided when corn threshing.
The intelligent control of the harvester in China is in a starting stage, the defect of a real-time detection control strategy of the grain breakage rate severely restricts the intelligent and automatic development of the corn combine harvester in China, and the multi-parameter combined regulation and control and the corn low-loss threshing self-adaptive control are realized aiming at the low-loss threshing requirement. Therefore, a precise and efficient monitoring system for grain crushing of the combine harvester needs to be researched, the corn harvesting quality is accurately and reliably detected, and relevant working parameters of the combine harvester are changed in real time. How to accurately and rapidly collect the corn kernel breaking rate is a precondition for closed-loop control in the harvesting process.
Disclosure of Invention
Aiming at the problems of low detection precision, poor corn kernel monolayer effect and long detection period existing in the online detection process of corn kernel breakage rate, the invention aims to provide the online detection device and the online detection method for corn kernel breakage rate, which have the advantages of good corn kernel monolayer effect, high image acquisition quality, high detection speed and high detection accuracy.
On one hand, the invention provides an online detection device for corn kernel breakage rate, which comprises:
detection case, feeding system, single layering device, driving system, image acquisition system and image processing system, its characterized in that:
the detection box comprises a top cover 4, a box body 1 and a base 18, wherein an opening at the upper part of the top cover 4 is used for fixing the feeding system, an opening 2 of the base 18 is used for discharging detected corn kernels, and the box body 1 is fixed with the top cover 4 and the base 18 through bolts;
the feeding system comprises a collecting hopper 9, a blanking groove 13, a sieve plate 12 and a flow regulating plate 11; the collecting hopper 9 and the blanking groove 13 are fixed on the outer wall of the top cover 4 through bolts, the sieve plate 12 is arranged in the blanking groove 13 and is fixed on the inner wall of the blanking groove 13 through bolts and hinges, and the flow regulating plate 11 is fixed on the inner wall of the blanking groove 13 through bolts and guide rails;
the single-layer device comprises a baffle plate frame 10 and a single-layer baffle plate 5; the baffle frame 10 is fixed on the outer wall of the blanking groove 13 through a baffle frame connecting piece, and the single-layer baffle 5 is arranged on the baffle frame 10;
the power system comprises a driving motor 8, a driving shaft 14, a driven shaft 17, a tensioning wheel 16, a synchronous belt 3, a baffle 15, a driving shaft belt pulley 21 and a driven shaft belt pulley 22; the driving shaft 14, the driven shaft 17 and the tensioning wheel 16 are arranged on the side wall of the box body 1, the driving shaft 14 is connected with the driving motor 8 through a coupler, and is connected with the driving shaft belt wheel 21 through a key; the driven shaft 17 is connected with a driven shaft belt wheel 22 through a key; the synchronous belt 3 is arranged at the lower part of the blanking groove 13 and used for bearing corn kernels to be detected; the baffle 15 is arranged at the side of the synchronous belt 3 and is fixed on the side wall of the box body 1 through bolts;
the image acquisition system comprises a CCD color industrial camera 7, a strip-shaped LED light source and a camera bracket 6; the camera support 6 is fixed above the top cover 4 through bolts; the CCD color industrial camera 7 is connected with the camera bracket 6 through a bolt, and a strip-shaped LED light source is arranged on the side wall of the box body 1;
the image processing system comprises an industrial control computer and a vehicle-mounted terminal display device which are connected with each other; the industrial control computer is connected with the CCD color industrial camera (7) through a Gige network port.
Preferably, three single-layer baffles 5-1, 5-1 and 5-3 are arranged on the baffle frame 10, the gaps between the bottom ends of the single-layer baffles 5-1, 5-1 and 5-3 and the synchronous belt 3 are sequentially reduced, and the gap between the bottom end of the single-layer baffle 5-3 farthest from the blanking groove 13 and the synchronous belt 3 is the smallest and is the thickness of one corn grain.
Preferably, the baffle frame 10 is provided with a long groove 19 for adjusting the gap between the baffle frame 10 and the synchronous belt 3.
Preferably, the single-layer baffle plates (5-1, 5-3) are V-shaped, and openings 20 of 10cm x 20cm are arranged between the two ends of the single-layer baffle plates and the baffle plate frame 10 for the excessive corn kernels to pass through.
Preferably, the angle of the screen plate 12 and the position of the flow rate adjusting plate 11 on the blanking tank are adjusted according to the feeding amount of the corn kernels so as to adjust the quantity of the corn kernels entering the detecting box.
On the other hand, the invention provides an online detection method for the corn kernel breaking rate, which comprises the following steps:
step 1, acquiring a corn kernel image by using a CCD color industrial camera, acquiring a corn kernel data set, dividing the corn kernel data set into a training set and a verification set, and preprocessing;
step 2, constructing a corn kernel breakage rate detection model based on a convolutional neural network, inputting a training set sample into the detection model, and training by adopting a counter propagation method to obtain a trained corn kernel breakage rate detection model;
and step 3, inputting the preprocessed corn kernel image to be identified into a trained corn kernel breakage rate detection model, identifying broken corn kernels and calculating the corn kernel breakage rate.
Specifically, step 2 specifically includes the following steps:
step 2.1: inputting the preprocessed training data set into an eight-layer convolutional neural network model for back propagation;
step 2.2: and repeating the iteration process until the set iteration times are reached, and then training is terminated to obtain a trained corn kernel breakage rate detection model.
Preferably, the convolutional neural network takes a LeNet-5 network as a framework, and comprises 1 input layer, 3 convolutional layers, 2 pooling layers, 1 full-connection layer and 1 output layer, wherein pictures start from the input layer, pass through the 3 convolutional layers, the 2 pooling layers and the 1 full-connection layer, and an output result of the network is obtained through the output layer; wherein,
the first layer is an input layer: the pixel value of the input image is 32 x 32;
the second layer is a convolution layer C1: the convolution kernel size is 5*5, the step length is 1 pixel, and the output size is 28 x 28;
the third layer is a pooling layer S2: the convolution kernel size is 2 x 2, and the output size is 14 x 14;
the fourth layer is a convolution layer C3: the convolution kernel size is 5*5, the step length is 1 pixel, and the output size is 10 x 10;
the fifth layer is a pooling layer S4: the convolution kernel size is 2 x 2, and the output size is 5*5;
the sixth layer is a convolution layer C5: the convolution kernel size is 5*5, the step size is 1 pixel;
the seventh layer is a full connection layer F6: the layer has 84 neurons;
the eighth time is the output layer: the output layer uses a softmax function for a total of 2 outputs.
Preferably, the pooling layer adopts a maximum pooling mode, the convolution layer adopts a ReLU function as an activation function, and the initial learning rate of the convolution neural network is 0.01.
Preferably, the preprocessing is to segment out a single grain image by identifying the location of each grain in the image.
The invention has the beneficial effects that: the device has reasonable structural design and small volume, is convenient to install, and is internally provided with the sieve plate for sieving impurities in grains, so that the accuracy of a detection result is improved, the sieve plate can realize secondary drainage, the impact of corn grains on the synchronous belt is weakened, and the service life of parts is prolonged; the corn kernel three-time single-layering treatment can be realized by adopting the single-layering device, so that the corn kernels are effectively prevented from being stacked and adhered, and the requirement of image analysis on the single-layering of the kernels is met; a closed detection box device is adopted, a strip-shaped LED light source is arranged in the closed detection box device, and stable illumination is provided for image detection. The online detection method of the invention adopts a convolutional neural network technology to establish a corn kernel breakage rate detection model, can carry out online breakage rate analysis on corn kernels, and has the advantages of high detection precision, high recognition speed and good repeatability.
Drawings
FIG. 1 is a block diagram of an online detection device for corn kernel breakage rate according to the present invention;
FIG. 2 is a diagram showing the structure of an online detection device for corn kernel breakage rate according to the invention;
FIG. 3 is a block diagram of a feeding system and a monolayer device of the corn kernel breakage rate online detection device of the invention;
FIG. 4 is a diagram showing the structure of a transmission system of the online detection device for corn kernel breakage rate;
FIG. 5 is a diagram showing the relative positions of a color industrial camera and a synchronous belt in an on-line detection device for corn kernel breakage rate according to the present invention;
FIG. 6 is a flow chart of a method for online detection of corn kernel breakage rate according to the present invention.
Detailed Description
In order to make the above objects, advantages and technical solutions of the present invention more clear, the following detailed description of the embodiments of the present invention will be given with reference to the accompanying drawings.
Aiming at the problems of the corn kernel breakage rate detection device of the corn combine harvester in China, the invention provides a method for identifying broken corn kernels based on a convolutional neural network technology and an image processing technology and calculating the average breakage rate of the corn kernels.
Fig. 1 shows a block diagram of an on-line detection device for corn kernel breakage rate, which comprises a mechanical device and an image detection device. Wherein the mechanical device comprises: the device comprises a detection box, a feeding system, a layering device and a power system; the image detection device comprises an image acquisition device and an image processing device. The detection box provides a closed detection environment for detecting the breakage rate of the corn kernels in the image; the feeding system collects corn kernels and transmits the corn kernels to the synchronous belt; the corn kernels are tiled by the single layering device, so that stacking and adhesion phenomena of the corn kernels are prevented, the good kernel single layering treatment improves the subsequent image processing efficiency, the operation speed of an image analysis algorithm is high, and the kernel breakage rate detection period is shortened; the power system drives the synchronous belt to rotate and conveys corn kernels; the image acquisition device provides illumination with stable illumination intensity for the detection box, and the CCD color industrial camera acquires corn kernel images; the image processing device processes and analyzes the collected color images, calculates the corn kernel breakage rate and transmits the result to the vehicle-mounted terminal for display.
Fig. 2 is a block diagram of an on-line detection device for corn kernel breakage rate according to the present invention. The on-line detection device of the present invention includes:
the detection box comprises a box body 1, a top cover 4 and a base 18; the upper opening of the top cover 4 is connected with a feeding system, an opening 2 is arranged on the base 18 and used for discharging and detecting corn kernels, and the box 1 is fixed with the top cover 4 and the base 18 through bolts.
As shown in fig. 3, which is a block diagram of the feeding system and the singulation apparatus in the in-line inspection apparatus shown in fig. 2.
The feeding system comprises a collecting hopper 9, a blanking groove 13, a sieve plate 12 and a flow regulating plate 11; the collecting hopper 9 and the blanking groove 13 are fixed on the outer wall of the top cover 4 through bolts, the sieve plate 12 is arranged inside the blanking groove 13, and the sieve plate is fixed on the inner wall of the blanking groove 13 through bolts and hinges. The angle of the screen deck 12 is adjusted by adjusting the position of the screen deck 12 on the screws 23. The corn kernels passing through the blanking groove 13 drop to the synchronous belt 3. The blanking groove 13 is of a closed structure all around, so that corn kernels are prevented from falling to the outside of the synchronous belt 3, the sieve plate 12 can buffer the impact of the corn kernels on the synchronous belt 3, and the service life of parts is prolonged. The flow regulating plate 11 is fixed on the inner wall of the blanking groove 13 through bolts and guide rails; the size of the feeding hole is controlled by adjusting the position of the flow adjusting plate 11 in the blanking groove 13, so that corn kernels entering the detection box under the condition of different kernel feeding amounts are ensured to be in a reasonable range.
The single-layer device comprises a baffle frame 10, a single-layer baffle 5 and baffle frame connectors; the baffle frame 10 is fixed on the outer wall of the blanking groove 13 through bolts, and a long groove 19 is formed in the baffle frame 10 and can be used for adjusting the gap between the baffle frame 10 and the synchronous belt 3; the monolayer baffle 5 is arranged on the baffle frame 10 and is used for monolayer arrangement of the corn kernels on the synchronous belt 3. The baffle frame 10 is provided with three single-layer baffles 5-1, 5-2 and 5-3, the gaps between the bottom ends of the three single-layer baffles and the synchronous belt 3 are sequentially reduced, namely, the farther the baffle frame is away from the blanking groove 13, the smaller the gap between the single-layer baffles 5 and the synchronous belt 3 is, namely, the distance between the bottom of the single-layer baffles 5-3 and the synchronous belt 3 is the smallest and is the height of a corn kernel, the corn kernel can be tiled on the synchronous belt 3 through three single-layer corn kernel treatment, and the corn kernel is prevented from stacking and adhering. The monolayer baffle 5 is V-shaped, and a gap 20 of 10cm x 20cm is arranged between the baffle frame 10 for the passage of redundant corn seeds, so that the phenomenon of blocking the corn seeds in the process of monolayer corn seeds is avoided.
Fig. 4 is a diagram showing a construction of a power system in the in-line detecting apparatus of fig. 2. The power system comprises a driving motor 8, a driving shaft 14, a driven shaft 17, a tensioning wheel 16, a synchronous belt 3 and a baffle 15; the driving shaft 14 is fixed on the side wall of the box body 1, is connected with the driving motor 8 through a coupling, and is connected with the driving shaft belt wheel 21 through a key. The driven shaft 17 is connected with the driven shaft belt pulley 22 through a key, the driven shaft 17 and the tensioning wheel 16 are both fixed on the side wall of the box body, the driving shaft 14 and the driven shaft 17 are connected through the synchronous belt 3, and the tensioning wheel 16 is arranged below the synchronous belt 3 and used for tensioning the synchronous belt 3. The hold-in range 3 sets up in blanking groove 13 lower part and is used for bearing the corn kernel that waits to detect. Baffle 15 sets up in hold-in range 3 side, fixes on box 1 lateral wall through the bolt for prevent that the maize seed grain from dropping to hold-in range 3 inside. A strip-shaped LED light source (not shown) is provided on the side wall of the case 1.
The image acquisition system comprises a CCD color industrial camera 7, a strip-shaped LED light source and a camera bracket 6; the CCD color industrial camera 7 is connected with the camera bracket 6 through a bolt, the camera bracket 6 is fixed on the top cover 4 through a bolt, and the strip-shaped LED light source is arranged on the side wall of the box body 1; fig. 5 shows the relative positions of a color industrial camera and a timing belt.
An image processing system including an industrial control computer; the industrial control computer is respectively connected with the CCD color industrial camera. The CCD color industrial camera 7 is connected with the camera bracket through a tightening bolt and is connected with an industrial control computer through a Gige network port. Preferably, the industrial control computer employs an Intel celron J1900 processor.
Based on the above, the invention provides a method for realizing the online detection of the corn kernel breakage rate by using the online detection device, as shown in fig. 6.
Step 1, acquiring corn kernel images, including complete kernel images and broken kernel images, by using a color industrial camera to obtain a corn kernel data set. The corn kernel dataset is divided into a training set and a validation set, and then the corn kernel dataset is preprocessed.
The collecting hopper 9 collects part of the seeds thrown out by the auger, the sieve plate 12 screens impurities in the seeds and discharges the impurities out of the detection box, the blanking groove 13 is arranged right above the synchronous belt 3, and the corn seeds fall above the synchronous belt 3 through the sieve plate 12. The driving shaft 14 is connected with the driving motor 8 through a coupler, drives the synchronous belt 3 to transport corn kernels to be detected to pass through the single-layer device, and sequentially passes through the three single-layer baffles 5 to tile the corn kernels to be detected on the synchronous belt. The CCD color industrial camera 7 collects corn kernel images tiled on the synchronous belt, and transmits the images to an industrial control computer through a Gige network port. The industrial control computer preprocesses the color image and separates out a single corn kernel image.
Corn kernel images acquired by color industrial cameras are 1600 x 1200 in size, and preprocessing is performed by identifying the position of each kernel in the image and separating out a single kernel image, for example, extracting a single kernel image with a size of 150 x 150 from the original image, and then setting the image to 32 x 32.
And 2, constructing a corn kernel breakage rate detection model, inputting a training set sample into the detection model, and training by adopting a counter propagation method.
Preferably, the preprocessed training data set is input into an eight-layer convolutional neural network model for back propagation; the picture starts from the input layer, passes through 3 convolution layers, 2 pooling layers and 1 full connection layer, and obtains the output result of the network through the output layer.
Preferably, the convolutional neural network is framed by a LeNet-5 network and comprises 1 input layer, 3 convolutional layers, 2 pooling layers, 1 full connection layer and 1 output layer; wherein,
the first layer is an input layer: the pixel value of the input image is 32 x 32;
the second layer is a convolution layer C1: the convolution kernel size is 5*5, the step length is 1 pixel, and the output size is 28 x 28;
the third layer is a pooling layer S2: the convolution kernel size is 2 x 2, and the output size is 14 x 14;
the fourth layer is a convolution layer C3: the convolution kernel size is 5*5, the step length is 1 pixel, and the output size is 10 x 10;
the fifth layer is a pooling layer S4: the convolution kernel size is 2 x 2, and the output size is 5*5;
the sixth layer is a convolution layer C5: the convolution kernel size is 5*5, the step size is 1 pixel;
the seventh layer is a full connection layer F6: the layer has 84 neurons;
the eighth time is the output layer: the output layer uses a softmax function for a total of 2 outputs.
The model adopts a moment-carrying random gradient descent method solver, a pooling layer adopts a maximum pooling mode, and a convolution layer adopts a ReLU function as an activation function. The initial learning rate of the convolutional neural network is 0.01.
The images in the training set are randomly ordered according to 7: and 3, dividing the training set and the verification set. And repeating the iteration process until the set iteration times, such as 50 times, are reached, and training is terminated, so that a trained model is obtained.
Step 3, corn kernel breaking rate detection: inputting the pretreated corn kernel image to be identified into a corn kernel breakage rate detection model, identifying broken corn kernels and calculating the corn kernel breakage rate.
The corn kernel online detection device provided by the invention has the advantages of reasonable structural design, small volume and convenience in vehicle-mounted installation. The inside sieve that is equipped with of blanking groove is arranged in sieving impurity in the seed grain, improves testing result accuracy, and the sieve can realize secondary drainage, weakens the impact of maize seed grain to the hold-in range, improves spare part's life. And corn kernels are subjected to three-time monolayer treatment, so that corn kernels are effectively prevented from being stacked and adhered. The detection method of the device adopts the convolutional neural network technology to analyze the breakage rate of corn kernels, and has the advantages of high detection speed and high accuracy.
The present invention is not limited to the preferred embodiments, and any changes or substitutions that would be apparent to one skilled in the art within the scope of the present invention are intended to be included in the scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.
Claims (11)
1. Corn kernel breakage rate on-line measuring device, characterized by, include: detection case, feeding system, single layering device, driving system, image acquisition system and image processing system, its characterized in that:
the detection box comprises a top cover (4), a box body (1) and a base (18), wherein an opening at the upper part of the top cover (4) is used for fixing the feeding system, an opening (2) of the base (18) is used for discharging detected corn kernels, and the box body (1) is fixed with the top cover (4) and the base (18) through bolts;
the feeding system comprises a collecting hopper (9), a blanking groove (13), a sieve plate (12) and a flow regulating plate (11); the collecting hopper (9) and the blanking groove (13) are fixed on the outer wall of the top cover (4) through bolts, the sieve plate (12) is arranged inside the blanking groove (13) and is fixed on the inner wall of the blanking groove (13) through bolts and hinges, and the flow regulating plate (11) is fixed on the inner wall of the blanking groove (13) through bolts and guide rails;
the single-layer device comprises a baffle plate frame (10) and a single-layer baffle plate (5); the baffle frame (10) is fixed on the outer wall of the blanking groove (13) through a baffle frame connecting piece, and the single-layer baffle (5) is arranged on the baffle frame (10);
the power system comprises a driving motor (8), a driving shaft (14), a driven shaft (17), a tensioning wheel (16), a synchronous belt (3), a baffle plate (15), a driving shaft belt wheel (21) and a driven shaft belt wheel (22); the driving shaft (14), the driven shaft (17) and the tensioning wheel (16) are arranged on the side wall of the box body (1), the driving shaft (14) is connected with the driving motor (8) through a coupler, and is connected with the driving shaft belt wheel (21) through a key; the driven shaft (17) is connected with the driven shaft belt wheel (22) through a key; the synchronous belt (3) is arranged at the lower part of the blanking groove (13) and used for bearing corn kernels to be detected; the baffle (15) is arranged at the side of the synchronous belt (3) and is fixed on the side wall of the box body (1) through bolts;
the image acquisition system comprises a CCD color industrial camera (7), a strip-shaped LED light source and a camera bracket (6); the camera support (6) is fixed above the top cover (4) through bolts; the CCD color industrial camera (7) is connected with the camera bracket (6) through bolts, and the strip-shaped LED light source is arranged on the side wall of the box body (1);
the image processing system comprises an industrial control computer and a vehicle-mounted terminal display device which are connected with each other; the industrial control computer is connected with the CCD color industrial camera (7) through a Gige network port.
2. The online detection device for corn kernel breakage rate according to claim 1, wherein three single-layer baffles (5-1, 5-1 and 5-3) are arranged on the baffle frame (10), gaps between the bottom ends of the single-layer baffles (5-1, 5-1 and 5-3) and the synchronous belt (3) are sequentially reduced, and gaps between the bottom ends of the single-layer baffles (5-3) farthest from the blanking groove (13) and the synchronous belt (3) are the smallest and are the thickness of one corn kernel.
3. An on-line detection device for corn kernel breakage rate according to claim 1, characterized in that the baffle frame (10) is provided with a long groove (19) for adjusting the gap between the baffle frame (10) and the synchronous belt (3).
4. The online detection device for corn kernel breakage rate according to claim 1, wherein the single-layer baffle plates (5-1, 5-3) are in a V shape, and openings (20) of 10cm x 20cm are arranged between two ends of each single-layer baffle plate and the baffle plate frame (10) for passing redundant corn kernels.
5. An on-line detection apparatus for corn kernel breakage rate according to claim 1, wherein the number of corn kernels entering the detection box is adjusted by adjusting the angle of the screen plate (12) and the position of the flow rate adjusting plate (11) on the blanking tank according to the feeding amount of corn kernels.
6. An on-line corn kernel breakage rate detection device as claimed in claim 5, characterized in that the angle of the screen plate (12) is adjusted by adjusting the position of the screen plate (12) on the screw (23).
7. A method for detecting the breakage rate of corn kernels by using the online detection device for breakage rate of corn kernels according to any one of claims 1 to 6, comprising the steps of:
step 1, acquiring a corn kernel image by using a CCD color industrial camera, acquiring a corn kernel data set, dividing the corn kernel data set into a training set and a verification set, and preprocessing;
step 2, constructing a corn kernel breakage rate detection model based on a convolutional neural network, inputting a training set sample into the detection model, and training by adopting a counter propagation method to obtain a trained corn kernel breakage rate detection model;
and step 3, inputting the preprocessed corn kernel image to be identified into a trained corn kernel breakage rate detection model, identifying broken corn kernels and calculating the corn kernel breakage rate.
8. A method for detecting the breakage rate of corn kernels by using the online detection device for breakage rate of corn kernels according to any one of claims 1 to 6, wherein the step 2 comprises the following steps:
step 2.1: inputting the preprocessed training data set into an eight-layer convolutional neural network model for back propagation;
step 2.2: and repeating the iteration process until the set iteration times are reached, and then training is terminated to obtain a trained corn kernel breakage rate detection model.
9. The method for detecting the corn kernel breakage rate by using the online corn kernel breakage rate detection device according to any one of claims 1 to 6, wherein the convolutional neural network takes a LeNet-5 network as a framework and comprises 1 input layer, 3 convolutional layers, 2 pooling layers, 1 full-connection layer and 1 output layer, and a picture starts from the input layer, passes through the 3 convolutional layers, the 2 pooling layers and the 1 full-connection layer, and obtains an output result of the network through the output layer; wherein,
the first layer is an input layer: the pixel value of the input image is 32 x 32;
the second layer is a convolution layer C1: the convolution kernel size is 5*5, the step length is 1 pixel, and the output size is 28 x 28;
the third layer is a pooling layer S2: the convolution kernel size is 2 x 2, and the output size is 14 x 14;
the fourth layer is a convolution layer C3: the convolution kernel size is 5*5, the step length is 1 pixel, and the output size is 10 x 10;
the fifth layer is a pooling layer S4: the convolution kernel size is 2 x 2, and the output size is 5*5;
the sixth layer is a convolution layer C5: the convolution kernel size is 5*5, the step size is 1 pixel;
the seventh layer is a full connection layer F6: the layer has 84 neurons;
the eighth time is the output layer: the output layer uses a softmax function for a total of 2 outputs.
10. A method for detecting corn kernel breakage rate by using the online corn kernel breakage rate detection device according to any one of claims 1-6, wherein the pooling layer adopts a maximum pooling mode, the convolution layer adopts a ReLU function as an activation function, and the initial learning rate of the convolution neural network is 0.01.
11. A method of corn kernel breakage rate detection using a corn kernel breakage rate online detection apparatus according to any one of claims 1 to 6, wherein the preprocessing is to divide a single kernel image by recognizing the position of each kernel in the image.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010818460.0A CN112001283B (en) | 2020-08-14 | 2020-08-14 | Corn kernel breakage rate online detection device and detection method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010818460.0A CN112001283B (en) | 2020-08-14 | 2020-08-14 | Corn kernel breakage rate online detection device and detection method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112001283A CN112001283A (en) | 2020-11-27 |
CN112001283B true CN112001283B (en) | 2024-01-26 |
Family
ID=73472933
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010818460.0A Active CN112001283B (en) | 2020-08-14 | 2020-08-14 | Corn kernel breakage rate online detection device and detection method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112001283B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113581798A (en) * | 2021-07-07 | 2021-11-02 | 中国农业科学院深圳农业基因组研究所 | Seed flattening device |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2013076644A1 (en) * | 2011-11-22 | 2013-05-30 | San Lucio S.r.l. | Food based on popcorn and cheese, and respective production method |
CN109085170A (en) * | 2017-06-14 | 2018-12-25 | 中国科学院沈阳自动化研究所 | Synchronous belt type corn kernel percentage of damage on-line measuring device and method |
CN110501222A (en) * | 2019-08-26 | 2019-11-26 | 中国农业大学 | A kind of cereal bulk solid anti-pressure ability detection method |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2016165082A1 (en) * | 2015-04-15 | 2016-10-20 | 中国科学院自动化研究所 | Image stego-detection method based on deep learning |
-
2020
- 2020-08-14 CN CN202010818460.0A patent/CN112001283B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2013076644A1 (en) * | 2011-11-22 | 2013-05-30 | San Lucio S.r.l. | Food based on popcorn and cheese, and respective production method |
CN109085170A (en) * | 2017-06-14 | 2018-12-25 | 中国科学院沈阳自动化研究所 | Synchronous belt type corn kernel percentage of damage on-line measuring device and method |
CN110501222A (en) * | 2019-08-26 | 2019-11-26 | 中国农业大学 | A kind of cereal bulk solid anti-pressure ability detection method |
Also Published As
Publication number | Publication date |
---|---|
CN112001283A (en) | 2020-11-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN113820325B (en) | Corn grain direct-harvest impurity-containing rate and breakage rate online detection system and method | |
US5917927A (en) | Grain inspection and analysis apparatus and method | |
CN101905215B (en) | Digitalized paddy rice seed testing machine | |
CN100506401C (en) | Pearl real time detection and classifying system based on mechanical vision | |
CN102284428A (en) | Corn real-time detection and grading machine vision system | |
CN105009731B (en) | Corn seed investigating method and its system | |
CN1804620A (en) | Method and apparatus for detecting surface quality of egg | |
CN104646315A (en) | Intelligent agricultural product sorting machine with aflatoxin detection function | |
CN101486033A (en) | Device for sorting fruits and method for sorting Chinese chestnut | |
CN102253052A (en) | Grain quality on-line detection apparatus based on field programmable gate array (FPGA), and method thereof | |
CN201041551Y (en) | A real time detection and classification system for pearl based on machine vision | |
CN108719424A (en) | A kind of aquatic products sorting technique and system based on machine vision | |
CN202591136U (en) | Automatic walnut sorting machine based on relative specific weight | |
CN108655021B (en) | Online damaged discernment sorter of maize seed based on machine vision | |
CN112001283B (en) | Corn kernel breakage rate online detection device and detection method | |
CN108120715B (en) | Corn kernel breaking rate online detection method and device based on machine vision | |
CN100365411C (en) | Detection method suited on surface of spherical fruit triggered to collect images based on need and equipment | |
CN206876589U (en) | Synchronous belt type corn kernel percentage of damage on-line measuring device | |
CN202114018U (en) | Machine vision system for grain real-time detection and grading | |
CN202083645U (en) | FPGA (field programmable gate array)-based food quality on-line detector | |
Liu et al. | Impurity monitoring study for corn kernel harvesting based on machine vision and CPU-Net | |
CN108204977A (en) | Corn kernel quality automatic detection device based on machine vision | |
CN205538564U (en) | Broken on -line monitoring system of cereal in combine grain tank | |
CN211043154U (en) | Machine vision-based device for detecting and evaluating internal and external quality of post-harvest cabbages | |
CN214794500U (en) | Online detection device for wheat appearance quality |
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 |