CN115393650A - Beer yeast powder intelligent production method and system - Google Patents
Beer yeast powder intelligent production method and system Download PDFInfo
- Publication number
- CN115393650A CN115393650A CN202211115228.6A CN202211115228A CN115393650A CN 115393650 A CN115393650 A CN 115393650A CN 202211115228 A CN202211115228 A CN 202211115228A CN 115393650 A CN115393650 A CN 115393650A
- Authority
- CN
- China
- Prior art keywords
- feature map
- scale
- neural network
- yeast powder
- differential
- 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.)
- Withdrawn
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B02—CRUSHING, PULVERISING, OR DISINTEGRATING; PREPARATORY TREATMENT OF GRAIN FOR MILLING
- B02C—CRUSHING, PULVERISING, OR DISINTEGRATING IN GENERAL; MILLING GRAIN
- B02C13/00—Disintegrating by mills having rotary beater elements ; Hammer mills
- B02C13/02—Disintegrating by mills having rotary beater elements ; Hammer mills with horizontal rotor shaft
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B02—CRUSHING, PULVERISING, OR DISINTEGRATING; PREPARATORY TREATMENT OF GRAIN FOR MILLING
- B02C—CRUSHING, PULVERISING, OR DISINTEGRATING IN GENERAL; MILLING GRAIN
- B02C23/00—Auxiliary methods or auxiliary devices or accessories specially adapted for crushing or disintegrating not provided for in preceding groups or not specially adapted to apparatus covered by a single preceding group
- B02C23/08—Separating or sorting of material, associated with crushing or disintegrating
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/24—Aligning, centring, orientation detection or correction of the image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Databases & Information Systems (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Food Science & Technology (AREA)
- Image Analysis (AREA)
Abstract
The method comprises the steps of enabling a crushed image of yeast raw materials crushed by a hammer mill and collected by a camera to pass through a first convolution neural network with a cavity convolution kernel with a first voidage and a second convolution neural network with a cavity convolution kernel with a second voidage respectively to obtain a first scale characteristic diagram and a second scale characteristic diagram, correcting characteristic values of all positions in a difference characteristic diagram between the first scale characteristic diagram and the second scale characteristic diagram to obtain a corrected difference characteristic diagram, and enabling the corrected difference characteristic diagram to pass through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the crushing degree of the yeast raw materials meets a preset requirement or not. Through such mode, carry out accurate aassessment to beer yeast powder's crushing state to optimize the crushing packing link in the beer yeast powder production process.
Description
Technical Field
The present application relates to the field of intelligent processing, and more particularly, to an intelligent production method and system for beer yeast powder.
Technical Field
The beer yeast powder contains rich vitamin B group, multiple vitamins and mineral substances, more than 50 percent of protein and complete amino acid group, and is the best source for supplementing high-quality protein. The beer yeast powder also contains rich dietary fiber, and is helpful for relieving constipation.
The preparation process flow of the beer yeast powder comprises the steps of stirring raw materials, drying in a roller, drying in a drying room, crushing and packaging, dedusting by a crusher, inspecting and warehousing. In recent years, a plurality of manufacturers optimize the preparation process flow of the beer yeast powder so as to improve the preparation efficiency and the preparation effect of the beer yeast powder. However, in the preparation process flow of the beer yeast powder, the optimization of the crushing and packaging links is not sufficient all the time.
Specifically, in the crushing and packaging link, materials enter a hammer type crusher through a primary cleaning screen, a lifting machine and a spiral feeder to be crushed, the materials are conveyed to a cyclone separator through a centrifugal fan, powdery yeast is placed into a finished product packaging bag through a continuous discharging device at the lower section, and a sample is sampled from a finished product for testing. In this link, if the pulverizing effect of the hammer mill is insufficient, the test of the sampled sample in the finished product may be failed, and if the pulverizing is too sufficient, more part of the beer yeast powder may be discharged in the form of dust. That is, it is important to evaluate the pulverization state of the beer yeast powder in the pulverization packaging step.
Therefore, an optimized beer yeast powder intelligent production scheme is expected, which can accurately evaluate the crushing state of the beer yeast powder so as to optimize the production efficiency and the quality of the finished product of the beer yeast powder.
Disclosure of Invention
The present application is proposed to solve the above-mentioned technical problems. The embodiment of the application provides an intelligent beer yeast powder production method and system, wherein a crushed image of a yeast material crushed by a hammer mill and acquired by a camera is respectively passed through a first convolution neural network with a cavity convolution kernel with a first voidage and a second convolution neural network with a cavity convolution kernel with a second voidage to obtain a first scale characteristic diagram and a second scale characteristic diagram, characteristic values of all positions in a difference characteristic diagram between the first scale characteristic diagram and the second scale characteristic diagram are corrected to obtain a corrected difference characteristic diagram, and then a classification result can be obtained through a classifier, wherein the classification result is used for indicating whether the crushing degree of the yeast material meets a preset requirement. Through such a mode, carry out accurate aassessment to the crushing state of beer yeast powder to optimize the crushing packing link in the beer yeast powder production process, and promote the production efficiency and the finished product quality of beer yeast powder.
Accordingly, according to an aspect of the present application, there is provided an intelligent production method of beer yeast powder, comprising:
acquiring a crushed image of the yeast material crushed by the hammer crusher and collected by a camera;
obtaining a first scale feature map by using a first convolution neural network of a hole convolution kernel with a first hole rate;
passing the pulverized image through a second convolutional neural network using a hole convolutional kernel with a second void rate to obtain a second scale feature map, wherein the first void rate is different from the second void rate;
calculating a difference feature map between the first scale feature map and the second scale feature map;
correcting the feature values of all positions in the differential feature map based on the scale of the differential feature map to obtain a corrected differential feature map, wherein the scale of the differential feature map is obtained by multiplying the width of the differential feature map by the height and then multiplying by the number of channels; and
and passing the corrected differential characteristic diagram through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the crushing degree of the yeast material meets the preset requirement or not.
In the above beer yeast powder intelligent production method, the obtaining a first scale feature map of the pulverized image by using a first convolution neural network of a cavity convolution kernel having a first cavity rate includes: performing convolution processing, pooling processing and nonlinear activation processing based on the hole convolution kernel with the first hole rate on input data respectively in forward pass of layers by using each layer of the first convolution neural network to output the first scale feature map from the last layer of the first convolution neural network, wherein the input of the first layer of the first convolution neural network is the crushed image.
In the above beer yeast powder intelligent production method, the passing the pulverized image through a second convolutional neural network using a cavity convolutional kernel with a second voidage to obtain a second scale feature map, where the first voidage is different from the second voidage, includes: and performing convolution processing, pooling processing and nonlinear activation processing based on the hole convolution kernel with the second hole rate on input data respectively in forward transmission of layers by using the layers of the second convolutional neural network to output the second scale feature map from the last layer of the second convolutional neural network, wherein the input of the first layer of the second convolutional neural network is the crushed image.
In the above method for intelligently producing beer yeast powder, the calculating a difference feature map between the first scale feature map and the second scale feature map includes: calculating the difference by position between the first scale feature map and the second scale feature map according to the following formula to obtain the difference feature map; wherein the formula is:
F 1 is a first scale feature map, F 2 Is a second scale feature map, F d Is a differential characteristic diagram.
In the above method for intelligently producing beer yeast powder, the correcting the feature value at each position in the difference feature map based on the scale of the difference feature map to obtain a corrected difference feature map includes: based on the scale of the differential feature map, correcting the feature value of each position in the differential feature map by the following formula to obtain the corrected differential feature map; wherein the formula is
f i Is the feature value of each location in the difference feature map, μ and σ are the mean and variance of the feature values of all locations in the difference feature map, and N is the scale of the difference feature map.
In the above method for intelligently producing beer yeast powder, the step of passing the corrected difference feature map through a classifier to obtain a classification result includes: processing the differential feature map by using the classifier according to the following formula to obtain the classification result, wherein the formula is as follows:
O＝softmax{(W n ,B n ):…:(W 1 ,B 1 )|Project(F)}
wherein Project (F) represents projecting the corrected difference feature map as a vector, W 1 To W n As a weight matrix for each fully connected layer, B 1 To B n A bias matrix representing the layers of the fully connected layer.
In the above beer yeast powder intelligent production method, the beer yeast powder intelligent production method further includes: and generating a centrifugal fan starting instruction in response to the classification result that the crushing degree of the yeast material meets the preset requirement, wherein the centrifugal fan starting instruction is used for starting the centrifugal fan.
According to another aspect of the present application, there is also provided an intelligent production system of beer yeast powder, comprising:
the image acquisition unit is used for acquiring a crushed image of the yeast material crushed by the hammer crusher and acquired by the camera;
the first convolution neural network unit is used for obtaining a first scale characteristic diagram of the crushed image through a first convolution neural network with a hole convolution kernel with a first hole rate;
a second convolutional neural network unit for passing the pulverized image through a second convolutional neural network using a hole convolutional kernel having a second void rate to obtain a second scale feature map, wherein the first void rate is different from the second void rate;
a difference feature map generation unit which calculates a difference feature map between the first scale feature map and the second scale feature map;
the corrected differential feature map generation unit is used for correcting the feature values of all positions in the differential feature map based on the scale of the differential feature map to obtain a corrected differential feature map, wherein the scale of the differential feature map is obtained by multiplying the width of the differential feature map by the height and then multiplying the width by the number of channels; and
and the evaluation result generating unit is used for enabling the corrected differential characteristic diagram to pass through a classifier to obtain a classification result, and the classification result is used for indicating whether the crushing degree of the yeast material meets the preset requirement or not.
In the above beer yeast powder intelligent production system, the first convolutional neural network unit is further configured to: performing convolution processing, pooling processing and nonlinear activation processing based on the hole convolution kernel with the first hole rate on input data respectively in forward pass of layers by using each layer of the first convolution neural network to output the first scale feature map from the last layer of the first convolution neural network, wherein the input of the first layer of the first convolution neural network is the crushed image.
In the above beer yeast powder intelligent production system, the second convolutional neural network unit is further configured to: performing convolution processing, pooling processing and nonlinear activation processing based on the hole convolution kernel with the second hole rate on input data in forward pass of layers respectively by using layers of the second convolutional neural network to output the second scale feature map by a last layer of the second convolutional neural network, wherein the input of the first layer of the second convolutional neural network is the crushed image.
In the above beer yeast powder intelligent production system, the difference characteristic map generating unit is further configured to: calculating the difference by position between the first scale feature map and the second scale feature map according to the following formula to obtain the difference feature map; wherein the formula is:
F 1 is a first scale feature map, F 2 Is a second scale feature map, F d Is a differential characteristic diagram.
In the above beer yeast powder intelligent production system, the corrected difference characteristic map generating unit is further configured to: based on the scale of the differential feature map, correcting the feature value of each position in the differential feature map by the following formula to obtain the corrected differential feature map; wherein the formula is
f i Is the feature value of each location in the difference feature map, μ and σ are the mean and variance of the feature values of all locations in the difference feature map, and N is the scale of the difference feature map.
In the above beer yeast powder intelligent production system, the evaluation result generating unit is further configured to: processing the differential feature map by using the classifier according to the following formula to obtain the classification result, wherein the formula is as follows:
O＝softmax{(W n ,B n ):…:(W 1 ,B 1 )|Project(F)}
wherein Project (F) represents projecting the corrected difference feature map as a vector, W 1 To W n As a weight matrix for each fully connected layer, B 1 To B n A bias matrix representing the layers of the fully connected layer.
In above-mentioned beer yeast powder intelligent production system, beer yeast powder intelligent production system still further is used for: and generating a centrifugal fan starting instruction in response to the classification result that the crushing degree of the yeast material meets the preset requirement, wherein the centrifugal fan starting instruction is used for starting the centrifugal fan.
According to still another aspect of the present application, there is provided an electronic apparatus including: a processor; and a memory in which computer program instructions are stored, which, when executed by the processor, cause the processor to perform the beer yeast powder intelligent production method as described above.
According to yet another aspect of the present application, there is provided a computer readable medium having stored thereon computer program instructions, which, when executed by a processor, cause the processor to perform the beer yeast powder intelligent production method as described above.
Compared with the prior art, the beer yeast powder intelligent production method and the system thereof provided by the application have the advantages that the crushed images of the yeast material raw material crushed by the hammer mill and acquired by the camera respectively pass through the first convolution neural network with the cavity convolution kernel of the first voidage and the second convolution neural network with the cavity convolution kernel of the second voidage to obtain the first scale characteristic diagram and the second scale characteristic diagram, the characteristic values of all positions in the difference characteristic diagram between the first scale characteristic diagram and the second scale characteristic diagram are corrected to obtain the corrected difference characteristic diagram, then the corrected difference characteristic diagram is subjected to a classifier to obtain a classification result, and the classification result is used for indicating whether the crushing degree of the yeast material raw material meets the preset requirement or not. Through such mode, carry out accurate aassessment to beer yeast powder's crushing state to optimize the crushing packing link in the beer yeast powder production process, and promote beer yeast powder's production efficiency and finished product quality.
Drawings
The above and other objects, features and advantages of the present application will become more apparent by describing in more detail embodiments of the present application with reference to the attached drawings. The accompanying drawings are included to provide a further understanding of the embodiments of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the principles of the application. In the drawings, like reference numbers generally represent like parts or steps.
FIG. 1 illustrates a flow chart of a beer yeast production process according to an embodiment of the present application.
Fig. 2 illustrates a scene schematic diagram of an intelligent production method of beer yeast powder according to an embodiment of the present application.
Fig. 3 illustrates a flowchart of an intelligent production method of beer yeast powder according to an embodiment of the present application.
Fig. 4 illustrates a schematic architecture diagram of an intelligent production method of beer yeast powder according to an embodiment of the present application.
Fig. 5 illustrates a block diagram of an intelligent production system of beer yeast powder according to an embodiment of the present application.
FIG. 6 illustrates a block diagram of an electronic device in accordance with an embodiment of the present application.
Detailed Description
Hereinafter, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be understood that the described embodiments are only some embodiments of the present application and not all embodiments of the present application, and that the present application is not limited by the example embodiments described herein.
Summary of the application
FIG. 1 illustrates a flow chart of a beer yeast powder production process according to an embodiment of the present application. As shown in figure 1, the production process of the beer yeast powder comprises the following steps: stirring raw materials, adding beer yeast, yeast filter residue and yeast-containing feed liquid collected in the equipment cleaning process into a yeast stirring pool, and uniformly stirring; drying in a roller, opening a roller steam valve and a rotating motor, preheating the roller for about 30 minutes, pumping yeast liquid, observing the drying degree by an operator, properly adjusting the rotating speed and the steam consumption of the roller, naturally cooling the flaky powdery yeast obtained by drying in the roller, filling the cooled flaky powdery yeast into a woven bag, collecting a certain amount of flaky powdery yeast, and then drying, crushing, inspecting and packaging the flaky powdery yeast; drying in a drying room, wherein the flaky powdery yeast obtained by roller drying is unstable in water due to multiple collection, and needs to be dried in the drying room at 100-180 ℃ after reaching a certain amount, and the qualified yeast subjected to sampling and testing enters the next step of crushing; crushing and packaging, feeding materials into a hammer type crusher through a primary cleaning sieve, a lifter and a spiral feeder to be crushed, conveying the crushed materials to a cyclone separator through a centrifugal fan, putting powdered yeast into a finished product packaging bag through a continuous discharger at the lower end, extracting a sample from a finished product to be tested, weighing and measuring the net weight of the packaging to 25 kg/bag, covering a batch number (production date) and the like on a label, tying a plastic bag in the packaging, tying a composite woven bag outside the packaging bag by using a packaging line, sealing the packaging bag by using a bag sewing machine, and moving the packaging bag to a finished product warehouse to be detected by an operator in time; the upper port of the cyclone separator is connected with the pulse dust collector, and compressed air is adopted in the pulse dust collector to blow regularly so as to recover yeast dust in the cloth bag and reduce emission; inspecting, wherein the product is sampled according to sampling regulations by a quality inspector in the production process and after the product is finished, and an inspector inspects the product according to inspection operation rules after receiving the sample, judges the product according to product standards and issues an inspection report; and warehousing, after the product is qualified, moving the product into a qualified product area or removing the mark to be detected, and handling warehousing procedures. In addition, because the yeast comes from breweries, the normal-temperature storage time of the wet beer yeast is short, and the long-distance transportation and warehouse storage are not facilitated, so that brewers for producing beer yeast powder can determine the distance according to the distance condition when purchasing raw materials, and purchase wet raw materials in nearby areas for roller drying, crushing and packaging; the dry raw materials dried by the local beer mill roller are purchased in a remote area, the raw materials are put into a crushing package after being inspected to be qualified by a beer yeast powder production factory, and the raw materials with higher water content are dried in a drying room and then crushed and packaged.
As mentioned above, in the crushing and packaging process, the materials enter the hammer type crusher through the pre-cleaning sieve, the lifting machine and the screw feeder to be crushed, and are conveyed to the cyclone separator through the centrifugal fan, the powdered yeast is put into a finished product packaging bag through the continuous discharging device at the lower section, and the finished product is sampled and tested. In this link, if the pulverizing effect of the hammer mill is insufficient, the test of the sampled sample in the final product may be failed, and if the pulverizing is too sufficient, more part of the beer yeast powder may be discharged in the form of dust. That is, it is important to evaluate the pulverization state of the beer yeast powder in the pulverization packaging step. Therefore, an optimized beer yeast powder intelligent production scheme is expected, which can accurately evaluate the crushing state of the beer yeast powder so as to optimize the production efficiency and the finished product quality of the beer yeast powder.
Accordingly, the evaluation of the pulverization state of the beer yeast powder can be converted into an image-based classification problem, that is, image features are extracted from the pulverized image of the yeast material pulverized by the hammer mill using the convolutional neural network model as a feature extractor, and the obtained image features are passed through a classifier to obtain a classification result indicating whether the pulverization degree of the yeast material meets a predetermined requirement.
However, since the yeast material is pulverized into fine particles in the pulverized image, which brings a challenge to image acquisition, it is difficult for a general camera to extract a high-resolution image. Even if the images with high resolution can be collected, the absolute value of the particle size number in the image career can be counted through the image semantic segmentation technology and the like, so that the crushing degree of the yeast material is evaluated, however, the crushed particles are extremely small-size objects, the image semantic segmentation technology cannot achieve accurate segmentation easily, and the accurate evaluation of the crushing degree based on absolute value statistics is low.
In particular, in the technical scheme of the application, the relative index of the crushed particles rather than the absolute index is used as the entry point of the crushed degree evaluation to classify and judge the crushed degree of the yeast material. Specifically, if the raw material of the yeast material is pulverized relatively well, the distribution state of the dust particles is relatively similar in different visual fields, and if the raw material of the yeast material is pulverized relatively poorly, the distribution state of the dust particles is greatly different in different visual fields.
Based on this, in the technical scheme of this application, gather the post-crushing image of yeast material raw materials through hammer mill crushing by the camera at first. Here, need not carry out absolute value statistics to the crushing granule of yeast material raw materials, consequently, need not adopt high definition digtal camera, only need ordinary camera can. Then, to haveAnd respectively carrying out feature extraction on the crushed image by using a first convolution neural network and a second convolution neural network of the cavity convolution kernels with different cavity rates to obtain a first scale feature map and a second scale feature map. In particular, hole convolution is an extension of the convolution kernel, and the added portion of the convolution kernel expands the size of the entire convolution kernel by filling in 0, which can expand the field of view without pooling. That is, the network can be further expanded without losing the resolution to obtain multi-scale information under high resolution, for example, the hole convolution kernel can be written as
Here, the hole convolution kernels having different void ratios have substantially different sized receptive fields, which enable to capture dust particle distribution status characteristics under different sized fields in the pulverized image.
Then, a difference feature map between the first scale feature map and the second scale feature map is calculated, for example, in a specific example of the present application, a difference by position between the first scale feature map and the second scale feature map is calculated to obtain the difference feature map, that is, a difference between features of the pulverized image in different visual fields is represented by a pixel-by-pixel difference between the first scale feature map and the second scale feature map. Then, the differential characteristic diagram is passed through a classifier to obtain a classification result for indicating whether the crushing degree of the yeast material meets the predetermined requirement.
In particular, in the technical solution of the present application, since the first scale feature map and the second scale feature map are obtained from the same source image through a convolutional neural network having hole convolution kernels with different hole rates, the first scale feature map and the second scale feature map have inconsistent distribution in spatial dimension, and therefore there are feature points outside the distribution formed by abnormal feature values in the calculated difference feature map.
Therefore, preferably, the difference feature map is optimized to suppress the influence of the out-of-distribution feature points, specifically:
f i is the differential feature map, e.g. the feature value for each position denoted F, μ and σ are the feature set F i E mean and variance of F, and N is the scale of the difference feature map F, i.e. width x height x number of channels.
Here, by taking the feature set as an adaptive instance, performing dynamic generation type information normalization on a single feature value based on intrinsic prior (intrinsic priorities) information of statistical features of the feature set, and performing invariant description in a set distribution domain by taking normalized mode length information of the feature set as a bias, feature optimization is achieved that shields disturbance distribution of a particular instance as much as possible, that is, statistical information normalization of the adaptive instance is achieved to suppress the influence of out-of-distribution feature points. Like this, can carry out more accurate aassessment to beer yeast powder's crushing state to optimize beer yeast powder's production efficiency and finished product quality.
In the beer yeast powder intelligent production scheme, a centrifugal fan starting instruction can be generated when the crushing degree of the yeast material raw material is detected to meet the preset requirement, and the centrifugal fan starting instruction is used for starting the centrifugal fan.
Based on this, the application provides an intelligent production method of beer yeast powder, which comprises the following steps: acquiring a crushed image of the yeast material crushed by the hammer crusher and collected by a camera; obtaining a first scale feature map by passing the crushed image through a first convolution neural network using a hole convolution kernel with a first hole rate; passing the pulverized image through a second convolutional neural network using a hole convolutional kernel with a second void rate to obtain a second scale feature map, wherein the first void rate is different from the second void rate; calculating a difference feature map between the first scale feature map and the second scale feature map; based on the scale of the differential feature map, correcting the feature value of each position in the differential feature map to obtain a corrected differential feature map, wherein the scale of the differential feature map is obtained by multiplying the width of the differential feature map by the height and then multiplying the channel number; and passing the corrected difference characteristic diagram through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the crushing degree of the yeast material meets the preset requirement or not.
Fig. 2 illustrates a scene schematic diagram of an intelligent production method of beer yeast powder according to an embodiment of the present application. As shown in fig. 2, in an application scenario of the beer yeast powder intelligent production method, firstly, a camera (e.g., C as illustrated in fig. 2) collects a crushed image of a yeast material crushed by a hammer mill (e.g., cr as illustrated in fig. 2). Further, the pulverized image of the yeast material pulverized by the hammer mill is input into a server (e.g., S as illustrated in fig. 2) deployed with an intelligent production algorithm of beer yeast powder, wherein the server is capable of processing the pulverized image of the yeast material pulverized by the hammer mill by the intelligent production algorithm of beer yeast powder to obtain a classification result indicating whether the pulverization degree of the yeast material meets a predetermined requirement, and generating a centrifugal fan start instruction for turning on the centrifugal fan (e.g., ce as illustrated in fig. 2) in response to the classification result indicating that the pulverization degree of the yeast material meets the predetermined requirement.
Having described the general principles of the present application, various non-limiting embodiments of the present application will now be described with reference to the accompanying drawings.
Exemplary method
Fig. 3 illustrates a flowchart of an intelligent production method of beer yeast powder according to an embodiment of the present application. As shown in fig. 3, the method for intelligently producing beer yeast powder according to the embodiment of the present application includes: s110, acquiring a crushed image of the yeast material crushed by the hammer mill and collected by a camera; s120, obtaining a first scale characteristic map by the crushed image through a first convolution neural network with a cavity convolution kernel with a first cavity rate; s130, enabling the crushed image to pass through a second convolution neural network with a cavity convolution kernel with a second cavity rate to obtain a second scale feature map, wherein the first cavity rate is different from the second cavity rate; s140, calculating a difference feature map between the first scale feature map and the second scale feature map; s150, correcting the characteristic value of each position in the differential characteristic diagram based on the scale of the differential characteristic diagram to obtain a corrected differential characteristic diagram, wherein the scale of the differential characteristic diagram is the product of the width of the differential characteristic diagram and the height of the differential characteristic diagram and then the product of the width of the differential characteristic diagram and the number of channels; and S160, passing the corrected differential characteristic diagram through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the crushing degree of the yeast material meets the preset requirement or not.
Fig. 4 illustrates a schematic architecture diagram of an intelligent production method of beer yeast powder according to an embodiment of the present application. As shown in fig. 4, a pulverized image of the yeast material pulverized by the hammer mill, which is captured by the camera, is first acquired. Then, the crushed image is passed through a first convolution neural network using a hole convolution kernel with a first void rate to obtain a first scale feature map, and the crushed image is passed through a second convolution neural network using a hole convolution kernel with a second void rate to obtain a second scale feature map, wherein the first void rate is different from the second void rate. Then, a difference feature map between the first scale feature map and the second scale feature map is calculated. And then, based on the scale of the differential feature map, correcting the feature value of each position in the differential feature map to obtain a corrected differential feature map, wherein the scale of the differential feature map is obtained by multiplying the width of the differential feature map by the height and then multiplying the channel number. And then, passing the corrected differential characteristic diagram through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the crushing degree of the yeast material meets the preset requirement or not.
In step S110, a pulverized image of the yeast material pulverized by the hammer mill, which is captured by the camera, is acquired. In the crushing and packaging link, the materials enter a hammer type crusher through a primary cleaning sieve, a lifter and a spiral feeder to be crushed and are conveyed to a cyclone separator through a centrifugal fan, the powdery yeast is put into a finished product packaging bag through a continuous discharger at the lower section, and a sample is sampled and tested in a finished product. In this link, if the pulverizing effect of the hammer mill is insufficient, the test of the sampled sample in the final product may be failed, and if the pulverizing is too sufficient, more part of the beer yeast powder may be discharged in the form of dust. That is, it is important to evaluate the pulverization state of beer yeast powder in the pulverization and packaging step. Accordingly, the evaluation of the pulverization state of the beer yeast powder can be converted into an image-based classification problem, that is, image features are extracted from the pulverized image of the yeast material pulverized by the hammer mill using the convolutional neural network model as a feature extractor, and the obtained image features are passed through a classifier to obtain a classification result indicating whether the pulverization degree of the yeast material meets a predetermined requirement. However, since the yeast material is pulverized into fine particles in the pulverized image, which brings a challenge to image acquisition, it is difficult for a general camera to extract a high-resolution image. Even if a high-resolution image can be acquired, the absolute value of the particle size number in an image career can be counted through image semantic segmentation and other technologies, so that the crushing degree of the yeast material is evaluated, however, the accurate segmentation is difficult to achieve through the image semantic segmentation technology due to the fact that the crushed particles are extremely small-size objects, and the accurate evaluation of the crushing degree based on the absolute value statistics is not high. In particular, in the technical scheme of the application, the relative index of the crushed particles rather than the absolute index is used as the entry point for the crushing degree evaluation to classify and judge the crushing degree of the yeast material. Specifically, if the raw material of the yeast material is pulverized relatively well, the distribution states of the dust particles are relatively similar in different visual fields, and if the raw material of the yeast material is pulverized relatively poorly, the distribution states of the dust particles are greatly different in different visual fields. Therefore, absolute value statistics of the pulverized particles of the yeast material is not required, that is, a high-definition camera is not required, and only a common camera is required.
In step S120, the crushed image is passed through a first convolution neural network using a hole convolution kernel having a first hole rate to obtain a first scale feature map. In particular, hole convolution is an extension of the convolution kernel, with the added portion of the convolution kernel enlarging the size of the entire convolution kernel by filling in 0's, which can enlarge the field of view without using pooling, e.g., the hole convolution kernel can be written as
That is, the network can further expand the reception field without losing the resolution, so as to obtain the multi-scale information under high resolution. By utilizing the hole convolution, the invariance of the relative space position of the characteristic layer and the information integrity can be kept. Here, the hole convolution kernels having different void ratios have substantially different sized receptive fields, which enable to capture dust particle distribution status characteristics under different sized fields in the pulverized image.
In one example, in the above beer yeast powder intelligent production method, the passing the crushed image through a first convolution neural network using a hole convolution kernel with a first hole rate to obtain a first scale feature map includes: performing convolution processing, pooling processing and nonlinear activation processing based on the hole convolution kernel with the first hole rate on input data respectively in forward pass of layers by using each layer of the first convolution neural network to output the first scale feature map from the last layer of the first convolution neural network, wherein the input of the first layer of the first convolution neural network is the crushed image.
In step S130, the crushed image is passed through a second convolutional neural network using a hole convolutional kernel having a second void rate to obtain a second scale feature map, wherein the first void rate is different from the second void rate.
In one example, in the above beer yeast powder intelligent production method, the passing the crushed image through a second convolutional neural network using a hole convolutional kernel with a second voidage to obtain a second scale feature map, where the first voidage is different from the second voidage, includes: and performing convolution processing, pooling processing and nonlinear activation processing based on the hole convolution kernel with the second hole rate on input data respectively in forward transmission of layers by using the layers of the second convolutional neural network to output the second scale feature map from the last layer of the second convolutional neural network, wherein the input of the first layer of the second convolutional neural network is the crushed image.
In step S140, a difference feature map between the first scale feature map and the second scale feature map is calculated. That is, the difference in position between the first scale feature map and the second scale feature map is calculated to obtain the difference feature map, that is, the difference between the features of the pulverized image in different visual fields is represented by the pixel-by-pixel difference between the first scale feature map and the second scale feature map.
In one example, in the above beer yeast powder intelligent production method, the calculating a difference feature map between the first scale feature map and the second scale feature map includes: calculating the difference by position between the first scale feature map and the second scale feature map according to the following formula to obtain the difference feature map; wherein the formula is:
F 1 is a first scale feature map, F 2 Is a second scale feature map, F d Is a differential characteristic diagram.
In step S150, feature values at various positions in the differential feature map are corrected based on a scale of the differential feature map to obtain a corrected differential feature map, where the scale of the differential feature map is obtained by multiplying a width of the differential feature map by a height and then by a channel number. Since the first scale feature map and the second scale feature map are obtained from the same source image through the convolutional neural network with the hole convolution kernels with different hole rates, the first scale feature map and the second scale feature map have inconsistent distribution on the spatial dimension, and therefore, there are feature points outside the distribution formed by abnormal feature values in the calculated difference feature map. Therefore, the differential feature map is optimized to suppress the influence of the out-of-distribution feature points.
In an example, in the above beer yeast powder intelligent production method, the correcting the feature values at the positions in the difference feature map based on the scale of the difference feature map to obtain a corrected difference feature map includes: based on the scale of the differential feature map, correcting the feature value of each position in the differential feature map by the following formula to obtain the corrected differential feature map; wherein the formula is
f i Is the feature value of each position in the difference feature map, μ and σ are the mean and variance of the feature values of all positions in the difference feature map, and N is the scale of the difference feature map, i.e., width x height x number of channels.
Here, by taking the feature set as an adaptive instance, performing dynamic generation type information normalization on a single feature value based on intrinsic prior (intrinsic priorities) information of statistical features of the feature set, and performing invariant description in a set distribution domain by taking normalized mode length information of the feature set as a bias, feature optimization is achieved that shields disturbance distribution of a particular instance as much as possible, that is, statistical information normalization of the adaptive instance is achieved to suppress the influence of out-of-distribution feature points. Like this, can carry out more accurate aassessment to beer yeast powder's crushing state to optimize beer yeast powder's production efficiency and finished product quality.
In step S160, the corrected differential feature map is passed through a classifier to obtain a classification result, where the classification result is used to indicate whether the pulverization degree of the yeast material meets a predetermined requirement.
In one example, in the above beer yeast powder intelligent production method, the passing the corrected differential feature map through a classifier to obtain a classification result includes: processing the differential feature map by using the classifier according to the following formula to obtain the classification result, wherein the formula is as follows: o = softmax { (W) n ,B n ):…:(W 1 ,B 1 ) L Project (F), where Project (F) represents the projection of the corrected differential feature map as a vector, W 1 To W n As a weight matrix for each fully connected layer, B 1 To B n A bias matrix representing the layers of the fully connected layer.
In an example, in the above beer yeast powder intelligent production method, the beer yeast powder intelligent production method further includes: and responding to the classification result that the crushing degree of the yeast raw material meets the preset requirement, and generating a centrifugal fan starting instruction, wherein the centrifugal fan starting instruction is used for starting the centrifugal fan. That is, when it is detected that the degree of pulverization of the yeast material meets a predetermined requirement, a centrifugal fan start-up instruction for turning on the centrifugal fan may be generated.
In summary, the beer yeast powder intelligent production method based on the embodiment of the application is clarified, the method includes that a first scale feature map and a second scale feature map are obtained by respectively passing a crushed image of a yeast material crushed by a hammer mill and acquired by a camera through a first convolution neural network with a cavity convolution kernel of a first voidage and a second convolution neural network with a cavity convolution kernel of a second voidage, feature values of positions in a difference feature map between the first scale feature map and the second scale feature map are corrected to obtain a corrected difference feature map, and then a classification result is obtained through a classifier, wherein the classification result is used for indicating whether the crushing degree of the yeast material meets a preset requirement. Through such a mode, carry out accurate aassessment to the crushing state of beer yeast powder to optimize the crushing packing link in the beer yeast powder production process, and promote the production efficiency and the finished product quality of beer yeast powder.
Exemplary System
Fig. 5 illustrates a block diagram of an intelligent production system of beer yeast powder according to an embodiment of the present application. As shown in fig. 5, the system 100 for intelligently producing beer yeast powder according to the embodiment of the present application includes: an image acquisition unit 110 for acquiring a pulverized image of the yeast material pulverized by the hammer mill acquired by the camera; a first convolution neural network unit 120, which passes the crushed image through a first convolution neural network using a hole convolution kernel with a first hole rate to obtain a first scale feature map; a second convolutional neural network unit 130, which passes the crushed image through a second convolutional neural network using a hole convolutional kernel having a second void rate to obtain a second scale feature map, wherein the first void rate is different from the second void rate; a difference feature map generation unit 140 that calculates a difference feature map between the first scale feature map and the second scale feature map; the corrected differential feature map generation unit 150 corrects the feature values of each position in the differential feature map based on the scale of the differential feature map to obtain a corrected differential feature map, wherein the scale of the differential feature map is obtained by multiplying the width of the differential feature map by the height and then by the number of channels; and an evaluation result generation unit 160, which passes the corrected difference feature map through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the crushing degree of the yeast material meets the preset requirement.
In an example, in the above beer yeast powder intelligent production system 100, the first convolutional neural network unit 120 is further configured to: performing convolution processing, pooling processing and nonlinear activation processing based on the hole convolution kernel with the first hole rate on input data respectively in forward pass of layers by using each layer of the first convolution neural network to output the first scale feature map from the last layer of the first convolution neural network, wherein the input of the first layer of the first convolution neural network is the crushed image.
In one example, in the above beer yeast intelligent production system 100, the second convolutional neural network unit 130 is further configured to: and performing convolution processing, pooling processing and nonlinear activation processing based on the hole convolution kernel with the second hole rate on input data respectively in forward transmission of layers by using the layers of the second convolutional neural network to output the second scale feature map from the last layer of the second convolutional neural network, wherein the input of the first layer of the second convolutional neural network is the crushed image.
In an example, in the above beer yeast powder intelligent production system 100, the difference feature map generating unit 140 is further configured to: calculating the difference by position between the first scale feature map and the second scale feature map according to the following formula to obtain the difference feature map; wherein the formula is:
F 1 is a first scale feature map, F 2 Is a second scale feature map, F d Is a differential characteristic diagram.
In an example, in the above beer yeast powder intelligent production system 100, the corrected difference feature map generating unit 150 is further configured to: based on the scale of the differential feature map, correcting the feature value of each position in the differential feature map by the following formula to obtain the corrected differential feature map; wherein the formula is
f i Is the feature value of each location in the difference feature map, μ and σ are the mean and variance of the feature values of all locations in the difference feature map, and N is the scale of the difference feature map.
In an example, in the above beer yeast intelligent production system 100, the evaluation result generating unit 160 is further configured to: processing the differential feature map by using the classifier according to the following formula to obtain the classification result, wherein the formula is as follows:
O＝softmax{(W n ,B n ):…:(W 1 ,B 1 )|Project(F)}
wherein Project (F) represents projecting the corrected difference feature map as a vector, W 1 To W n As a weight matrix for each fully connected layer, B 1 To B n A bias matrix representing the layers of the fully connected layer.
In an example, in the above beer yeast powder intelligent production system 100, the beer yeast powder intelligent production system is further configured to: and generating a centrifugal fan starting instruction in response to the classification result that the crushing degree of the yeast material meets the preset requirement, wherein the centrifugal fan starting instruction is used for starting the centrifugal fan.
In summary, the beer yeast powder intelligent production system 100 based on the embodiment of the present application is illustrated, in which the crushed images of the yeast material crushed by the hammer mill and collected by the camera are respectively passed through the first convolution neural network with the cavity convolution kernel having the first voidage and the second convolution neural network with the cavity convolution kernel having the second voidage to obtain the first scale feature map and the second scale feature map, the feature values of the positions in the difference feature map between the first scale feature map and the second scale feature map are corrected to obtain the corrected difference feature map, and then the corrected difference feature map is passed through the classifier to obtain the classification result, where the classification result is used to indicate whether the crushing degree of the yeast material meets the predetermined requirement. Through such mode, carry out accurate aassessment to beer yeast powder's crushing state to optimize the crushing packing link in the beer yeast powder production process, and promote beer yeast powder's production efficiency and finished product quality.
As described above, the beer yeast powder intelligent production system 100 according to the embodiment of the present application can be implemented in various terminal devices, such as a server for the beer yeast powder intelligent production method. In one example, the beer yeast intelligent production system 100 according to the embodiment of the present application can be integrated into a terminal device as a software module and/or a hardware module. For example, the beer yeast powder intelligent production system 100 can be a software module in the operating system of the terminal device, or can be an application program developed for the terminal device; of course, the beer yeast powder intelligent production system 100 can also be one of the hardware modules of the terminal device.
Alternatively, in another example, the beer yeast powder intelligent production system 100 and the terminal device may be separate devices, and the beer yeast powder intelligent production system 100 may be connected to the terminal device through a wired and/or wireless network and transmit the interactive information according to the agreed data format.
Exemplary electronic device
Next, an electronic apparatus according to an embodiment of the present application is described with reference to fig. 6. FIG. 6 illustrates a block diagram of an electronic device in accordance with an embodiment of the present application.
As shown in fig. 6, the electronic device 10 includes one or more processors 11 and memory 12.
The processor 11 may be a Central Processing Unit (CPU) or other form of processing unit having data processing capabilities and/or instruction execution capabilities, and may control other components in the electronic device 10 to perform desired functions.
In one example, the electronic device 10 may further include: an input device 13 and an output device 14, which are interconnected by a bus system and/or other form of connection mechanism (not shown).
The input device 13 may include, for example, a keyboard, a mouse, and the like.
The output device 14 can output various information including the classification result to the outside. The output devices 14 may include, for example, a display, speakers, a printer, and a communication network and its connected remote output devices, among others.
Of course, for simplicity, only some of the components of the electronic device 10 relevant to the present application are shown in fig. 6, and components such as buses, input/output interfaces, and the like are omitted. In addition, the electronic device 10 may include any other suitable components depending on the particular application.
Exemplary computer program product and computer-readable storage Medium
In addition to the above-described methods and apparatus, embodiments of the present application may also be a computer program product comprising computer program instructions that, when executed by a processor, cause the processor to perform the steps in the functions of the beer yeast powder intelligent production method according to the various embodiments of the present application described in the "exemplary methods" section of this specification, above.
The computer program product may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages, for carrying out operations according to embodiments of the present application. The program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device, or entirely on the remote computing device or server.
Furthermore, embodiments of the present application may also be a computer readable storage medium having stored thereon computer program instructions, which, when executed by a processor, cause the processor to perform the steps in the functions in the beer yeast intelligent production method according to various embodiments of the present application described in the "exemplary methods" section above in this specification.
The computer-readable storage medium may take any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may include, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
The foregoing describes the general principles of the present application in conjunction with specific embodiments, however, it is noted that the advantages, effects, etc. mentioned in the present application are merely examples and are not limiting, and they should not be considered essential to the various embodiments of the present application. Furthermore, the foregoing disclosure of specific details is for the purpose of illustration and description and is not intended to be limiting, since the foregoing disclosure is not intended to be exhaustive or to limit the disclosure to the precise details disclosed.
The block diagrams of devices, apparatuses, systems referred to in this application are only given as illustrative examples and are not intended to require or imply that the connections, arrangements, configurations, etc. must be made in the manner shown in the block diagrams. These devices, apparatuses, devices, systems may be connected, arranged, configured in any manner, as will be appreciated by those skilled in the art. Words such as "including," "comprising," "having," and the like are open-ended words that mean "including, but not limited to," and are used interchangeably therewith. As used herein, the words "or" and "refer to, and are used interchangeably with, the word" and/or, "unless the context clearly dictates otherwise. The word "such as" is used herein to mean, and is used interchangeably with, the phrase "such as but not limited to".
It should also be noted that in the devices, apparatuses, and methods of the present application, the components or steps may be decomposed and/or recombined. These decompositions and/or recombinations are to be considered as equivalents of the present application.
The previous description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other aspects without departing from the scope of the application. Thus, the present application is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
The foregoing description has been presented for purposes of illustration and description. Furthermore, the description is not intended to limit embodiments of the application to the form disclosed herein. While a number of example aspects and embodiments have been discussed above, those of skill in the art will recognize certain variations, modifications, alterations, additions and sub-combinations thereof.
Claims (10)
1. An intelligent production method of beer yeast powder is characterized by comprising the following steps:
acquiring a crushed image of the yeast material crushed by the hammer crusher and collected by a camera;
obtaining a first scale feature map by passing the crushed image through a first convolution neural network using a hole convolution kernel with a first hole rate;
passing the pulverized image through a second convolutional neural network using a hole convolutional kernel with a second void rate to obtain a second scale feature map, wherein the first void rate is different from the second void rate;
calculating a difference feature map between the first scale feature map and the second scale feature map;
correcting the feature values of all positions in the differential feature map based on the scale of the differential feature map to obtain a corrected differential feature map, wherein the scale of the differential feature map is obtained by multiplying the width of the differential feature map by the height and then multiplying by the number of channels; and
and passing the corrected differential characteristic diagram through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the crushing degree of the yeast material meets the preset requirement or not.
2. The intelligent beer yeast powder production method according to claim 1, wherein the step of obtaining the first scale feature map by using a first convolutional neural network with a first void convolutional kernel with a first void rate comprises the following steps:
performing convolution processing, pooling processing and nonlinear activation processing based on the hole convolution kernel with the first hole rate on input data respectively in forward pass of layers by using each layer of the first convolution neural network to output the first scale feature map from the last layer of the first convolution neural network, wherein the input of the first layer of the first convolution neural network is the crushed image.
3. The intelligent production method of beer yeast powder according to claim 2, wherein the passing the pulverized image through a second convolutional neural network using a hole convolutional kernel with a second voidage to obtain a second scale feature map, wherein the first voidage is different from the second voidage, and comprises:
and performing convolution processing, pooling processing and nonlinear activation processing based on the hole convolution kernel with the second hole rate on input data respectively in forward transmission of layers by using the layers of the second convolutional neural network to output the second scale feature map from the last layer of the second convolutional neural network, wherein the input of the first layer of the second convolutional neural network is the crushed image.
4. The intelligent production method of beer yeast powder according to claim 3, wherein the calculating a difference feature map between the first scale feature map and the second scale feature map comprises:
calculating the difference by position between the first scale feature map and the second scale feature map according to the following formula to obtain the difference feature map;
wherein the formula is:
F 1 is a first scale feature map, F 2 Is a second scale feature map, F d Is a differential characteristic diagram.
5. The intelligent production method of beer yeast powder according to claim 4, wherein the step of correcting the feature value of each position in the difference feature map based on the scale of the difference feature map to obtain a corrected difference feature map comprises:
based on the scale of the differential feature map, correcting the feature value of each position in the differential feature map by the following formula to obtain the corrected differential feature map;
wherein the formula is
f i Is the feature value of each location in the difference feature map, μ and σ are the mean and variance of the feature values of all locations in the difference feature map, and N is the scale of the difference feature map.
6. The intelligent production method of beer yeast powder according to claim 5, wherein the step of passing the corrected difference feature map through a classifier to obtain a classification result comprises:
processing the differential feature map by using the classifier according to the following formula to obtain the classification result, wherein the formula is as follows: o = softmax { (W) n ,B n ):…:(W 1 ,B 1 ) L Project (F), where Project (F) represents the projection of the corrected differential feature map as a vector, W 1 To W n As a weight matrix for each fully connected layer, B 1 To B n A bias matrix representing the layers of the fully connected layer.
7. The intelligent production method of beer yeast powder according to claim 6, wherein the intelligent production method of beer yeast powder further comprises:
and responding to the classification result that the crushing degree of the yeast raw material meets the preset requirement, and generating a centrifugal fan starting instruction, wherein the centrifugal fan starting instruction is used for starting the centrifugal fan.
8. The utility model provides a beer yeast powder intelligent production system which characterized in that includes:
the image acquisition unit is used for acquiring a crushed image of the yeast material crushed by the hammer mill and acquired by the camera;
the first convolution neural network unit is used for obtaining a first scale characteristic diagram of the crushed image through a first convolution neural network with a hole convolution kernel with a first hole rate;
a second convolutional neural network unit for passing the pulverized image through a second convolutional neural network using a hole convolutional kernel having a second void rate to obtain a second scale feature map, wherein the first void rate is different from the second void rate;
a difference feature map generation unit which calculates a difference feature map between the first scale feature map and the second scale feature map;
the corrected differential feature map generation unit is used for correcting the feature values of all positions in the differential feature map based on the scale of the differential feature map to obtain a corrected differential feature map, wherein the scale of the differential feature map is obtained by multiplying the width of the differential feature map by the height and then multiplying the width by the number of channels; and
and the evaluation result generating unit is used for enabling the corrected differential characteristic diagram to pass through a classifier to obtain a classification result, and the classification result is used for indicating whether the crushing degree of the yeast material meets the preset requirement or not.
9. The intelligent beer yeast powder production system according to claim 8, wherein the first convolutional neural network unit is further configured to:
performing convolution processing, pooling processing and nonlinear activation processing based on the hole convolution kernel with the first hole rate on input data respectively in forward pass of layers by using each layer of the first convolution neural network to output the first scale feature map from the last layer of the first convolution neural network, wherein the input of the first layer of the first convolution neural network is the crushed image.
10. The intelligent production system of beer yeast powder according to claim 9, further comprising:
and generating a centrifugal fan starting instruction in response to the classification result that the crushing degree of the yeast material meets the preset requirement, wherein the centrifugal fan starting instruction is used for starting the centrifugal fan.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211115228.6A CN115393650A (en) | 2022-09-14 | 2022-09-14 | Beer yeast powder intelligent production method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211115228.6A CN115393650A (en) | 2022-09-14 | 2022-09-14 | Beer yeast powder intelligent production method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115393650A true CN115393650A (en) | 2022-11-25 |
Family
ID=84126730
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211115228.6A Withdrawn CN115393650A (en) | 2022-09-14 | 2022-09-14 | Beer yeast powder intelligent production method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115393650A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115849519A (en) * | 2022-12-30 | 2023-03-28 | 浙江致远环境科技股份有限公司 | Organic modularization electrocatalytic oxidation treatment device |
-
2022
- 2022-09-14 CN CN202211115228.6A patent/CN115393650A/en not_active Withdrawn
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115849519A (en) * | 2022-12-30 | 2023-03-28 | 浙江致远环境科技股份有限公司 | Organic modularization electrocatalytic oxidation treatment device |
| CN115849519B (en) * | 2022-12-30 | 2024-03-22 | 浙江致远环境科技股份有限公司 | Organic modularized electrocatalytic oxidation treatment device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11295173B2 (en) | Image identification apparatus, image identification method, and recording medium | |
| CN104138854B (en) | Ore separation system and method based on pseudo-dual intensity radial imaging | |
| Majumdar et al. | Classification of bulk samples of cereal grains using machine vision | |
| CN109598715B (en) | Material granularity online detection method based on machine vision | |
| CN115393650A (en) | Beer yeast powder intelligent production method and system | |
| CN107966454A (en) | A kind of end plug defect detecting device and detection method based on FPGA | |
| CN107240134A (en) | A kind of industrial products rapid classification method and device based on online colour recognition | |
| CN102982305A (en) | Information processing apparatus and method of processing information, storage medium and program | |
| Yusoff et al. | Real-time hevea leaves diseases identification using Sobel edge algorithm on FPGA: A preliminary study | |
| CN111144151A (en) | High-speed dynamic bar code real-time detection method based on image recognition | |
| CN107066959A (en) | A kind of hyperspectral image classification method based on Steerable filter and linear space correlation information | |
| Koh et al. | Utilising convolutional neural networks to perform fast automated modal mineralogy analysis for thin-section optical microscopy | |
| CN112257711B (en) | Method for detecting damage fault of railway wagon floor | |
| CN114140684A (en) | Method, device and equipment for detecting coal blockage and coal leakage and storage medium | |
| WO2023093087A1 (en) | Image recognition-based coal conveying amount monitoring method and apparatus, device, and storage medium | |
| Zhang et al. | Online double-sided identification and eliminating system of unclosed-glumes rice seed based on machine vision | |
| CN117055249B (en) | Sampling quality inspection analysis system for liquid crystal display screen assembly processing line | |
| US8472696B2 (en) | Observation condition determination support device and observation condition determination support method | |
| CN116012293A (en) | Large-block identification system and method based on feature matching | |
| JP2007018176A (en) | Learning device, learning method, learning program, recording medium, and device and method for pattern recognition | |
| CN116342540A (en) | Packaging film defect detection method, device, equipment and storage medium | |
| CN115818166A (en) | Unattended automatic control method and system for wheel hopper continuous system | |
| EP4194108A1 (en) | Small-grain agricultural product color selection method combining area scan and line scan photoelectric features | |
| WO2020189043A1 (en) | Learning model generation method, learning model, inspection device, abnormality detection method, and computer program | |
| CN114611863A (en) | E-commerce product packaging quality detection method based on big data |
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 | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20221125 |