Flour humidifying method based on chaos state single view reconstruction of flour particle groups

Through a single-view depth estimation method based on loop prediction and point cloud density algorithm, combined with the improved error correction network and flour humidifier design, the problems of occlusion area depth estimation and powder drying in flour processing are solved, and flour granules are realized Precise reconstruction and humidification treatment improve processing efficiency and product quality.

CN118982469APending Publication Date: 2024-11-19CANGZHOU DEV ZONE HENGXIANGSHENG GRAIN MASCH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411057736.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-02
Publication Date
2024-11-19

AI Technical Summary

Technical Problem

During the flour processing process, the traditional single-view reconstruction method cannot effectively estimate the depth information of the blocked flour granules, resulting in poor reconstruction effect, and the grinding of powder is too dry to affect subsequent processing, which can easily lead to the silence and blockage of the processing equipment Phenomenon.

Method used

Using a single-view depth estimation method based on loop prediction, combining the pixel-to-pixel conversion network model and point cloud density algorithm, sparse point cloud density is densely processed through normal and uniform distribution generation methods, and the network is corrected using improved error correction. , realize the precise reconstruction of the chaotic state of flour particles, and at the same time design a flour humidifier to cooperate with spray components and mix components for humidification.

Benefits of technology

It improves the accuracy and reconstruction effect of the chaotic interface morphology analysis of flour granules, avoids processing equipment problems caused by excessive drying of powder, and ensures the smooth progress of the processing process and product quality.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118982469A_ABST
    Figure CN118982469A_ABST
Patent Text Reader

Abstract

The flour humidifier comprises a machine body, an upper cover, a motor, a mixing assembly, a spraying assembly and an intelligent control assembly, an opening is formed in the top of the machine body, and the upper cover is detachably installed and fixed to the top of the machine body and covers the opening of the machine body; the motor is mounted on the upper cover, and the output end of the motor is in driving connection with the upper end of the mixing assembly, so that the mixing assembly is rotatably arranged in the inner cavity of the machine body. Through cooperative use of the mixing assembly and the spraying assembly, when powder such as flour is processed, water supplementing and humidifying operation can be conducted on the powder, and therefore the situation that the subsequent processing process is affected due to the fact that the powder is too dry can be avoided; the intelligent control assembly is used for controlling cooperation of the mixing assembly and the spraying assembly, powder is replenished with water and humidified evenly, powder caking and caking generated in the powder replenishing and humidifying process can be reduced, the problem that in a single-view reconstruction method, a shielded area cannot be effectively reconstructed is effectively solved, and the detail reconstruction method has remarkable advantages in the detail reconstruction effect.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of grain processing machinery, and in particular to a flour humidification method, a system and a flour humidifier based on chaotic single-view reconstruction of flour particle groups. Background Art

[0002] In the processing of grain powders such as flour, impurities are usually removed by screening first, and then the small hair tufts and attached dirt on the surface of the material are cleaned; then the purified grain particles are soaked in water for 18 to 72 hours to become tough, and the bran is easily separated from the particles, at which time the particles can be ground; after coarse grinding, they can be passed through a selector to retain pure endosperm particles by utilizing different specific gravities. The light bran will float up, and the remaining heavier endosperm powder will first pass through a smooth roller mill to flatten the endosperm and grind it into fine powder. The fine powder is then separated into types and grades by a multi-layer flat rotary sifter.

[0003] In the process of grinding wheat into powder, the semi-finished powder needs to lose moisture and dry, which affects the progress of the next process and the quality of the product. It may even make the processing machinery prone to blockage and stalling in the subsequent processing, resulting in long downtime and cleaning time and low work efficiency. For this reason, we proposed a flour humidifier to solve the above problems.

[0004] With the vigorous development of digital processing technology, the demand for high efficiency, automation and digitalization in flour processing management has become increasingly prominent. Among them, the product interface reconstruction technology in the flour processing process, as an important part of digital flour processing management, provides flour processing managers with technical support such as product interface morphology analysis, product interface morphology change image monitoring and intelligent humidification. In the target state reconstruction technology, the reconstruction method based on a single view has the advantage of simpler data acquisition and processing compared with the multi-view method because it only requires one RGB image as input data. However, the existing research on target state reconstruction mainly focuses on the reconstruction of physical objects in virtual simulation scenes, and does not consider the complexity and occlusion interference of the chaotic state of physical particle clusters in real scenes. In actual scenes, the mutual occlusion between particles increases the complexity and difficulty of reconstruction. Solving the occlusion problem is crucial to the complete reconstruction of the flour particle cluster structure, which not only helps to achieve more accurate analysis of the chaotic state interface morphology of flour particle clusters, but also supports management decisions such as monitoring of the chaotic state interface morphology changes of flour particle clusters and intelligent pruning. Therefore, it is particularly important to study new single-view reconstruction methods to achieve complete reconstruction of the chaotic state of flour particle clusters in real scenes. Summary of the invention

[0005] 1. Technical issues to be solved

[0006] In view of the above problems, the present invention proposes a flour humidification method and system based on single-view reconstruction of the chaotic state of flour particle clusters. In this method, firstly, a single-view depth estimation method based on cyclic prediction is proposed, which solves the problem that the traditional single-view reconstruction method cannot effectively estimate the depth information of occluded flour particle clusters; secondly, a point cloud density algorithm based on the analysis of the chaotic state morphological characteristics of flour particle clusters and a point cloud correction method based on point-by-point correction are proposed, which solve the problem that the preliminary obtained flour particle cluster point cloud is sparse and inaccurate in the single-view depth estimation method. Comparison with the traditional single-view reconstruction method and the multi-view reconstruction method shows that the single-view reconstruction method of the present invention shows good performance, is superior to the traditional single-view reconstruction method in reconstruction effect, and can effectively solve the problem that the occluded area cannot be effectively reconstructed in the traditional single-view reconstruction method; compared with the multi-view reconstruction method, the single-view reconstruction method proposed in the present invention shows significant advantages in the reconstruction effect of the chaotic state details of flour particle clusters.

[0007] In view of the above problems, the present invention also provides a flour humidifier, which solves the problem that the grain ground into powder is relatively dry and the powder is likely to affect the normal use of processing equipment in the subsequent processing process.

[0008] (II) Technical solution

[0009] To achieve the above object, the present invention provides the following technical solutions:

[0010] The first aspect of the present invention is to provide a flour humidification method based on chaotic state single view reconstruction of flour particle clusters, comprising:

[0011] S1, obtaining multiple two-dimensional images of the chaotic state of flour particles;

[0012] S2, inputting the multiple two-dimensional images into a pixel-to-pixel conversion network model, thereby solving the problem that the traditional depth estimation network cannot effectively estimate the depth information of the occluded area;

[0013] S3, establishing a point cloud densification algorithm based on normal distribution and uniform distribution, analyzing the morphological characteristics of the flour particle group based on the point cloud densification algorithm, and densifying the sparse point cloud using the normal and uniform distribution generation methods;

[0014] S4, establishing an improved error correction network, wherein the improved error correction network is used to correct errors caused by the pixel-to-pixel conversion network model and the point cloud dense algorithm, inputting the coordinates of the point cloud into the error correction network, and outputting the corrected coordinates at the corresponding position;

[0015] S5, reconstructing a chaotic single image of the flour particle cluster based on the corrected coordinates at the corresponding position;

[0016] S6, obtaining the humidity of the current flour particle mass based on the single view, and controlling the time and intensity of flour humidification based on the humidity of the current flour particle mass.

[0017] Preferably, S2 includes:

[0018] S21, building a pixel-to-pixel conversion network model;

[0019] S22, hierarchically storing the three-dimensional information of the chaotic state of the flour particle cluster into the plurality of two-dimensional images based on a point cloud three-dimensional information hierarchical storage strategy;

[0020] S23, based on the pixel-to-pixel conversion network model, cyclically predicting the depth information of the two-dimensional image that stores the three-dimensional information layer by layer, wherein the two-dimensional image depth information corresponds to the structural information of all occluded areas in the flour particle group; fusing the two-dimensional image depth information with the multiple two-dimensional images to obtain the morphological characteristics of the flour particle group.

[0021] Preferably, S3 includes:

[0022] S31, preprocessing and classifying the point cloud data based on the morphological characteristics of the flour particle group to obtain different types of points;

[0023] S32, different densification methods and parameters are used for the different types of points. The different densification methods include a method based on spatial patch diffusion and a method based on contour constraints. Based on certain spatial geometric rules and constraints, the existing sparse point cloud information is densified through deduction and calculation.

[0024] Preferably, the improved error correction network consists of a convolutional layer, a batch normalization layer and a fully connected layer. A data vector of size 1*4 passes through three convolutional layers (Conv-64, Conv-128 and Conv-256 layers) in sequence, and a batch normalization layer is interspersed in the convolutional layer to prevent the risk of gradient disappearance; then passes through four fully connected layers in sequence, and finally outputs a value of size 1*1 as the corrected y coordinate; multiple convolutional layers gradually extract deep-level features of point cloud data in a progressive manner; after a series of convolutional and batch normalization layer processing, the data is passed to the fully connected layer for further processing, so that the previously extracted features are globally integrated and mapped to the output space.

[0025] The second aspect of the present invention is to provide a flour humidification system based on single-view reconstruction of chaotic state of flour particle clusters, comprising:

[0026] An image acquisition module, used to acquire multiple two-dimensional images of the chaotic state of flour particle clusters;

[0027] A depth information estimation module, used to input the multiple two-dimensional images into a pixel-to-pixel conversion network model, thereby solving the problem that the traditional depth estimation network cannot effectively estimate the depth information of the occluded area;

[0028] A point cloud densification processing module is used to establish a point cloud densification algorithm based on normal distribution and uniform distribution, analyze the morphological characteristics of the flour particle group based on the point cloud densification algorithm, and perform densification processing on the sparse point cloud using the normal and uniform distribution generation methods;

[0029] An error correction module is used to establish an improved error correction network, wherein the improved error correction network is used to correct errors caused by the pixel-to-pixel conversion network model and the point cloud dense algorithm, input the coordinates of the point cloud to the error correction network, and output the corrected coordinates at the corresponding position;

[0030] A single view reconstruction module is used to reconstruct a chaotic single view of a flour particle group based on the corrected coordinates at the corresponding position;

[0031] The flour humidification module is used to obtain the humidity of the current flour particle group based on the single view, and control the time and intensity of flour humidification based on the humidity of the current flour particle group.

[0032] The third aspect of the present invention is to provide a flour humidifier, including a body, an upper cover, a motor, a mixing component, a spray component and an intelligent control component, wherein an opening is formed on the top of the body, and the upper cover is detachably mounted and fixed on the top of the body and covers the opening of the body; the motor is mounted on the upper cover, and the output end of the motor drives the upper end of the mixing component so that the mixing component can be rotatably arranged in the inner cavity of the body; the spray component is mounted on the upper cover, and the output end of the spray component extends into the top of the body, water and gas can be input into the spray component, and water mist can be sprayed into the body by using the output end of the spray component; a feeding part connected to the inner cavity of the body is formed on the upper cover, and the material fed through the feeding part can be dispersed by swinging the rotating mixing component, and the dispersed material and the water mist sprayed by the spray component are mixed under the stirring of the rotating mixing component; the intelligent control component is used to implement the control of the coordination of the mixing component and the spray component based on the method of the first aspect.

[0033] Preferably, the mixing assembly includes a rotating drum, a plurality of stirring blades and a connecting shaft, the plurality of stirring blades are all arranged on the outer wall of the rotating drum, and the stirring blades can be arranged to be inclined downward and upward, so that the stirring blades can be welded or fixed to the outer wall of the rotating drum at a certain angle, and the plurality of stirring blades are evenly distributed along the radial and axial directions of the rotating drum; the connecting shaft is fixed at the center of the upper surface of the vertical slats; the mixing assembly also includes a plurality of vertical slats and annular slats, the plurality of vertical slats are evenly distributed and fixed at the outer edge of the upper surface of the rotating drum, and the annular slats are respectively fixedly connected to the upper ends of the plurality of vertical slats, and a gap through which the material can pass is formed between two adjacent vertical slats.

[0034] Preferably, the spray assembly includes a pipe body, a water inlet, an air inlet and a plurality of nozzles, the water inlet and the air inlet are connected to the pipe body at one end facing the pipe body, water and gas can be respectively input into the pipe body by using the water inlet and the air inlet, the plurality of nozzles are connected and fixed at the bottom of the pipe body, and the plurality of nozzles are located above the mixing assembly.

[0035] Preferably, a transmission belt is connected to the output end of the motor, and the other end of the transmission belt is connected to the upper end of the connecting shaft, so that the motor drives the rotation of the mixing assembly; the outer cover of the transmission belt is provided with the transmission belt protection cover; the top of the upper cover is formed with a C-shaped cover, and the C-shaped cover covers the tube body on its inner side; the top of the upper cover is also fixedly connected with an end sleeve, and the end sleeve covers the upper end of the connecting shaft on its inner side.

[0036] Preferably, the machine body includes a hood, and the cross-section of the inner cavity of the hood is set to a polygon; an inspection port is formed on the outer wall of the hood and communicates with the inner cavity of the hood, and a plurality of threaded columns welded and fixed to the hood are arranged on the outside of the inspection port, and an inspection cover for covering the inspection port is provided, and the plurality of threaded columns pass through the inspection cover and are connected to nuts; the machine body also includes a cone bucket and a plurality of supporting legs, the cone bucket is fixed at the lower end of the hood and communicates with the inner cavity of the hood, a discharge port is formed at the bottom of the cone bucket, and a plurality of supporting legs are welded and fixed to the outer wall of the cone bucket.

[0037] (III) Beneficial effects

[0038] Compared with the prior art, the present invention provides a flour humidifier, which has the following beneficial effects:

[0039] 1. Establish a pixel-to-pixel conversion network model to solve the problem that the traditional depth estimation network cannot effectively estimate the depth information of the occluded area. This network model combines the point cloud three-dimensional information layered storage strategy, firstly stores the three-dimensional information of the chaotic state of flour particles in multiple two-dimensional images, and then predicts the depth information layer by layer through the improved pixel-to-pixel conversion network model; through the layer-by-layer prediction processing method, the network can more accurately capture the structural information of the occluded area, improving the accuracy of depth estimation.

[0040] 2. Establish a point cloud densification algorithm based on normal distribution and uniform distribution to solve the problem of sparse chaotic point cloud of flour particles after the depth image is predicted by the single-view depth estimation method. The algorithm analyzes the morphological characteristics of flour particles and densifies the sparse point cloud using the normal and uniform distribution generation methods.

[0041] 3. Establish an error correction network to correct the errors caused by the pixel-to-pixel conversion network model and the point cloud densification algorithm. The present invention adopts a layered cyclic prediction method when performing depth prediction, which causes the error to gradually accumulate in the network layer, making the final depth estimation result inaccurate; in addition, when performing point cloud densification, due to the complex structure of flour particles, the flour particles obtained by densification are not fine enough; in order to reduce the errors caused by depth estimation and densification, the present invention proposes an error correction network, which inputs the coordinates of the point cloud and outputs the corrected coordinates at the corresponding position. The role of the correction network is to make more precise corrections to the existing point cloud coordinates to reduce the accumulated errors and improve the accuracy of the overall reconstruction.

[0042] 4. In the present invention, by using the mixing component in conjunction with the spraying component, when processing powder materials such as flour, the powder materials can be replenished with water and humidified, thereby preventing the powder materials from being too dry and affecting the subsequent processing process; in addition, the mixing component and the spraying component can be used in conjunction to replenish water and humidify the powder materials more evenly, and the agglomeration and caking of the powder materials during the replenishment and humidification process of the powder materials can be reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a structural schematic diagram of a flour humidifier according to the implementation scheme;

[0044] Figure 2 for Figure 1 Schematic diagram of the structure of the flour humidifier after opening the inspection cover;

[0045] Figure 3 is a schematic structural diagram of a machine body according to an implementation scheme;

[0046] Figure 4 is a schematic structural diagram of an upper cover according to an implementation scheme;

[0047] Figure 5 is a schematic structural diagram of a mixing assembly according to an embodiment.

[0048] In the figure: 10, machine body; 11, machine cover; 12, inspection cover; 13, cone bucket; 14, supporting leg; 20, upper cover; 21, feeding part; 22, C-shaped cover; 23, end sleeve; 30, motor; 31, transmission belt protection cover; 40, mixing component; 41, rotating drum; 42, stirring blade; 43, vertical slats; 44, annular slats; 45, connecting shaft; 50, spray component; 51, water inlet; 52, air inlet; 53, nozzle. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] Embodiment 1

[0051] A flour humidification method based on chaotic state single-view reconstruction of flour particle clusters, comprising:

[0052] S1, obtaining multiple two-dimensional images of the chaotic state of flour particles;

[0053] S2, inputting the plurality of two-dimensional images into a pixel-to-pixel conversion network model, thereby solving the problem that the depth information of the occluded area cannot be effectively estimated in the traditional depth estimation network; comprising:

[0054] S21, building a pixel-to-pixel conversion network model;

[0055] S22, hierarchically storing the three-dimensional information of the chaotic state of the flour particle cluster into the plurality of two-dimensional images based on a point cloud three-dimensional information hierarchical storage strategy;

[0056] S23, based on the pixel-to-pixel conversion network model, cyclically predicting the depth information of the two-dimensional image that stores the three-dimensional information layer by layer, wherein the two-dimensional image depth information corresponds to the structural information of all occluded areas in the flour particle group; fusing the two-dimensional image depth information with the multiple two-dimensional images to obtain the morphological characteristics of the flour particle group;

[0057] S3, establishing a point cloud densification algorithm based on normal distribution and uniform distribution, analyzing the morphological characteristics of the flour particle group based on the point cloud densification algorithm, and densifying the sparse point cloud using the normal and uniform distribution generation methods;

[0058] As a preferred implementation, S3 includes:

[0059] S31, preprocessing and classifying the point cloud data based on the morphological characteristics of the flour particle group to obtain different types of points;

[0060] S32, using different densification methods and parameters for the different types of points, the different densification methods include a method based on spatial patch diffusion and a method based on contour constraints, based on certain spatial geometric rules and constraints, through deduction and calculation, the existing sparse point cloud information is densified. This classification processing method enables the algorithm to better adapt to the complexity and diversity of tree trunk morphology. This helps to improve the accuracy and reliability of the densification results, making the densified point cloud data closer to the actual tree trunk morphology;

[0061] S4, establishing an improved error correction network, wherein the improved error correction network is used to correct errors caused by the pixel-to-pixel conversion network model and the point cloud dense algorithm, inputting the coordinates of the point cloud into the error correction network, and outputting the corrected coordinates at the corresponding position;

[0062] As a preferred embodiment, the improved error correction network consists of a convolutional layer, a batch normalization layer and a fully connected layer. The data vector of size 1*4 passes through three convolutional layers (Conv-64, Conv-128 and Conv-256 layers) in sequence, and the batch normalization layer is interspersed in the convolutional layer to prevent the risk of gradient disappearance; then it passes through four fully connected layers in sequence, and finally outputs a value of size 1*1 as the corrected y coordinate. Multiple convolutional layers gradually extract the deep features of point cloud data in a progressive manner. Specifically, the data vector of size 1*4 first passes through the Conv-64 layer to preliminarily capture the local features of the data; then, the data passes through Conv-128 and Conv-256 in sequence to learn more abstract and advanced feature representations. Between the convolutional layers, the network is interspersed with batch normalization layers, so that by normalizing each batch of data, the gradient disappearance problem that may occur during training is solved, thereby accelerating the convergence speed of the network and improving the stability of the network. After a series of convolution and batch normalization layers, the data is passed to the fully connected layer for further processing. The function of the fully connected layer is to globally integrate the features extracted previously and map them to the output space. In this network, four fully connected layers, FC-512, FC-256, FC-128 and FC-1, are set up in sequence. By reducing the dimension layer by layer, the feature vector is gradually mapped to the final output space. After being processed by the fully connected layer, the network outputs a feature vector of size 512*4. Finally, this feature vector is output as a value of size 1*1, that is, the corrected y coordinate.

[0063] S5, reconstructing a chaotic single image of the flour particle cluster based on the corrected coordinates at the corresponding position;

[0064] S6, obtaining the humidity of the current flour particle mass based on the single view, and controlling the time and intensity of flour humidification based on the humidity of the current flour particle mass.

[0065] Embodiment 2

[0066] This embodiment provides a flour humidification system based on single-view reconstruction of a chaotic state of flour particles, including:

[0067] An image acquisition module, used to acquire multiple two-dimensional images of the chaotic state of flour particle clusters;

[0068] A depth information estimation module, used to input the multiple two-dimensional images into a pixel-to-pixel conversion network model, thereby solving the problem that the traditional depth estimation network cannot effectively estimate the depth information of the occluded area;

[0069] A point cloud densification processing module is used to establish a point cloud densification algorithm based on normal distribution and uniform distribution, analyze the morphological characteristics of the flour particle group based on the point cloud densification algorithm, and perform densification processing on the sparse point cloud using the normal and uniform distribution generation methods;

[0070] An error correction module is used to establish an improved error correction network, wherein the improved error correction network is used to correct errors caused by the pixel-to-pixel conversion network model and the point cloud dense algorithm, input the coordinates of the point cloud to the error correction network, and output the corrected coordinates at the corresponding position;

[0071] A single view reconstruction module is used to reconstruct a chaotic single view of a flour particle group based on the corrected coordinates at the corresponding position;

[0072] The flour humidification module is used to obtain the humidity of the current flour particle group based on the single view, and control the time and intensity of flour humidification based on the humidity of the current flour particle group.

[0073] Embodiment 3

[0074] A flour humidifier comprises a machine body, an upper cover, a motor, a mixing component, a spray component and an intelligent control component. An opening is formed on the top of the machine body, and the upper cover is detachably mounted and fixed on the top of the machine body and covers the opening of the machine body; the motor is mounted on the upper cover, and the output end of the motor drives and connects the upper end of the mixing component, so that the mixing component can be rotatably arranged in the inner cavity of the machine body; the spray component is mounted on the upper cover, and the output end of the spray component extends into the top of the machine body, water and gas can be input into the spray component, and water mist can be sprayed into the machine body by using the output end of the spray component; a feeding part connected to the inner cavity of the machine body is formed on the upper cover, and the material fed through the feeding part can be swung and dispersed by using the rotating mixing component, and the dispersed material and the water mist sprayed by the spray component are mixed under the stirring of the rotating mixing component; the intelligent control component is used to implement the control of the coordination of the mixing component and the spray component based on the method of the first aspect.

[0075] As a preferred embodiment, the mixing assembly includes a rotating drum, a plurality of stirring blades and a connecting shaft, the plurality of stirring blades are arranged on the outer wall of the rotating drum, and the stirring blades can be arranged to be inclined downward and upward, so that the stirring blades can be welded or fixed to the outer wall of the rotating drum at a certain angle, and the plurality of stirring blades are evenly distributed along the radial and axial directions of the rotating drum; the connecting shaft is fixed at the center of the upper surface of the vertical strips; the mixing assembly also includes a plurality of vertical strips and annular strips, the plurality of vertical strips are evenly fixed at the outer edge of the upper surface of the rotating drum, and the annular strips are respectively fixedly connected to the upper ends of the plurality of vertical strips, and a gap through which the material can pass is formed between two adjacent vertical strips.

[0076] As a preferred embodiment, the spray assembly includes a pipe body, a water inlet, an air inlet and a plurality of nozzles, wherein the water inlet and the air inlet are connected to the pipe body at one end facing the pipe body, and water and gas can be respectively input into the pipe body by using the water inlet and the air inlet, and the plurality of nozzles are connected and fixed to the bottom of the pipe body, and the plurality of nozzles are located above the mixing assembly.

[0077] As a preferred embodiment, a transmission belt is connected to the output end of the motor, and the other end of the transmission belt is connected to the upper end of the connecting shaft, so that the motor drives the rotation of the mixing assembly; the outer cover of the transmission belt is provided with the transmission belt protection cover; the top of the upper cover is formed with a C-shaped cover, and the C-shaped cover covers the tube body on its inner side; the top of the upper cover is also fixedly connected to an end sleeve, and the end sleeve covers the upper end of the connecting shaft on its inner side.

[0078] As a preferred embodiment, the machine body includes a hood, and the cross-section of the inner cavity of the hood is set to a polygon; an inspection port is formed on the outer wall of the hood and communicates with the inner cavity of the hood, and a plurality of threaded columns welded and fixed to the hood are arranged on the outside of the inspection port, and an inspection cover for covering the inspection port is provided, and the plurality of threaded columns pass through the inspection cover and are connected to nuts; the machine body also includes a cone bucket and a plurality of supporting legs, the cone bucket is fixed at the lower end of the hood and communicates with the inner cavity of the hood, a discharge port is formed at the bottom of the cone bucket, and a plurality of supporting legs are welded and fixed to the outer wall of the cone bucket.

[0079] See also Figure 1 and Figure 2 A flour humidifier includes a body 10, an upper cover 20, a motor 30, a mixing assembly 40 and a spray assembly 50. An opening is formed on the top of the body 10, and the upper cover 20 is detachably mounted and fixed on the top of the body 10 and covers the opening of the body 10. A feeding portion 21 communicating with the inner cavity of the body 10 is formed on the upper cover 20, and a discharge port is formed at the bottom of the body 10, so that a user can use the feeding portion 21 to add grain powder such as flour to the body 10, and the discharge port can be used to discharge the humidified powder. The motor 30 is mounted on the upper cover 20, and the output end of the motor 30 is driven to connect the upper end of the mixing assembly 40, so that the mixing assembly 40 is rotatably arranged in the inner cavity of the body 10. The spray assembly 50 is mounted on the upper cover 20, and the output end of the spray assembly 50 extends into the top of the body 10. Water and gas can be input into the spray assembly 50, and water mist can be sprayed into the body 10 by using the output end of the spray assembly 50. When humidifying grain powder such as flour, the material fed through the feeding part 21 can be shaken and dispersed by the rotating mixing assembly 40, and the dispersed material and the water mist sprayed by the spray assembly 50 can be mixed and contacted under the stirring of the rotating mixing assembly 40, thereby humidifying the powder.

[0080] See also Figure 3The machine body 10 may include a hood 11, an inspection cover 12, a cone bucket 13 and a plurality of supporting legs 14. The hood 11 may be made of a steel plate that is welded after being bent multiple times to form a hood body with a polygonal inner cavity. The inner cavity cross section of the hood 11 is set to a polygon, so that the powder that is swung and dispersed is in contact with the inner wall of the hood 11 and the mixing component 40 is stirred, so that the powder that is agglomerated or agglomerated is more easily broken. An inspection port is formed on the outer wall of the hood 11 and communicates with the inner cavity of the hood 11. By opening the inspection port (not shown in the figure), the user can easily clean the mixing component 40. A plurality of threaded columns are welded and fixed to the hood 11 on the outside of the inspection port, and an inspection cover 12 is provided at the inspection port to cover it. Under normal conditions, a plurality of threaded columns can be used to pass through the inspection cover 12 and be screwed and connected with a nut, so that the inspection cover 12 is fixed and blocks the inspection port, thereby preventing the powder from being thrown out during the processing process. The cone bucket 13 is fixed at the lower end of the hood 11 and communicates with the inner cavity of the hood 11. A discharge port is formed at the bottom of the cone bucket 13, and the humidified powder can be discharged through the discharge port. A plurality of support legs 14 are welded and fixed on the outer wall of the cone bucket 13, and the support legs 14 can support the entire device on the ground, and the discharge port has a certain distance from the ground.

[0081] See also Figure 4 A C-shaped cover 22 and an end sleeve 23 may also be fixed to the top of the upper cover 20. The C-shaped cover 22 is used to cover the pipe body in the spray assembly 50 on its inner side; the end sleeve 23 can be used to cover the upper end of the connecting shaft 45 in the mixing assembly 40 on its inner side.

[0082] See also Figure 4 The output end of the motor 30 may be connected with a transmission belt (not shown in the figure). The transmission belt is a common transmission structure in the prior art, which is generally composed of two transmission wheels and a transmission belt. The two transmission wheels may be connected to the output end of the motor 30 and the upper end of the connecting shaft 45 in the mixing assembly 40, respectively, so that the mixing assembly 40 can be driven to rotate as a whole. The outer side of the transmission belt may be covered with a transmission belt protection cover 31 to shield and protect it.

[0083] See also Figure 5The mixing assembly 40 may include a rotating drum 41, a plurality of stirring blades 42, a plurality of vertical slats 43, an annular slat 44, and a connecting shaft 45. The connecting shaft 45 is fixed at the center of the upper surface of the vertical slat 43 and is used to connect with the motor 30 so that the mixing assembly 40 can be driven to rotate. The plurality of vertical slats 43 are evenly distributed and fixed at the outer edge of the upper surface of the rotating drum 41, and the annular slat 44 is fixedly connected to the upper ends of the plurality of vertical slats 43, respectively, and a gap through which the material can pass is formed between two adjacent vertical slats 43, so that the powder can pass through the above gap and continue to fall downward from between the rotating drum 41 and the hood 11. A plurality of stirring blades 42 are fixedly connected to the outer wall of the rotating cylinder 41, and each stirring blade 42 can be welded or fixed to the outer wall of the rotating cylinder 41 by bolt connection at a certain angle (the angle between the stirring blades 42 and the horizontal plane is 0° to 90°, and does not include 0° and 90°) downward or tilted downward. The plurality of stirring blades 42 are evenly distributed along the radial and axial directions of the rotating cylinder 41, and the powder advancement speed (falling speed) can be changed by changing the distribution ratio of the two directions of the stirring blades 42. The powder fed through the feeding part 21 will fall on the upper surface of the rotating cylinder 41, and under the action of centrifugal force, the powder will rotate outward and disperse and fall downward through the gaps between adjacent vertical strips 43. The stirring blades 42 can stir the powder falling from the top of the rotating cylinder 41 and mix it evenly with the water mist, and make the powder stirred collide with the inner wall of the hood 11 to reduce agglomeration and caking.

[0084] See also Figure 4 The spray assembly 50 may include a tube body (not shown in the figure), a water inlet 51, an air inlet 52 and a plurality of nozzles 53. The water inlet 51 and the air inlet 52 are connected to the tube body at one end facing the tube body. The water inlet 51 and the air inlet 52 can be used to input water and gas into the tube body respectively, and the water and gas are mixed in the tube body. The plurality of nozzles 53 are connected and fixed at the bottom of the tube body, and the plurality of nozzles 53 are located above the mixing assembly 40. The nozzles 53 can be used to spray water mist downwards, and the water mist can be used to humidify the powder. An adjustment knob can be installed on the top of each nozzle 53, and rotating the adjustment knob can control the opening and closing of the spraying of the nozzle 53 and the flow rate.

[0085] In all the schemes mentioned above, the connection between two parts can be selected according to actual conditions by welding, bolt and nut matching connection, bolt or screw connection or other well-known connection methods, which will not be described one by one here. In the above, all fixed connections are preferably welded. Although the embodiments of the present invention have been shown and described, it can be understood by ordinary technicians in this field that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principle and spirit of the present invention. The scope of the present invention is defined by the attached claims and their equivalents.

Claims

1. A flour humidification method based on chaotic state single-view reconstruction of flour particle clusters, characterized in that: include: S1, obtaining multiple two-dimensional images of the chaotic state of flour particles; S2, inputting the multiple two-dimensional images into a pixel-to-pixel conversion network model, thereby solving the problem that the traditional depth estimation network cannot effectively estimate the depth information of the occluded area; S3, establishing a point cloud densification algorithm based on normal distribution and uniform distribution, analyzing the morphological characteristics of the flour particle group based on the point cloud densification algorithm, and densifying the sparse point cloud using the normal and uniform distribution generation methods; S4, establishing an improved error correction network, wherein the improved error correction network is used to correct errors caused by the pixel-to-pixel conversion network model and the point cloud dense algorithm, inputting the coordinates of the point cloud into the error correction network, and outputting the corrected coordinates at the corresponding position; S5, reconstructing a chaotic single image of the flour particle cluster based on the corrected coordinates at the corresponding position; S6, obtaining the humidity of the current flour particle mass based on the single view, and controlling the time and intensity of flour humidification based on the humidity of the current flour particle mass.

2. The flour humidification method based on chaotic state single view reconstruction of flour particle clusters according to claim 1 is characterized in that: The S2 includes: S21, building a pixel-to-pixel conversion network model; S22, hierarchically storing the three-dimensional information of the chaotic state of the flour particle cluster into the plurality of two-dimensional images based on a point cloud three-dimensional information hierarchical storage strategy; S23, based on the pixel-to-pixel conversion network model, cyclically predicting the depth information of the two-dimensional image that stores the three-dimensional information layer by layer, wherein the two-dimensional image depth information corresponds to the structural information of all occluded areas in the flour particle group; fusing the two-dimensional image depth information with the multiple two-dimensional images to obtain the morphological characteristics of the flour particle group.

3. The flour humidification method based on chaotic state single view reconstruction of flour particle clusters according to claim 2 is characterized in that: The S3 includes: S31, preprocessing and classifying the point cloud data based on the morphological characteristics of the flour particle group to obtain different types of points; S32, different densification methods and parameters are used for the different types of points. The different densification methods include a method based on spatial patch diffusion and a method based on contour constraints. Based on certain spatial geometric rules and constraints, the existing sparse point cloud information is densified through deduction and calculation.

4. The flour humidification method based on chaotic state single view reconstruction of flour particle clusters according to claim 3 is characterized in that: The improved error correction network consists of a convolutional layer, a batch normalization layer and a fully connected layer. A data vector of size 1*4 passes through three convolutional layers (Conv-64, Conv-128 and Conv-256 layers) in sequence, and a batch normalization layer is interspersed in the convolutional layer to prevent the risk of gradient disappearance; then it passes through four fully connected layers in sequence, and finally outputs a value of size 1*1 as the corrected y coordinate; multiple convolutional layers gradually extract deep-level features of point cloud data in a progressive manner; after a series of convolutional and batch normalization layer processing, the data is passed to the fully connected layer for further processing, so that the previously extracted features are globally integrated and mapped to the output space.

5. A flour humidification system based on single-view reconstruction of chaotic state of flour particle clusters, used to implement any method of claims 1-4, characterized in that: include: An image acquisition module, used to acquire multiple two-dimensional images of the chaotic state of flour particle clusters; A depth information estimation module, used to input the multiple two-dimensional images into a pixel-to-pixel conversion network model, thereby solving the problem that the traditional depth estimation network cannot effectively estimate the depth information of the occluded area; A point cloud densification processing module is used to establish a point cloud densification algorithm based on normal distribution and uniform distribution, analyze the morphological characteristics of the flour particle group based on the point cloud densification algorithm, and perform densification processing on the sparse point cloud using the normal and uniform distribution generation methods; An error correction module is used to establish an improved error correction network, wherein the improved error correction network is used to correct errors caused by the pixel-to-pixel conversion network model and the point cloud dense algorithm, input the coordinates of the point cloud to the error correction network, and output the corrected coordinates at the corresponding position; A single view reconstruction module is used to reconstruct a chaotic single view of a flour particle group based on the corrected coordinates at the corresponding position; The flour humidification module is used to obtain the humidity of the current flour particle group based on the single view, and control the time and intensity of flour humidification based on the humidity of the current flour particle group.

6. A flour humidifier, characterized in that: The invention comprises a machine body (10), an upper cover (20), a motor (30), a mixing assembly (40), a spray assembly (50) and an intelligent control assembly, wherein an opening is formed on the top of the machine body (10), and the upper cover (20) is detachably mounted and fixed on the top of the machine body (10) and covers the opening of the machine body (10); the motor (30) is mounted on the upper cover (20), and the output end of the motor (30) is driven to connect the upper end of the mixing assembly (40), so that the mixing assembly (40) is rotatably arranged in the inner cavity of the machine body (10); the spray assembly (50) is mounted on the upper cover (20), and the output end of the spray assembly (50) extends into the top of the machine body (10), water and gas can be input into the spray assembly (50), and water mist can be sprayed into the machine body (10) by using the output end of the spray assembly (50); The upper cover (20) is provided with a feeding portion (21) which is in communication with the inner cavity of the machine body (10); the material fed through the feeding portion (21) can be dispersed by shaking the rotating mixing component (40), and the dispersed material and the water mist sprayed by the spray component (50) are mixed under the stirring of the rotating mixing component (40); The intelligent control component is used to control the coordination of the mixing component and the spraying component by implementing any one of the methods described in claims 1-4.

7. A flour humidifier according to claim 6, characterized in that: The mixing assembly (40) comprises a rotating cylinder (41), a plurality of stirring blades (42) and a connecting shaft (45); the plurality of stirring blades (42) are fixedly connected to the outer wall of the rotating cylinder (41), and the plurality of stirring blades (42) are evenly distributed along the radial direction and the axial direction of the rotating cylinder (41); the connecting shaft (45) is fixed at the center of the upper surface of the vertical strip (43); each stirring blade (42) is welded or bolted to the outer wall of the rotating cylinder (41) at a certain angle upward or downward; the mixing assembly (40) also comprises a plurality of vertical strips (43), annular strips (44), and a plurality of The vertical slats (43) are evenly distributed and fixed at the outer edge of the upper surface of the rotating cylinder (41), and the annular slats (44) are respectively fixedly connected to the upper ends of the plurality of vertical slats (43), and a gap through which the material can pass is formed between two adjacent vertical slats (43); the machine body (10) also includes a cone bucket (13) and a plurality of supporting legs (14), the cone bucket (13) is fixed at the lower end of the machine cover (11) and communicates with the inner cavity of the machine cover (11), a discharge port is formed at the bottom of the cone bucket (13), and a plurality of supporting legs (14) are welded and fixed to the outer wall of the cone bucket (13).

8. A flour humidifier according to claim 7, characterized in that: A transmission belt is connected to the output end of the motor (30), and the other end of the transmission belt is connected to the upper end of the connecting shaft (45), so that the motor (30) drives the mixing assembly (40) to rotate; the outer cover of the transmission belt is provided with the transmission belt protection cover (31).

9. A flour humidifier according to claim 8, characterized in that: The spray assembly (50) comprises a pipe body, a water inlet (51), an air inlet (52) and a plurality of nozzles (53); the water inlet (51) and the air inlet (52) are connected to the pipe body at one end facing the pipe body; water and gas can be respectively input into the pipe body by means of the water inlet (51) and the air inlet (52); the plurality of nozzles (53) are connected and fixed to the bottom of the pipe body, and the plurality of nozzles (53) are located above the mixing assembly (40).

10. A flour humidifier according to claim 9, characterized in that: The machine body (10) comprises a hood (11), the inner cavity cross-section of the hood (11) being arranged in a polygonal shape; an inspection port is formed on the outer wall of the hood (11) and communicates with the inner cavity of the hood (11); a plurality of threaded columns welded and fixed to the hood (11) are arranged on the outer side of the inspection port, and an inspection cover (12) is arranged at the inspection port to cover the inspection port, and the plurality of threaded columns all pass through the inspection cover (12) and are connected to nuts.