CN115520786A - Aluminum material discharging auxiliary transfer equipment and transfer control method - Google Patents

Aluminum material discharging auxiliary transfer equipment and transfer control method Download PDF

Info

Publication number
CN115520786A
CN115520786A CN202211024805.0A CN202211024805A CN115520786A CN 115520786 A CN115520786 A CN 115520786A CN 202211024805 A CN202211024805 A CN 202211024805A CN 115520786 A CN115520786 A CN 115520786A
Authority
CN
China
Prior art keywords
matrix
sequence
transfer
gravity
gravity center
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.)
Granted
Application number
CN202211024805.0A
Other languages
Chinese (zh)
Other versions
CN115520786B (en
Inventor
刘建伟
冯大兵
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Jiangsu Guangkun Aluminum Co ltd
Original Assignee
Jiangsu Guangkun Aluminum Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Jiangsu Guangkun Aluminum Co ltd filed Critical Jiangsu Guangkun Aluminum Co ltd
Priority to CN202211024805.0A priority Critical patent/CN115520786B/en
Publication of CN115520786A publication Critical patent/CN115520786A/en
Application granted granted Critical
Publication of CN115520786B publication Critical patent/CN115520786B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B66—HOISTING; LIFTING; HAULING
    • B66C—CRANES; LOAD-ENGAGING ELEMENTS OR DEVICES FOR CRANES, CAPSTANS, WINCHES, OR TACKLES
    • B66C13/00—Other constructional features or details
    • B66C13/18—Control systems or devices
    • B66C13/48—Automatic control of crane drives for producing a single or repeated working cycle; Programme control
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B65—CONVEYING; PACKING; STORING; HANDLING THIN OR FILAMENTARY MATERIAL
    • B65G—TRANSPORT OR STORAGE DEVICES, e.g. CONVEYORS FOR LOADING OR TIPPING, SHOP CONVEYOR SYSTEMS OR PNEUMATIC TUBE CONVEYORS
    • B65G47/00—Article or material-handling devices associated with conveyors; Methods employing such devices
    • B65G47/74—Feeding, transfer, or discharging devices of particular kinds or types
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B66—HOISTING; LIFTING; HAULING
    • B66C—CRANES; LOAD-ENGAGING ELEMENTS OR DEVICES FOR CRANES, CAPSTANS, WINCHES, OR TACKLES
    • B66C15/00—Safety gear
    • B66C15/06—Arrangements or use of warning devices
    • B66C15/065—Arrangements or use of warning devices electrical
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/60—Analysis of geometric attributes
    • G06T7/66—Analysis of geometric attributes of image moments or centre of gravity
    • 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
    • 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
    • G06V20/40—Scenes; Scene-specific elements in video content
    • G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
    • G—PHYSICS
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
    • G08B21/18—Status alarms
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10016—Video; Image sequence
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Computation (AREA)
  • Software Systems (AREA)
  • Multimedia (AREA)
  • Mechanical Engineering (AREA)
  • Databases & Information Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Computing Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Automation & Control Theory (AREA)
  • Emergency Management (AREA)
  • Business, Economics & Management (AREA)
  • Computational Linguistics (AREA)
  • Geometry (AREA)
  • Image Analysis (AREA)

Abstract

The application relates to the technical field of aluminum material transfer, and particularly discloses an aluminum material discharge auxiliary transfer device and a transfer control method, wherein a sequence of first gravity center data of a fixing mechanism and a sequence of second gravity center data of an aluminum material are extracted from a transfer monitoring video collected by a monitoring camera, a difference value between the gravity center data of each corresponding time point in the sequence of the first gravity center data and the sequence of the second gravity center data is calculated to obtain a sequence of gravity center difference values, then a convolutional neural network model is used for carrying out appropriate coding on the sequence of the first gravity center data, the sequence of the second gravity center data and the sequence of the gravity center difference values, and a classifier is used for decoding to obtain a classification result for representing whether a drop risk early warning prompt is generated or not.

Description

Aluminum product discharge auxiliary transfer equipment and transfer control method
Technical Field
The application relates to the technical field of aluminum product transfer, in particular to an aluminum product discharge auxiliary transfer device and a transfer control method.
Background
Articles made of aluminum and other alloying elements. Usually, the steel is processed into casting products, forging products, foils, plates, strips, pipes, bars, section bars and the like, and then the steel is manufactured by the processes of cold bending, saw cutting, drilling, assembling, coloring and the like. The main metal element is aluminum, and some alloy elements are added, so that the performance of the aluminum material is improved, and the aluminum material needs to be transported in the production process of the aluminum material.
Need fix when current aluminum product is transported, current transfer device's fixed establishment need be fixed and dismantlement through the manual work, and is more troublesome, and because it is artifical fixed, fixed dynamics is unchangeable, and in order to prevent that the aluminum product from being out of shape, general fixed dynamics can not be too big, when the trend of landing appears, can not increase fixed dynamics voluntarily for the aluminum product drops, causes the damage on aluminum product and ground, probably causes the damage and the casualties of machine even.
Therefore, an aluminum material production discharge transfer device with a transfer stability detection function is expected to generate a warning signal when the risk of dropping is detected.
Disclosure of Invention
The present application is proposed to solve the above-mentioned technical problems. The embodiment of the application provides an aluminum product discharging auxiliary transfer device and a transfer control method, wherein a sequence of first gravity center data of a fixing mechanism and a sequence of second gravity center data of an aluminum product are extracted from a transfer monitoring video collected by a monitoring camera, the difference between the gravity center data of each corresponding time point in the sequence of the first gravity center data and the sequence of the second gravity center data is calculated to obtain a sequence of gravity center difference values, then a convolutional neural network model is used for properly coding the sequence of the first gravity center data, the sequence of the second gravity center data and the sequence of the gravity center difference values, and the classification result is decoded by a classifier to obtain a classification result for representing whether an early warning prompt for the existence of a falling risk is generated.
According to one aspect of the application, an aluminum product discharge auxiliary transfer device is provided, which comprises: the transfer monitoring video acquisition module is used for acquiring a transfer monitoring video acquired by the monitoring camera; the gravity center data extraction module is used for extracting a sequence of first gravity center data of a fixing mechanism and a sequence of second gravity center data of an aluminum material from the transfer monitoring video, wherein the fixing mechanism is used for fixing the aluminum material; a center of gravity difference module, configured to calculate a difference between center of gravity data at each corresponding time point in the sequence of the first center of gravity data and the sequence of the second center of gravity data to obtain a sequence of center of gravity differences; the gravity center change feature extraction module is used for enabling the sequence of the gravity center difference value to pass through a time sequence encoder comprising a one-dimensional convolution layer so as to obtain a first feature vector; an absolute gravity center feature extraction module, configured to pass the sequence of the first gravity center data and the sequence of the second gravity center data through the time-series encoder including the one-dimensional convolutional layer, respectively, to obtain a second feature vector and a third feature vector; a first transfer module to calculate a first transfer matrix of the first eigenvector relative to the second eigenvector; a second transfer module for calculating a second transfer matrix of the first eigenvector relative to the third eigenvector; the characteristic correction module is used for respectively correcting the characteristic values of the first transfer matrix and the second transfer matrix based on the distance between corresponding positions in the first transfer matrix and the second transfer matrix so as to obtain a corrected first transfer matrix and a corrected second transfer matrix; the information fusion module is used for fusing the corrected first transfer matrix and the corrected second transfer matrix to obtain a classification characteristic matrix; and the transfer risk prompting module is used for enabling the classification characteristic matrix to pass through a classifier to obtain a classification result, and the classification result is used for indicating whether an early warning prompt for the existence of the falling risk is generated.
According to another aspect of the application, an aluminum material discharge auxiliary transfer control method is provided, which includes: acquiring a transfer monitoring video acquired by a monitoring camera; extracting a sequence of first gravity center data of a fixing mechanism and a sequence of second gravity center data of an aluminum material from the transfer monitoring video, wherein the fixing mechanism is used for fixing the aluminum material; calculating differences between the gravity center data of each corresponding time point in the sequence of the first gravity center data and the sequence of the second gravity center data to obtain a sequence of gravity center differences; passing the sequence of gravity center difference values through a time sequence encoder comprising a one-dimensional convolutional layer to obtain a first feature vector; respectively passing the sequence of the first gravity center data and the sequence of the second gravity center data through the time sequence encoder containing the one-dimensional convolutional layer to obtain a second feature vector and a third feature vector; calculating a first transfer matrix of the first eigenvector relative to the second eigenvector; calculating a second transfer matrix of the first eigenvector relative to the third eigenvector; respectively correcting the eigenvalues of the first transfer matrix and the second transfer matrix based on the distance between corresponding positions in the first transfer matrix and the second transfer matrix to obtain a corrected first transfer matrix and a corrected second transfer matrix; fusing the corrected first transfer matrix and the corrected second transfer matrix to obtain a classification feature matrix; and enabling the classification characteristic matrix to pass through a classifier to obtain a classification result, wherein the classification result is used for indicating whether an early warning prompt for existence of a falling risk is generated.
Compared with the prior art, the aluminum product discharging auxiliary transfer equipment and the transfer control method have the advantages that the sequence of the first gravity center data of the fixing mechanism and the sequence of the second gravity center data of the aluminum product are extracted from the transfer monitoring video collected by the monitoring camera, the difference value between the gravity center data of each corresponding time point in the sequence of the first gravity center data and the sequence of the second gravity center data is calculated to obtain the sequence of the gravity center difference value, then the sequence of the first gravity center data, the sequence of the second gravity center data and the sequence of the gravity center difference value are properly coded by using the convolutional neural network model, and the classification result of the early warning prompt for indicating whether the falling risk exists is obtained through decoding by the classifier, so that the early warning can be accurately carried out when the falling risk exists in the aluminum product, the aluminum product is prevented from falling in the transfer process, and the normal transfer and processing of the aluminum product are guaranteed.
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 an application scenario of an aluminum material discharge auxiliary transfer device according to an embodiment of the application.
Fig. 2 illustrates a block diagram schematic diagram of an aluminum material discharge auxiliary transfer device according to an embodiment of the application.
Fig. 3 illustrates a block diagram of a gravity center change feature extraction module in an aluminum material discharge auxiliary transfer device according to an embodiment of the application.
Fig. 4 illustrates a block diagram of an absolute gravity center feature extraction module in an aluminum material discharge auxiliary transfer device according to an embodiment of the application.
Fig. 5 illustrates a flow chart of an aluminum material discharge auxiliary transfer control method according to an embodiment of the present application.
Fig. 6 illustrates a schematic diagram of a system architecture of an aluminum material discharge-assisted transfer control method according to 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.
Overview of a scene
As described above, the inventor of the present application considers that when transferring an aluminum material, since the fixing force of the fixing device to the aluminum material is not too large to prevent the aluminum material from deforming, which may easily cause the aluminum material to drop, it is necessary to monitor the stability of transferring the aluminum material in real time to prevent the aluminum material from dropping when transferring the production discharge of the aluminum material. The inventor of the present application has also found that, in the transfer process of an aluminum material, when a positional deviation of the aluminum material occurs, causing a risk of dropping thereof, a movement characteristic of the aluminum material in a time dimension or a movement characteristic of the transfer mechanism in a time dimension may change, and a relative positional relationship between the aluminum material and the transfer mechanism may also change. Therefore, in the technical scheme of this application, expect through synthesizing the aluminum product with the relative motion information characteristic of transport mechanism, the absolute motion information characteristic of transport mechanism and the absolute operation information characteristic of aluminum product carry out the transportation stability of aluminum product is judged, in order to guarantee the safe transportation of aluminum product.
Specifically, in the technical scheme of this application, firstly, consider that the aluminum product with the motion state characteristic of transport mechanism can all be drawed through surveillance video, consequently gather the surveillance video that the aluminum product transported through surveillance camera head. It should be understood that for the monitoring video of the aluminum material transfer, in order to better characterize the motion characteristics of the aluminum material and the fixing mechanism in the transfer mechanism, the respective gravity centers are selected as reference points to conduct data motion characteristic mining. Specifically, the center of gravity data of the fixing mechanism and the aluminum material at each time point are further extracted from the transfer monitoring video to obtain a sequence of first center of gravity data of the fixing mechanism and a sequence of second center of gravity data of the aluminum material, wherein the fixing mechanism is used for fixing the aluminum material.
Then, in order to describe the dynamic characteristic information of the relative position operational relationship of the fixing mechanism and the aluminum material within a predetermined time period of the surveillance video, first, a difference between the center of gravity data of each corresponding time point in the sequence of the first center of gravity data and the sequence of the second center of gravity data is further calculated to obtain a sequence of center of gravity differences.
It should be understood that, since the fixing mechanism and the aluminum material have a dynamic motion law within a predetermined time period of the surveillance video, in order to be able to sufficiently extract an absolute motion implicit associated feature of the fixing mechanism, an absolute motion implicit associated feature of the aluminum material, and a relative motion implicit associated feature of the aluminum material and the fixing mechanism in a time dimension, a time-series encoder including a one-dimensional convolution layer is further used to encode the sequence of the gravity center difference values, the sequence of the first gravity center data, and the sequence of the second gravity center data, respectively, so as to be able to extract a dynamic change feature of the fixing mechanism, the motion feature of the aluminum material in the time-series dimension, and a dynamic change feature of the relative motion feature of the aluminum material and the fixing mechanism in the time-series dimension, respectively, so as to obtain a first feature vector, a second feature vector, and a third feature vector.
Further, considering that since there is a correlation between the absolute movement characteristic of the fixing mechanism and the absolute operation characteristic of the aluminum material and the relative movement characteristic of the aluminum material and the fixing mechanism, for example, when the relative movement characteristic of the aluminum material and the fixing mechanism changes, the absolute movement characteristic of the fixing mechanism or the absolute operation characteristic of the aluminum material also changes, if the fusion of the characteristics is performed only by a simple addition method, the characteristic distribution flow shape of the fused characteristics in the high-dimensional characteristic space becomes very complicated and irregular. Thus, a first transfer matrix of the first feature vector relative to the second feature vector is further calculated, and a second transfer matrix of the first feature vector relative to the third feature vector is calculated.
In particular, in the technical solution of the present application, since the first eigenvector is a time-series associated feature of the gravity center difference sequence, a first transition matrix of the first eigenvector relative to the second eigenvector and a second transition matrix of the first eigenvector relative to the third eigenvector are constrained in different directions on feature distribution, that is, the first transition matrix and the second transition matrix have anisotropy, which is embodied in that high-dimensional feature expressions of the first transition matrix and the second transition matrix respectively reside in a narrow subset of the whole high-dimensional feature space, so that when the first transition matrix and the second transition matrix are fused for classification, the lack of continuity of the fused feature expressions may cause degradation of the solution space of the classification problem, thereby affecting the classification effect.
Therefore, preferably before fusing the first transition matrix and the second transition matrix, the first transition matrix is optimized in the same direction, specifically:
Figure 772536DEST_PATH_IMAGE001
wherein
Figure 299332DEST_PATH_IMAGE002
And
Figure 764949DEST_PATH_IMAGE003
the first transition matrix and the second transition matrix respectively,
Figure 351788DEST_PATH_IMAGE004
and
Figure 702522DEST_PATH_IMAGE005
respectively is that
Figure 501851DEST_PATH_IMAGE006
A characteristic value of the position, and
Figure 353132DEST_PATH_IMAGE007
specifically, the distance between the first transition matrix and the second transition matrix may be set as the hyper-parameter.
Here, by performing the equidirectional optimization of the first transition matrix and the second transition matrix based on the comparison search space, the optimized first transition matrix and the optimized second transition matrix can be mapped to the isotropic and differentiated expression space, which improves the continuity of the fused feature expression, thereby suppressing the solution space degradation of the classification problem and improving the classification effect. Like this, can improve right the accuracy that there is the risk early warning that drops in the aluminum product is avoided the aluminum product drops in the transportation, guarantees the normal transportation and the processing of aluminum product.
And then, after the corrected first transfer matrix and the corrected second transfer matrix are obtained, further fusing the corrected first transfer matrix and the corrected second transfer matrix for classification so as to obtain a classification result of an early warning prompt for indicating whether the falling risk exists or not.
Based on this, the application provides an auxiliary transfer equipment of aluminum product ejection of compact, it includes: the transfer monitoring video acquisition module is used for acquiring a transfer monitoring video acquired by the monitoring camera; the gravity center data extraction module is used for extracting a sequence of first gravity center data of a fixing mechanism and a sequence of second gravity center data of an aluminum material from the transfer monitoring video, wherein the fixing mechanism is used for fixing the aluminum material; a center of gravity difference module, configured to calculate a difference between center of gravity data at each corresponding time point in the sequence of the first center of gravity data and the sequence of the second center of gravity data to obtain a sequence of center of gravity differences; the gravity center change feature extraction module is used for enabling the sequence of the gravity center difference value to pass through a time sequence encoder comprising a one-dimensional convolution layer so as to obtain a first feature vector; an absolute gravity center feature extraction module, configured to pass the sequence of the first gravity center data and the sequence of the second gravity center data through the time-series encoder including the one-dimensional convolutional layer, respectively, to obtain a second feature vector and a third feature vector; a first transfer module to calculate a first transfer matrix of the first eigenvector relative to the second eigenvector; a second transfer module for calculating a second transfer matrix of the first eigenvector relative to the third eigenvector; the characteristic correction module is used for respectively correcting the characteristic values of the first transfer matrix and the second transfer matrix based on the distance between corresponding positions in the first transfer matrix and the second transfer matrix so as to obtain a corrected first transfer matrix and a corrected second transfer matrix; the information fusion module is used for fusing the corrected first transfer matrix and the corrected second transfer matrix to obtain a classification characteristic matrix; and the transfer risk prompting module is used for enabling the classification characteristic matrix to pass through a classifier to obtain a classification result, and the classification result is used for indicating whether an early warning prompt for existence of a dropping risk is generated.
Fig. 1 illustrates an application scene diagram of an aluminum material discharge auxiliary transfer device according to an embodiment of the application. As shown in fig. 1, in this application scenario, first, a transfer monitoring video of an aluminum material (e.g., M as illustrated in fig. 1) is acquired by a monitoring camera (e.g., C as illustrated in fig. 1) disposed beside an aluminum material discharge auxiliary transfer device having a fixing mechanism (e.g., G as illustrated in fig. 1) for fixing the aluminum material. Then, the collected transportation monitoring video is input into a server (for example, S illustrated in fig. 1) deployed with an aluminum product discharge auxiliary transportation control algorithm, wherein the server can process the transportation monitoring video by using the aluminum product discharge auxiliary transportation control algorithm to generate an early warning result for indicating whether an early warning prompt indicating that a risk of dropping exists is generated.
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 System
Fig. 2 illustrates a block diagram schematic diagram of an aluminum material discharge auxiliary transfer device according to an embodiment of the application. As shown in fig. 2, the aluminum material discharge auxiliary transfer apparatus 100 according to the embodiment of the present application includes: a transfer surveillance video acquisition module 110, configured to acquire a transfer surveillance video acquired by a surveillance camera; a center of gravity data extraction module 120, configured to extract, from the transfer surveillance video, a sequence of first center of gravity data of a fixing mechanism and a sequence of second center of gravity data of an aluminum material, where the fixing mechanism is used to fix the aluminum material; a center of gravity difference module 130, configured to calculate a difference between the center of gravity data at each corresponding time point in the sequence of the first center of gravity data and the sequence of the second center of gravity data to obtain a sequence of center of gravity differences; a gravity center change feature extraction module 140, configured to pass the sequence of gravity center difference values through a time-series encoder including a one-dimensional convolutional layer to obtain a first feature vector; an absolute barycentric feature extraction module 150, configured to pass the sequence of the first barycentric data and the sequence of the second barycentric data through the time-series encoder including the one-dimensional convolutional layer, respectively, to obtain a second feature vector and a third feature vector; a first transfer module 160 for calculating a first transfer matrix of the first eigenvector relative to the second eigenvector; a second transfer module 170 for calculating a second transfer matrix of the first eigenvector relative to the third eigenvector; a feature correction module 180, configured to perform feature value correction on the first transfer matrix and the second transfer matrix respectively based on a distance between corresponding positions in the first transfer matrix and the second transfer matrix to obtain a corrected first transfer matrix and a corrected second transfer matrix; an information fusion module 190, configured to fuse the corrected first transfer matrix and the corrected second transfer matrix to obtain a classification feature matrix; and a transfer risk prompting module 200, configured to pass the classification feature matrix through a classifier to obtain a classification result, where the classification result is used to indicate whether an early warning prompt indicating that a risk of dropping exists is generated.
In this embodiment of the application, the transfer surveillance video capturing module 110 is configured to obtain a transfer surveillance video captured by a surveillance camera. As described above, the inventor of the present application considers that when transferring an aluminum material, since the fixing force of the fixing device on the aluminum material is not too large to prevent the aluminum material from deforming, which may easily cause the aluminum material to drop, it is necessary to monitor the stability of transferring the aluminum material in real time to prevent the aluminum material from dropping when transferring the production discharging of the aluminum material. The inventor of the present application has also found that, in the transfer process of an aluminum material, when a positional deviation of the aluminum material occurs, causing a risk of dropping thereof, a movement characteristic of the aluminum material in a time dimension or a movement characteristic of the transfer mechanism in a time dimension may change, and a relative positional relationship between the aluminum material and the transfer mechanism may also change. Therefore, in the technical solution of the present application, it is desirable to determine the transfer stability of the aluminum material by integrating the aluminum material and the relative motion information characteristic of the transfer mechanism, the absolute motion information characteristic of the transfer mechanism, and the absolute operation information characteristic of the aluminum material, so as to ensure the safe transfer of the aluminum material.
In the technical scheme of this application, consider the aluminum product with the motion state characteristic of transport mechanism can all extract through surveillance video, consequently gathers the surveillance video that the aluminum product transported through surveillance camera head.
In a specific embodiment of the application, a monitoring camera arranged beside an aluminum material discharge auxiliary transfer device is used for collecting a transfer monitoring video of an aluminum material, wherein the aluminum material discharge auxiliary transfer device is provided with a fixing mechanism for fixing the aluminum material.
In this embodiment, the center of gravity data extracting module 120 is configured to extract a sequence of first center of gravity data of a fixing mechanism and a sequence of second center of gravity data of an aluminum material from the transfer surveillance video, where the fixing mechanism is configured to fix the aluminum material. It should be understood that for the monitoring video of the aluminum material transfer, in order to better characterize the motion characteristics of the aluminum material and the fixing mechanism in the transfer mechanism, the respective gravity centers are selected as reference points to conduct data motion characteristic mining. That is, specifically, the center of gravity data of the fixing mechanism and the aluminum material at each time point is further extracted from the transfer surveillance video to obtain a sequence of first center of gravity data of the fixing mechanism and a sequence of second center of gravity data of the aluminum material, where the fixing mechanism is used for fixing the aluminum material.
In an embodiment of the present application, the center of gravity difference module 130 is configured to calculate a difference between the center of gravity data at each corresponding time point in the sequence of the first center of gravity data and the sequence of the second center of gravity data to obtain a sequence of center of gravity differences. It should be understood that, in order to describe the dynamic characteristic information of the relative positional operational relationship of the fixing mechanism and the aluminum material within the predetermined time period of the surveillance video, the difference between the gravity center data of each corresponding time point in the sequence of the first gravity center data and the sequence of the second gravity center data is further calculated to obtain the sequence of the gravity center differences.
In a specific embodiment of the present application, the center of gravity difference module 130 is further configured to: calculating a difference between the gravity center data of each corresponding time point in the sequence of the first gravity center data and the sequence of the second gravity center data by the following formula to obtain a sequence of the gravity center differences; wherein the formula is:
Figure 110873DEST_PATH_IMAGE008
wherein, the first and the second end of the pipe are connected with each other,
Figure 414815DEST_PATH_IMAGE009
a sequence representing the first center of gravity data,
Figure 17835DEST_PATH_IMAGE010
a sequence representing the second center of gravity data,
Figure 992131DEST_PATH_IMAGE011
a sequence representing the difference in the center of gravity,
Figure 920773DEST_PATH_IMAGE012
representing a position-wise subtraction between the barycentric data of respective corresponding time points in the barycentric sequence.
In this embodiment, the center of gravity change feature extraction module 140 and the absolute center of gravity feature extraction module 150 are configured to pass the sequence of center of gravity difference values through a time-series encoder including one-dimensional convolutional layers to obtain a first feature vector, and pass the sequence of the first center of gravity data and the sequence of the second center of gravity data through the time-series encoder including one-dimensional convolutional layers to obtain a second feature vector and a third feature vector, respectively. It should be understood that, since the fixing mechanism and the aluminum material have a dynamic motion law within a predetermined time period of the surveillance video, in order to be able to sufficiently extract an absolute motion implicit associated feature of the fixing mechanism, an absolute motion implicit associated feature of the aluminum material, and a relative motion implicit associated feature of the aluminum material and the fixing mechanism in a time dimension, a time-series encoder including a one-dimensional convolution layer is further used to encode the sequence of the gravity center difference values, the sequence of the first gravity center data, and the sequence of the second gravity center data, respectively, so as to be able to extract a dynamic change feature of the fixing mechanism, the motion feature of the aluminum material, and a dynamic change feature of the relative motion feature of the aluminum material and the fixing mechanism in the time-series dimension, respectively, so as to obtain the first feature vector, the second feature vector, and the third feature vector.
Fig. 3 illustrates a block diagram of a gravity center change feature extraction module in an aluminum material discharge auxiliary transfer device according to an embodiment of the application. As shown in fig. 3, in a specific embodiment of the present application, the center of gravity change feature extraction module 140 includes: an arrangement unit 141 configured to arrange the sequence of the gravity center difference values into a gravity center difference input vector; a full-concatenation unit 142, configured to perform full-concatenation coding on the barycentric difference input vector using a full-concatenation layer of the time sequence encoder according to the following formula to extract high-dimensional implicit features of feature values of each position in the barycentric difference input vector, where the formula is:
Figure 977591DEST_PATH_IMAGE013
in which
Figure 384301DEST_PATH_IMAGE014
Is the center of gravity difference input vector,
Figure 475754DEST_PATH_IMAGE015
is the output vector of the digital video signal,
Figure 309718DEST_PATH_IMAGE016
is a matrix of weights that is a function of,
Figure 856762DEST_PATH_IMAGE017
is a vector of the offset to the offset,
Figure 67163DEST_PATH_IMAGE018
represents a matrix multiplication; and a one-dimensional convolution unit 143, configured to perform one-dimensional convolution encoding on the barycentric difference input vector using a one-dimensional convolution layer of the time-series encoder according to the following formula to extract high-dimensional implicit correlation features between feature values of each position in the barycentric difference input vector, where the formula is:
Figure 481964DEST_PATH_IMAGE019
wherein, the first and the second end of the pipe are connected with each other,ais a convolution kernelxA width in the direction,
Figure 486829DEST_PATH_IMAGE020
Is a convolution kernel parameter vector,
Figure 987080DEST_PATH_IMAGE021
Is a matrix of local vectors operating with a convolution kernel,wis the size of the convolution kernel and,
Figure 735594DEST_PATH_IMAGE022
representing the barycentric difference input vector.
Fig. 4 illustrates a block diagram of an absolute gravity center feature extraction module in an aluminum material discharge auxiliary transfer device according to an embodiment of the application. In a specific embodiment of the present application, the absolute gravity center feature extraction module 150 includes: a barycentric data arrangement unit 151 for arranging the sequence of the first barycentric data and the sequence of the second barycentric data as a first barycentric input vector and a second barycentric input vector, respectively; a fully-concatenated encoding unit 152 for fully concatenating the first and second centroid input vectors using a fully-concatenated layer of the sequential encoder in the following formulaAnd coding to respectively extract high-dimensional implicit features of feature values of all positions in the first gravity center input vector and the second gravity center input vector, wherein the formula is as follows:
Figure 801639DEST_PATH_IMAGE023
wherein
Figure 992053DEST_PATH_IMAGE024
Is the first and second centroid input vectors,
Figure 245180DEST_PATH_IMAGE025
is the output vector of the output vector,
Figure 797384DEST_PATH_IMAGE026
is a matrix of the weights that is,
Figure 452357DEST_PATH_IMAGE027
is a vector of the offset to the offset,
Figure 799024DEST_PATH_IMAGE028
represents a matrix multiplication; and a one-dimensional convolution encoding unit 153 configured to perform one-dimensional convolution encoding on the first gravity center input vector and the second gravity center input vector using the one-dimensional convolution layer of the time-series encoder according to the following formula to extract high-dimensional implicit correlation features between feature values of respective positions in the first gravity center input vector and the second gravity center input vector, respectively, where the formula is:
Figure 805026DEST_PATH_IMAGE019
wherein, the first and the second end of the pipe are connected with each other,ais a convolution kernelxA width in the direction,
Figure 895342DEST_PATH_IMAGE029
Is a convolution kernel parameter vector,
Figure 407751DEST_PATH_IMAGE021
Is a matrix of local vectors operating with a convolution kernel,wis the size of the convolution kernel and,
Figure 925320DEST_PATH_IMAGE030
representing the first and second barycentric input vectors.
In an embodiment of the present application, the first transfer module 160 and the second transfer module 170 are configured to calculate a first transfer matrix of the first eigenvector relative to the second eigenvector, and calculate a second transfer matrix of the first eigenvector relative to the third eigenvector. It should be understood that, considering that the absolute movement characteristic of the fixing mechanism and the absolute operation characteristic of the aluminum material are related to the relative movement characteristic of the aluminum material and the fixing mechanism, for example, when the relative movement characteristic of the aluminum material and the fixing mechanism changes, the absolute movement characteristic of the fixing mechanism or the absolute operation characteristic of the aluminum material also changes, if the fusion of the characteristics is performed by a simple addition method, the characteristic distribution flow shape of the fused characteristics in a high-dimensional characteristic space becomes very complicated and irregular. Thus, a first transfer matrix of the first feature vector relative to the second feature vector is further calculated, and a second transfer matrix of the first feature vector relative to the third feature vector is calculated.
In a specific embodiment of the present application, the first transfer module 160 is further configured to: calculating the first transfer matrix of the first feature vector relative to the second feature vector in the following formula; wherein the formula is:
Figure 153039DEST_PATH_IMAGE031
=
Figure 781466DEST_PATH_IMAGE032
*
Figure 145451DEST_PATH_IMAGE033
wherein
Figure 833922DEST_PATH_IMAGE034
Representing the first feature vector in a first set of features,
Figure 283357DEST_PATH_IMAGE035
-representing the second feature vector by means of a second feature vector,
Figure 718405DEST_PATH_IMAGE036
representing the first transition matrix; the second transfer module 170 is further configured to: calculating the second transfer matrix of the first feature vector relative to the third feature vector with the following formula; wherein the formula is:
Figure 936897DEST_PATH_IMAGE037
=
Figure 61848DEST_PATH_IMAGE038
*
Figure 733001DEST_PATH_IMAGE039
wherein
Figure 234389DEST_PATH_IMAGE034
Representing the first feature vector in a first set of features,
Figure 307388DEST_PATH_IMAGE040
representing the third feature vector and the second feature vector,
Figure 603240DEST_PATH_IMAGE041
representing the second transition matrix.
In particular, in the technical solution of the present application, since the first eigenvector is a time-series associated feature of the gravity center difference sequence, a first transition matrix of the first eigenvector relative to the second eigenvector and a second transition matrix of the first eigenvector relative to the third eigenvector are constrained in different directions on feature distribution, that is, the first transition matrix and the second transition matrix have anisotropy, which is embodied in that high-dimensional feature expressions of the first transition matrix and the second transition matrix respectively reside in a narrow subset of the whole high-dimensional feature space, so that when the first transition matrix and the second transition matrix are fused for classification, the lack of continuity of the fused feature expressions may cause degradation of the solution space of the classification problem, thereby affecting the classification effect. Therefore, it is preferable to first optimize the first transition matrix and the second transition matrix in the same direction before fusing them.
In this embodiment, the characteristic correction module 180 is configured to perform characteristic value correction on the first transition matrix and the second transition matrix respectively based on a distance between corresponding positions in the first transition matrix and the second transition matrix to obtain a corrected first transition matrix and a corrected second transition matrix.
In a specific embodiment of the present application, the feature correction module includes: a first correction unit, configured to perform eigenvalue correction on the first transfer matrix based on a distance between corresponding positions in the first transfer matrix and the second transfer matrix to obtain a corrected first transfer matrix; wherein the formula is:
Figure 30197DEST_PATH_IMAGE042
wherein the content of the first and second substances,
Figure 804118DEST_PATH_IMAGE043
representing the second transition matrix in a second order,
Figure 997202DEST_PATH_IMAGE044
and
Figure 198377DEST_PATH_IMAGE045
respectively represent the first transition matrix and the second transition matrix
Figure 844122DEST_PATH_IMAGE046
The value of the characteristic of the location is,
Figure 421733DEST_PATH_IMAGE047
representing the first transition matrix and the second transition matrix
Figure 469324DEST_PATH_IMAGE048
A distance between the respective positions, and
Figure 98189DEST_PATH_IMAGE049
is a hyper-parameter;
a second correction unit, configured to perform eigenvalue correction on the second transfer matrix based on a distance between corresponding positions in the first transfer matrix and the second transfer matrix to obtain a corrected second transfer matrix; wherein the formula is:
Figure 496810DEST_PATH_IMAGE050
wherein the content of the first and second substances,
Figure 612533DEST_PATH_IMAGE051
respectively, represent the first transfer matrix and,
Figure 514630DEST_PATH_IMAGE052
and
Figure 323186DEST_PATH_IMAGE053
respectively represent the first transition matrix and the second transition matrix
Figure 677944DEST_PATH_IMAGE054
The value of the characteristic of the location is,
Figure 597359DEST_PATH_IMAGE055
representing the first transition matrix and the second transition matrix
Figure 888050DEST_PATH_IMAGE056
A distance between the respective positions, and
Figure 601928DEST_PATH_IMAGE057
is a hyper-parameter.
Here, by performing the equidirectional optimization of the first transition matrix and the second transition matrix based on the comparison search space, the optimized first transition matrix and the optimized second transition matrix can be mapped to the isotropic and differentiated expression space, which improves the continuity of the fused feature expression, thereby suppressing the solution space degradation of the classification problem and improving the classification effect. Like this, can improve right the accuracy that there is the risk early warning that drops in the aluminum product is avoided the aluminum product drops in the transportation, guarantees the normal transportation and the processing of aluminum product.
In this embodiment of the application, the information fusion module 190 is configured to fuse the corrected first transfer matrix and the corrected second transfer matrix to obtain a classification feature matrix. Namely, the aluminum material and the relative motion information characteristic of the transfer mechanism, the absolute motion information characteristic of the transfer mechanism and the absolute motion information characteristic of the aluminum material are fused to improve the characterization capability of the classification characteristic matrix, so that the accuracy of the transfer stability judgment of the aluminum material by equipment is improved.
In a specific embodiment of the present application, the information fusion module is further configured to: fusing the corrected first transfer matrix and the corrected second transfer matrix according to the following formula to obtain the classification feature matrix; wherein the formula is:
Figure 709562DEST_PATH_IMAGE058
wherein, in the process,
Figure 167088DEST_PATH_IMAGE059
for the purpose of the classification feature matrix,
Figure 778198DEST_PATH_IMAGE060
for the corrected first transfer matrix to be used,
Figure 928556DEST_PATH_IMAGE061
for the said corrected second transfer matrix,
Figure 257906DEST_PATH_IMAGE062
for the weighting parameters used for controlling the balance between the corrected first transition matrix and the corrected second transition matrix in the classification feature matrix,
Figure 253544DEST_PATH_IMAGE063
representing a position-wise addition of the matrix.
In this embodiment, the transfer risk prompting module 200 is configured to pass the classification feature matrix through a classifier to obtain a classification result, where the classification result is used to indicate whether an early warning prompt indicating that a risk of dropping exists is generated.
In a specific embodiment of the present application, the transfer risk prompting module is further configured to: processing the classification feature matrix using the classifier to generate a classification result with the following formula:
Figure 253249DEST_PATH_IMAGE064
in which
Figure 574509DEST_PATH_IMAGE065
Representing the projection of the classification feature matrix as a vector,
Figure 125576DEST_PATH_IMAGE066
to is that
Figure 456063DEST_PATH_IMAGE067
Is a weight matrix of the fully connected layers of each layer,
Figure 41765DEST_PATH_IMAGE068
to
Figure 533926DEST_PATH_IMAGE069
Representing fully-connected layers of each layerThe matrix is biased.
In summary, based on the aluminum product discharge auxiliary transfer device of the embodiment of the application, the sequence of the first gravity center data of the fixing mechanism and the sequence of the second gravity center data of the aluminum product are extracted from the transfer monitoring video collected by the monitoring camera, the difference between the gravity center data of each corresponding time point in the sequence of the first gravity center data and the sequence of the second gravity center data is calculated to obtain the sequence of the gravity center difference, then, the sequence of the first gravity center data, the sequence of the second gravity center data and the sequence of the gravity center difference are properly encoded by using the convolutional neural network model, and are decoded by the classifier to obtain the classification result of the early warning prompt indicating whether the aluminum product is dropped or not, so that the early warning can be accurately performed when the aluminum product is dropped, the aluminum product is prevented from dropping in the transfer process, and the normal transfer and processing of the aluminum product are ensured.
Exemplary method
Fig. 5 illustrates a flow chart of an aluminum material discharge auxiliary transfer control method according to an embodiment of the present application. As shown in fig. 5, the aluminum material discharge auxiliary transfer control method according to the embodiment of the application includes: s110, acquiring a transfer monitoring video collected by a monitoring camera; s120, extracting a sequence of first gravity center data of a fixing mechanism and a sequence of second gravity center data of an aluminum material from the transfer monitoring video, wherein the fixing mechanism is used for fixing the aluminum material; s130, calculating differences between the gravity center data of each corresponding time point in the sequence of the first gravity center data and the sequence of the second gravity center data to obtain a sequence of gravity center differences; s140, passing the sequence of the gravity center difference values through a time sequence encoder comprising a one-dimensional convolution layer to obtain a first feature vector; s150, respectively passing the sequence of the first gravity center data and the sequence of the second gravity center data through the time sequence encoder containing the one-dimensional convolutional layer to obtain a second eigenvector and a third eigenvector; s160, calculating a first transfer matrix of the first feature vector relative to the second feature vector; s170, calculating a second transfer matrix of the first feature vector relative to the third feature vector; s180, respectively correcting eigenvalues of the first transfer matrix and the second transfer matrix based on the distance between corresponding positions in the first transfer matrix and the second transfer matrix to obtain a corrected first transfer matrix and a corrected second transfer matrix; s190, fusing the corrected first transfer matrix and the corrected second transfer matrix to obtain a classification characteristic matrix; and S200, enabling the classification characteristic matrix to pass through a classifier to obtain a classification result, wherein the classification result is used for indicating whether an early warning prompt for the existence of the falling risk is generated.
Fig. 6 illustrates a schematic diagram of a system architecture of an aluminum material discharge auxiliary transfer control method according to an embodiment of the present application. As shown in fig. 6, in the system architecture of the aluminum product discharge auxiliary transfer control method according to the embodiment of the present application, first, a transfer surveillance video collected by a surveillance camera is obtained, and a sequence of first center-of-gravity data of a fixing mechanism and a sequence of second center-of-gravity data of an aluminum product are extracted from the transfer surveillance video. Then, a difference between the gravity center data of each corresponding time point in the sequence of the first gravity center data and the sequence of the second gravity center data is calculated to obtain a sequence of gravity center differences, and the sequence of gravity center differences is passed through a time-series encoder including one-dimensional convolutional layers to obtain a first feature vector. And meanwhile, respectively passing the sequence of the first gravity center data and the sequence of the second gravity center data through the time sequence encoder containing the one-dimensional convolution layer to obtain a second feature vector and a third feature vector. Then, a first transfer matrix of the first eigenvector relative to the second eigenvector is calculated, and a second transfer matrix of the first eigenvector relative to the third eigenvector is calculated at the same time. And then, respectively correcting the eigenvalues of the first transfer matrix and the second transfer matrix based on the distance between corresponding positions in the first transfer matrix and the second transfer matrix to obtain a corrected first transfer matrix and a corrected second transfer matrix. And finally, fusing the corrected first transfer matrix and the corrected second transfer matrix to obtain a classification characteristic matrix, and enabling the classification characteristic matrix to pass through a classifier to obtain a classification result, wherein the classification result is used for indicating whether an early warning prompt for the existence of a falling risk is generated.
In a specific embodiment of the application, the calculating a difference between the barycentric data of each corresponding time point in the sequence of the first barycentric data and the sequence of the second barycentric data to obtain a sequence of barycentric differences includes: calculating a difference between the gravity center data of each corresponding time point in the sequence of the first gravity center data and the sequence of the second gravity center data by the following formula to obtain a sequence of the gravity center differences; wherein the formula is:
Figure 837869DEST_PATH_IMAGE070
wherein the content of the first and second substances,
Figure 709397DEST_PATH_IMAGE071
a sequence representing the first center of gravity data,
Figure 680764DEST_PATH_IMAGE072
a sequence representing the second center of gravity data,
Figure 609406DEST_PATH_IMAGE073
a sequence representing the difference in the center of gravity,
Figure 135065DEST_PATH_IMAGE074
representing a subtraction by position between the barycentric data of respective corresponding time points in the barycentric sequence.
In a specific embodiment of the present application, the passing the sequence of gravity center difference values through a time-series encoder including one-dimensional convolutional layers to obtain a first feature vector includes: arranging the sequence of the gravity center difference values into a gravity center difference input vector; performing full concatenation encoding on the barycentric difference input vector by using a full concatenation layer of the time sequence encoder according to the following formula to extract high-dimensional implicit features of feature values of all positions in the barycentric difference input vector, wherein the formula is as follows:
Figure 541776DEST_PATH_IMAGE075
wherein
Figure 102070DEST_PATH_IMAGE014
Is the center of gravity difference input vector,
Figure 936034DEST_PATH_IMAGE015
is the output vector of the output vector,
Figure 307296DEST_PATH_IMAGE076
is a matrix of weights that is a function of,
Figure 517698DEST_PATH_IMAGE017
is a vector of the offset to be used,
Figure 198078DEST_PATH_IMAGE018
represents a matrix multiplication; and performing one-dimensional convolution encoding on the gravity center difference input vector by using a one-dimensional convolution layer of the time sequence encoder according to the following formula so as to extract high-dimensional implicit correlation features among feature values of all positions in the gravity center difference input vector, wherein the formula is as follows:
Figure 796418DEST_PATH_IMAGE077
wherein the content of the first and second substances,ais a convolution kernelxA width in the direction,
Figure 562249DEST_PATH_IMAGE078
Is a convolution kernel parameter vector,
Figure 110430DEST_PATH_IMAGE079
Is a matrix of local vectors operating with a convolution kernel,wis the size of the convolution kernel and,
Figure 645316DEST_PATH_IMAGE080
representing the barycentric difference input vector.
In a specific embodiment of the present application, the first center of gravity is setThe sequence of the first feature vector and the sequence of the first barycentric data are respectively passed through the time-sequence encoder containing the one-dimensional convolutional layer to obtain a first feature vector and a second feature vector, and the method comprises the following steps: arranging the sequence of the first barycentric data and the sequence of the second barycentric data as a first barycentric input vector and a second barycentric input vector, respectively; using a full-connection layer of the time sequence encoder to perform full-connection encoding on the first gravity center input vector and the second gravity center input vector by using the following formula to extract high-dimensional implicit features of feature values of each position in the first gravity center input vector and the second gravity center input vector respectively, wherein the formula is as follows:
Figure 86662DEST_PATH_IMAGE081
wherein
Figure 339788DEST_PATH_IMAGE080
Is the first and second centroid input vectors,
Figure 626413DEST_PATH_IMAGE082
is the output vector of the digital video signal,
Figure 281386DEST_PATH_IMAGE083
is a matrix of the weights that is,
Figure 630983DEST_PATH_IMAGE084
is a vector of the offset to the offset,
Figure 840247DEST_PATH_IMAGE085
represents a matrix multiplication; and performing one-dimensional convolutional encoding on the first gravity center input vector and the second gravity center input vector by using a one-dimensional convolutional layer of the time-series encoder according to the following formula so as to respectively extract high-dimensional implicit correlation features between feature values of positions in the first gravity center input vector and the second gravity center input vector, wherein the formula is as follows:
Figure 930563DEST_PATH_IMAGE086
wherein the content of the first and second substances,ais a convolution kernelxA width in the direction,
Figure 440042DEST_PATH_IMAGE087
Is a convolution kernel parameter vector,
Figure 957611DEST_PATH_IMAGE088
Is a matrix of local vectors operating with a convolution kernel,wis the size of the convolution kernel and,
Figure 185330DEST_PATH_IMAGE089
representing the first and second barycentric input vectors.
In a specific embodiment of the present application, the calculating a first transfer matrix of the first eigenvector relative to the second eigenvector includes: calculating the first transfer matrix of the first feature vector relative to the second feature vector with the following formula; wherein the formula is:
Figure 813757DEST_PATH_IMAGE090
=
Figure 912163DEST_PATH_IMAGE091
*
Figure 134722DEST_PATH_IMAGE092
wherein
Figure 318579DEST_PATH_IMAGE034
Representing the first feature vector in a first set of features,
Figure 281855DEST_PATH_IMAGE035
representing the second feature vector in the second set of feature vectors,
Figure 500347DEST_PATH_IMAGE093
representing the first transfer matrix(ii) a Calculating a second transfer matrix of the first eigenvector relative to the third eigenvector, comprising: calculating the second transfer matrix of the first feature vector relative to the third feature vector in the following formula; wherein the formula is:
Figure 625298DEST_PATH_IMAGE090
=
Figure 562030DEST_PATH_IMAGE094
*
Figure 54629DEST_PATH_IMAGE095
wherein
Figure 127628DEST_PATH_IMAGE034
Representing the first feature vector in a first set of features,
Figure 157900DEST_PATH_IMAGE096
-representing the third feature vector by means of a third feature vector,
Figure 316349DEST_PATH_IMAGE097
representing the second transition matrix.
In a specific embodiment of the present application, the performing, based on a distance between corresponding positions in the first transition matrix and the second transition matrix, eigenvalue correction on the first transition matrix and the second transition matrix respectively to obtain a corrected first transition matrix and a corrected second transition matrix includes: performing eigenvalue correction on the first transfer matrix based on the distance between corresponding positions in the first transfer matrix and the second transfer matrix to obtain a corrected first transfer matrix; wherein the formula is:
Figure 355850DEST_PATH_IMAGE098
wherein, the first and the second end of the pipe are connected with each other,
Figure 814513DEST_PATH_IMAGE099
representing the second transition matrix in a second order,
Figure 15687DEST_PATH_IMAGE100
and
Figure 664362DEST_PATH_IMAGE101
respectively represent the first transition matrix and the second transition matrix
Figure 773132DEST_PATH_IMAGE102
The value of the characteristic of the location is,
Figure 86302DEST_PATH_IMAGE103
representing the first transition matrix and the second transition matrix
Figure 458377DEST_PATH_IMAGE102
A distance between the respective positions, and
Figure 325839DEST_PATH_IMAGE104
is a hyper-parameter; performing eigenvalue correction on the second transfer matrix based on the distance between corresponding positions in the first transfer matrix and the second transfer matrix to obtain a corrected second transfer matrix; wherein the formula is:
Figure 707142DEST_PATH_IMAGE105
wherein the content of the first and second substances,
Figure 609239DEST_PATH_IMAGE106
respectively, represent the first transfer matrix and,
Figure 686303DEST_PATH_IMAGE107
and
Figure 306641DEST_PATH_IMAGE108
respectively representIn the first and second transition matrices
Figure 960476DEST_PATH_IMAGE109
The value of the characteristic of the location is,
Figure 248238DEST_PATH_IMAGE110
representing the first transition matrix and the second transition matrix
Figure 962116DEST_PATH_IMAGE111
A distance between the respective positions, and
Figure 69749DEST_PATH_IMAGE112
is a hyper-parameter.
Here, it can be understood by those skilled in the art that the detailed operations of the respective steps in the above aluminum material discharging auxiliary transfer apparatus have been described in detail in the above description of the aluminum material discharging auxiliary transfer control method with reference to fig. 1 to 4, and therefore, the repetitive description thereof will be omitted.

Claims (10)

1. The utility model provides an auxiliary transfer equipment of aluminum product ejection of compact which characterized in that includes: the transfer monitoring video acquisition module is used for acquiring a transfer monitoring video acquired by the monitoring camera; the center of gravity data extraction module is used for extracting a sequence of first center of gravity data of a fixing mechanism and a sequence of second center of gravity data of the aluminum material from the transfer monitoring video, wherein the fixing mechanism is used for fixing the aluminum material; a center of gravity difference module, configured to calculate a difference between center of gravity data of each corresponding time point in the sequence of the first center of gravity data and the sequence of the second center of gravity data to obtain a sequence of center of gravity differences; the gravity center change feature extraction module is used for enabling the sequence of the gravity center difference value to pass through a time sequence encoder comprising a one-dimensional convolution layer so as to obtain a first feature vector; an absolute gravity center feature extraction module, configured to pass the sequence of the first gravity center data and the sequence of the second gravity center data through the time-series encoder including the one-dimensional convolutional layer, respectively, to obtain a second feature vector and a third feature vector; a first transfer module to calculate a first transfer matrix of the first eigenvector relative to the second eigenvector; a second transfer module for calculating a second transfer matrix of the first eigenvector relative to the third eigenvector; the characteristic correction module is used for respectively correcting the characteristic values of the first transfer matrix and the second transfer matrix based on the distance between corresponding positions in the first transfer matrix and the second transfer matrix to obtain a corrected first transfer matrix and a corrected second transfer matrix; the information fusion module is used for fusing the corrected first transfer matrix and the corrected second transfer matrix to obtain a classification characteristic matrix; and the transfer risk prompting module is used for enabling the classification characteristic matrix to pass through a classifier to obtain a classification result, and the classification result is used for indicating whether an early warning prompt for the existence of the falling risk is generated.
2. Aluminum product discharge auxiliary transfer equipment according to claim 1, wherein the center of gravity difference module is further configured to: calculating a difference between the gravity center data of each corresponding time point in the sequence of the first gravity center data and the sequence of the second gravity center data by the following formula to obtain a sequence of the gravity center differences; wherein the formula is:
Figure 352490DEST_PATH_IMAGE001
wherein, in the process,
Figure 331947DEST_PATH_IMAGE002
a sequence representing the first center of gravity data,
Figure 174001DEST_PATH_IMAGE003
a sequence representing the second center of gravity data,
Figure 165615DEST_PATH_IMAGE004
a sequence representing the difference in the center of gravity,
Figure 42304DEST_PATH_IMAGE005
representing a position-wise subtraction between the barycentric data of respective corresponding time points in the barycentric sequence.
3. The aluminum material discharge auxiliary transfer equipment of claim 2, wherein the gravity center change feature extraction module comprises: an arrangement unit configured to arrange the sequence of gravity center difference values as a gravity center difference input vector; a full-connection unit, configured to perform full-connection coding on the barycentric difference input vector using a full-connection layer of the time sequence encoder according to the following formula to extract high-dimensional implicit features of feature values of each position in the barycentric difference input vector, where the formula is:
Figure 927084DEST_PATH_IMAGE006
wherein
Figure 522013DEST_PATH_IMAGE007
Is the center of gravity difference input vector,
Figure 783230DEST_PATH_IMAGE008
is the output vector of the output vector,
Figure 514426DEST_PATH_IMAGE009
is a matrix of weights that is a function of,
Figure 835686DEST_PATH_IMAGE010
is a vector of the offset to be used,
Figure 920841DEST_PATH_IMAGE011
represents a matrix multiplication; and a one-dimensional convolution unit, configured to perform one-dimensional convolution encoding on the barycentric difference input vector by using a one-dimensional convolution layer of the time sequence encoder according to the following formula to extract high-dimensional implicit correlation features between feature values of each position in the barycentric difference input vector, where the formula is:
Figure 985749DEST_PATH_IMAGE012
wherein, in the process,ais a convolution kernel inxA width in the direction,
Figure 571451DEST_PATH_IMAGE013
Is a convolution kernel parameter vector,
Figure 329192DEST_PATH_IMAGE014
Is a matrix of local vectors operating with a convolution kernel,wis the size of the convolution kernel and,
Figure 633134DEST_PATH_IMAGE015
representing the barycentric difference input vector.
4. The aluminum material discharge auxiliary transfer equipment of claim 3, wherein the absolute gravity center feature extraction module comprises: a center-of-gravity data arrangement unit configured to arrange the sequence of the first center-of-gravity data and the sequence of the second center-of-gravity data as a first center-of-gravity input vector and a second center-of-gravity input vector, respectively; a full-concatenation encoding unit, configured to perform full-concatenation encoding on the first gravity center input vector and the second gravity center input vector using a full-concatenation layer of the time-series encoder according to the following formula to extract high-dimensional implicit features of feature values of respective positions in the first gravity center input vector and the second gravity center input vector, respectively, where the formula is:
Figure 970574DEST_PATH_IMAGE006
wherein
Figure 676362DEST_PATH_IMAGE016
Is the first and second centroid input vectors,
Figure 342354DEST_PATH_IMAGE017
is the output vector of the digital video signal,
Figure 133593DEST_PATH_IMAGE018
is a matrix of the weights that is,
Figure 805883DEST_PATH_IMAGE019
is a vector of the offset to the offset,
Figure 631756DEST_PATH_IMAGE020
represents a matrix multiplication; and a one-dimensional convolution encoding unit configured to perform one-dimensional convolution encoding on the first gravity center input vector and the second gravity center input vector using the one-dimensional convolution layer of the time-series encoder by using the following formula to extract high-dimensional implicit correlation features between feature values of respective positions in the first gravity center input vector and the second gravity center input vector, respectively, where the formula is:
Figure 465720DEST_PATH_IMAGE012
wherein, in the process,ais a convolution kernelxWidth in the direction,
Figure 744255DEST_PATH_IMAGE013
Is a convolution kernel parameter vector,
Figure 689077DEST_PATH_IMAGE014
Is a matrix of local vectors operating with a convolution kernel,wis the size of the convolution kernel and,
Figure 626247DEST_PATH_IMAGE021
representing the first and second centroid input vectors.
5. The aluminum product discharge auxiliary transfer equipment of claim 4, wherein the first transfer module is further configured to: calculating the first transfer matrix of the first feature vector relative to the second feature vector in the following formula; wherein the formula is:
Figure 365533DEST_PATH_IMAGE022
=
Figure 131364DEST_PATH_IMAGE023
*
Figure 879877DEST_PATH_IMAGE024
wherein
Figure 414763DEST_PATH_IMAGE025
Representing the first feature vector in a first set of features,
Figure 590530DEST_PATH_IMAGE026
-representing the second feature vector by means of a second feature vector,
Figure 578077DEST_PATH_IMAGE027
representing the first transition matrix; the second transfer module is further configured to: calculating the second transfer matrix of the first feature vector relative to the third feature vector with the following formula; wherein the formula is:
Figure 398790DEST_PATH_IMAGE022
=
Figure 522604DEST_PATH_IMAGE028
*
Figure 869272DEST_PATH_IMAGE029
in which
Figure 609695DEST_PATH_IMAGE025
-representing the first feature vector by means of a first representation,
Figure 700011DEST_PATH_IMAGE030
representing the third feature vector and the second feature vector,
Figure 209489DEST_PATH_IMAGE031
representing the second transition matrix.
6. Aluminum product discharge auxiliary transfer equipment according to claim 5, wherein the characteristic correction module comprises: a first correction unit, configured to perform eigenvalue correction on the first transfer matrix based on a distance between corresponding positions in the first transfer matrix and the second transfer matrix to obtain a corrected first transfer matrix; wherein the formula is:
Figure 727058DEST_PATH_IMAGE032
wherein, the first and the second end of the pipe are connected with each other,
Figure 957707DEST_PATH_IMAGE033
representing the second transition matrix and the second transition matrix,
Figure 586134DEST_PATH_IMAGE034
and
Figure 481278DEST_PATH_IMAGE035
respectively represent the first transition matrix and the second transition matrix
Figure 169748DEST_PATH_IMAGE036
The value of the characteristic of the location is,
Figure 884764DEST_PATH_IMAGE037
representing the first transition matrix and the second transition matrix
Figure 316882DEST_PATH_IMAGE036
Distance between the respective positions, and
Figure 269794DEST_PATH_IMAGE038
is a hyper-parameter; second correctionA unit, configured to perform eigenvalue correction on the second transfer matrix based on a distance between corresponding positions in the first transfer matrix and the second transfer matrix to obtain a corrected second transfer matrix; wherein the formula is:
Figure 928833DEST_PATH_IMAGE039
wherein the content of the first and second substances,
Figure 599986DEST_PATH_IMAGE040
respectively, represent the first transfer matrix and,
Figure 835795DEST_PATH_IMAGE041
and
Figure 174373DEST_PATH_IMAGE042
respectively represent the first transition matrix and the second transition matrix
Figure 1383DEST_PATH_IMAGE043
The value of the characteristic of the location is,
Figure 783718DEST_PATH_IMAGE044
representing the first transition matrix and the second transition matrix
Figure 88798DEST_PATH_IMAGE045
Distance between the respective positions, and
Figure 547461DEST_PATH_IMAGE046
is a hyper-parameter.
7. The aluminum material discharge auxiliary transfer equipment of claim 6, wherein the information fusion module is further configured to: fusing the corrected first transition matrix and the corrected second transition matrix to obtain the final productThe classification characteristic matrix; wherein the formula is:
Figure 14214DEST_PATH_IMAGE047
wherein, in the process,
Figure 659959DEST_PATH_IMAGE048
for the purpose of the classification feature matrix,
Figure 237571DEST_PATH_IMAGE049
for the corrected first transfer matrix to be used,
Figure 285162DEST_PATH_IMAGE050
for the second transfer matrix after the correction, the first transfer matrix,
Figure 925746DEST_PATH_IMAGE051
for a weighting parameter for controlling a balance between the corrected first transition matrix and the corrected second transition matrix in the classification feature matrix,
Figure 58787DEST_PATH_IMAGE052
representing a position-wise addition of the matrix.
8. The aluminum material discharge auxiliary transfer equipment of claim 7, wherein the transfer risk prompting module is further configured to: processing the classification feature matrix using the classifier to generate a classification result with the following formula:
Figure 440090DEST_PATH_IMAGE053
wherein
Figure 607766DEST_PATH_IMAGE054
Representing the projection of the classification feature matrix as a vector,
Figure 150743DEST_PATH_IMAGE055
to
Figure 771080DEST_PATH_IMAGE056
Is a weight matrix of the fully connected layers of each layer,
Figure 690495DEST_PATH_IMAGE057
to is that
Figure 981186DEST_PATH_IMAGE058
A bias matrix representing the layers of the fully connected layer.
9. An aluminum product discharge auxiliary transfer control method is characterized by comprising the following steps: acquiring a transfer monitoring video acquired by a monitoring camera; extracting a sequence of first gravity center data of a fixing mechanism and a sequence of second gravity center data of an aluminum material from the transfer monitoring video, wherein the fixing mechanism is used for fixing the aluminum material; calculating differences between the gravity center data of each corresponding time point in the sequence of the first gravity center data and the sequence of the second gravity center data to obtain a sequence of gravity center differences; passing the sequence of gravity center difference values through a time sequence encoder comprising a one-dimensional convolutional layer to obtain a first feature vector; respectively passing the sequence of the first barycentric data and the sequence of the second barycentric data through the time-sequence encoder comprising the one-dimensional convolutional layer to obtain a second eigenvector and a third eigenvector; calculating a first transfer matrix of the first eigenvector relative to the second eigenvector; calculating a second transfer matrix of the first eigenvector relative to the third eigenvector; respectively correcting eigenvalues of the first transfer matrix and the second transfer matrix based on the distance between corresponding positions in the first transfer matrix and the second transfer matrix to obtain a corrected first transfer matrix and a corrected second transfer matrix; fusing the corrected first transfer matrix and the corrected second transfer matrix to obtain a classification feature matrix; and enabling the classification characteristic matrix to pass through a classifier to obtain a classification result, wherein the classification result is used for indicating whether an early warning prompt for existence of a falling risk is generated.
10. The aluminum material discharge auxiliary transfer control method as recited in claim 9, wherein the calculating a difference between the gravity center data at each corresponding time point in the sequence of the first gravity center data and the sequence of the second gravity center data to obtain the sequence of gravity center differences includes: calculating a difference between the gravity center data of each corresponding time point in the sequence of the first gravity center data and the sequence of the second gravity center data to obtain a sequence of the gravity center differences; wherein the formula is:
Figure 695064DEST_PATH_IMAGE001
wherein the content of the first and second substances,
Figure 68277DEST_PATH_IMAGE059
a sequence representing the first center of gravity data,
Figure 791382DEST_PATH_IMAGE060
a sequence representing the second center of gravity data,
Figure 668071DEST_PATH_IMAGE061
a sequence representing the difference in the center of gravity,
Figure 818430DEST_PATH_IMAGE062
representing a position-wise subtraction between the barycentric data of respective corresponding time points in the barycentric sequence.
CN202211024805.0A 2022-08-25 2022-08-25 Aluminum product discharging auxiliary transfer equipment and transfer control method Active CN115520786B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202211024805.0A CN115520786B (en) 2022-08-25 2022-08-25 Aluminum product discharging auxiliary transfer equipment and transfer control method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202211024805.0A CN115520786B (en) 2022-08-25 2022-08-25 Aluminum product discharging auxiliary transfer equipment and transfer control method

Publications (2)

Publication Number Publication Date
CN115520786A true CN115520786A (en) 2022-12-27
CN115520786B CN115520786B (en) 2024-02-27

Family

ID=84696971

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211024805.0A Active CN115520786B (en) 2022-08-25 2022-08-25 Aluminum product discharging auxiliary transfer equipment and transfer control method

Country Status (1)

Country Link
CN (1) CN115520786B (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110422767A (en) * 2019-06-27 2019-11-08 三一海洋重工有限公司 To the method, apparatus and system of suspender positioning
KR20200065595A (en) * 2018-11-30 2020-06-09 주식회사 무스마 System and method for analyzing cargo of crane
CN112241747A (en) * 2019-07-16 2021-01-19 顺丰科技有限公司 Object sorting method, device, sorting equipment and storage medium
CN112489087A (en) * 2020-12-13 2021-03-12 深圳市进致网络科技有限公司 Method for detecting shaking of suspension type operation platform for high-rise building construction
CN112800912A (en) * 2021-01-20 2021-05-14 江苏天幕无人机科技有限公司 Dynamic feature based label-based migration feature neural network training method

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20200065595A (en) * 2018-11-30 2020-06-09 주식회사 무스마 System and method for analyzing cargo of crane
CN110422767A (en) * 2019-06-27 2019-11-08 三一海洋重工有限公司 To the method, apparatus and system of suspender positioning
CN112241747A (en) * 2019-07-16 2021-01-19 顺丰科技有限公司 Object sorting method, device, sorting equipment and storage medium
CN112489087A (en) * 2020-12-13 2021-03-12 深圳市进致网络科技有限公司 Method for detecting shaking of suspension type operation platform for high-rise building construction
CN112800912A (en) * 2021-01-20 2021-05-14 江苏天幕无人机科技有限公司 Dynamic feature based label-based migration feature neural network training method

Also Published As

Publication number Publication date
CN115520786B (en) 2024-02-27

Similar Documents

Publication Publication Date Title
Yan et al. Perspective-guided convolution networks for crowd counting
CN109948524B (en) Traffic vehicle density estimation method based on space-based monitoring
WO2018105028A1 (en) Inspection device and inspection method
JP5719230B2 (en) Object recognition device, method for controlling object recognition device, and program
CN109977895B (en) Wild animal video target detection method based on multi-feature map fusion
CN103443823B (en) Shooting pattern recognition device, shooting image identification system and shooting image-recognizing method
JP2008234551A (en) Abnormality detection apparatus and abnormality detection program
US20110279685A1 (en) Method and system for automatic objects localization
CN114758304B (en) High-purity rounded quartz powder sieving equipment and sieving control method thereof
CN112381132A (en) Target object tracking method and system based on fusion of multiple cameras
US20110243385A1 (en) Moving object detection apparatus, moving object detection method, and program
EP2993621A1 (en) Method and apparatus for detecting shielding against object
EP1096799B1 (en) Camera signal processing device and camera signal processing method
US7616779B2 (en) Method for automatic key posture information abstraction
CN108537235B (en) Method for extracting image features by low-complexity scale pyramid
CN115520786A (en) Aluminum material discharging auxiliary transfer equipment and transfer control method
CN113886632B (en) Video retrieval matching method based on dynamic programming
CN116055460A (en) Image compression recovery method under complex communication condition
Pietikaeinen et al. Experiments with two industrial problems using texture classification based on feature distributions
CN110619626A (en) Image processing apparatus, system, method and device
JP2004054442A (en) Face detecting device, face detecting method, and face detecting program
CN113221971B (en) Multi-scale crowd counting method and system based on front and back feature fusion
WO2012148258A1 (en) Abrupt object movement detection system and method thereof
KR20220075442A (en) Scenario information detection method, apparatus, electronic device, medium and program
CN112750096A (en) Neural network training method for monitoring bagged cement capacity

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant