CN113836795A - Method and platform for building initial cargo placement attitude model - Google Patents
Method and platform for building initial cargo placement attitude model Download PDFInfo
- Publication number
- CN113836795A CN113836795A CN202111044006.5A CN202111044006A CN113836795A CN 113836795 A CN113836795 A CN 113836795A CN 202111044006 A CN202111044006 A CN 202111044006A CN 113836795 A CN113836795 A CN 113836795A
- Authority
- CN
- China
- Prior art keywords
- initial placement
- cargo
- goods
- neural network
- posture
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 48
- 230000001133 acceleration Effects 0.000 claims abstract description 83
- 238000003062 neural network model Methods 0.000 claims abstract description 37
- 230000009466 transformation Effects 0.000 claims abstract description 19
- 238000012549 training Methods 0.000 claims abstract description 18
- 238000013528 artificial neural network Methods 0.000 claims abstract description 10
- 238000002372 labelling Methods 0.000 claims abstract description 7
- 238000005070 sampling Methods 0.000 claims description 17
- 238000001514 detection method Methods 0.000 claims description 15
- 239000013598 vector Substances 0.000 claims description 13
- 238000004364 calculation method Methods 0.000 claims description 4
- 230000036544 posture Effects 0.000 description 49
- 238000004458 analytical method Methods 0.000 description 11
- 238000012360 testing method Methods 0.000 description 11
- 238000004806 packaging method and process Methods 0.000 description 10
- 230000008569 process Effects 0.000 description 10
- 238000010801 machine learning Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 5
- 238000013461 design Methods 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 3
- 238000004422 calculation algorithm Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 239000000470 constituent Substances 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000005484 gravity Effects 0.000 description 2
- 238000009863 impact test Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 238000005406 washing Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000002542 deteriorative effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000002779 inactivation Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000001902 propagating effect Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000035939 shock Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
- 230000002194 synthesizing effect Effects 0.000 description 1
- 238000010998 test method Methods 0.000 description 1
- 230000004580 weight loss Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
- G06Q10/083—Shipping
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Business, Economics & Management (AREA)
- Medical Informatics (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Economics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Hardware Design (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Human Resources & Organizations (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Geometry (AREA)
- Image Analysis (AREA)
Abstract
A method for constructing a cargo initial placement attitude model comprises the following steps: the method comprises the following steps: acquiring triaxial acceleration data when goods are initially placed, and determining a section when the goods are stably placed according to the triaxial acceleration data; acquiring the average value of the three-axis acceleration of the goods placed in the section when the goods are placed stably, and carrying out nonlinear transformation on the average value; comparing the average value of the triaxial acceleration with the initial placement posture of the goods, labeling the initial placement posture of the goods according to the comparison result to obtain labeled data, sending the average value of the triaxial acceleration after nonlinear transformation and the labeled data into a neural network for training, and determining a neural network model of the initial placement posture of the goods according to a preset requirement on the identification accuracy of the initial placement posture of the goods.
Description
Technical Field
The invention relates to the field of intelligent control, in particular to a method for building a model of an initial placement posture of goods, a platform and a method for recognizing the initial placement posture of the goods.
Background
Goods from factory to seller can go through the links of storage, transportation and transportation, and can be subjected to actions such as vibration, impact, falling and even violent handling actions, wherein the initial placement posture (face, edge and angle) of the goods is an important parameter for establishing the packaging and transportation standard in a laboratory. The information is accurately detected and identified, the package transportation test method, the product structure and the package design are favorably and pertinently improved, the cargo damage responsibility is positioned, and the transportation loss is reduced.
The above information disclosed in the background section is only for further understanding of the background of the invention and therefore it may contain information that does not form the prior art that is already known to a person of ordinary skill in the art.
Disclosure of Invention
The invention provides a method for constructing a model of initial placement postures of goods, a platform method for identifying the initial placement postures of the goods and a method for identifying the initial placement postures of the goods.
The invention provides a method for constructing a cargo initial placement attitude model, which comprises the following steps: the method comprises the following steps: s1: acquiring triaxial acceleration data when goods are initially placed, and determining a section when the goods are stably placed according to the triaxial acceleration data; s2: acquiring the average value of the three-axis acceleration of the goods placed in the section when the goods are placed stably, and carrying out nonlinear transformation on the average value; s3: comparing the average value of the triaxial acceleration with the initial placing posture of the goods, and labeling the placing posture of the goods according to the comparison result to obtain labeling data, S4: and sending the average value of the three-axis acceleration after the nonlinear transformation and the labeled data into a neural network for training, and determining a cargo initial placement posture neural network model according to a preset cargo initial placement posture identification accuracy requirement.
According to an embodiment of the present invention, wherein the S1 includes: the initial placement attitude of the cargo is acquired through a transportation environment detection device, wherein a three-axis acceleration sensor is included in the transportation environment detection device.
According to an embodiment of the present invention, in said S1, the peak position of the wave-shaped curve of the resultant vector acceleration a of the triaxial acceleration of the initial placement attitude of the cargo is determined, and the time coordinate of the peak is recorded as tmax, from the start position t of a0The step length N is traversed towards the direction of the wave crest side, and the interval [0, N × N ] is calculated in each traversal]Average value of vector acceleration a in interval
If it is not
Determining the interval as a candidate interval when traversing to tmaxWhen the goods are placed, the traversal is finished, and the candidate block section with the longest time length is selected as the block section when the goods are placed stably, wherein
ax、axAnd axThe acceleration in X, Y and Z directions when the triaxial acceleration sensor detects that the goods are initially placed is respectively, n is the number of traversal times, and b is a preset threshold value.
According to an embodiment of the present invention, in S1, when the stable time interval for determining the placement of the goods cannot be obtained at the end of the traversal, the time interval for determining the placement of the goods is set as a preset value, where the preset value is related to the three-axis acceleration data sampling rate, the sampling time length, and the pre-trigger time length.
According to an embodiment of the present invention, in the S2, the average value of the three-axis acceleration is
And
said non-linear transformation being
And
according to an embodiment of the present invention, in S3, the initial placement posture of the cargo is a plane, an edge and a corner, and the labeled data is: the plane of the initial placement posture of the load is set to 0, the edge is set to 1, and the angle is set to 2.
According to one embodiment of the invention, the preset initial cargo placement posture recognition accuracy is that the initial cargo placement posture is accurately recognized to be more than 90%.
According to an embodiment of the present invention, the input of the neural network model during training is: the average value of the three-axis acceleration after nonlinear transformation and the labeled data; the output of the neural network model during training is as follows: after training, initially placing a posture recognition neural network model, wherein the neural network model at least comprises the number of network layers of the neural network model, the number of nodes on each layer and relevant parameters of the neural network; the input of the neural network model in prediction is as follows: the average value of the three-axis acceleration after the nonlinear transformation in the section when the initial placement attitude of the goods is placed stably; the output of the neural network model in prediction is as follows: initial placement attitude of the cargo.
A second aspect of the present invention provides a platform for recognizing an initial placement posture of a cargo, comprising: the system comprises an input device, a neural network model and an output device, wherein the input device is used for collecting initial placement attitude data of cargos; the neural network model calculates the collected initial placement attitude data of the goods; and the output device outputs a posture report of initial placement of the cargo according to a calculation result of the neural network model, wherein the neural network model is obtained according to the method.
According to an embodiment of the present invention, the cargo initial placement posture report includes a cargo initial placement posture.
A third aspect of the present invention provides a method for identifying an initial placement posture of a cargo, comprising: and sending data related to the cargo eating and placing posture into the platform for identifying the initial placing posture of the cargo, and determining the initial placing posture of the cargo according to a report output by the platform.
According to the scheme of the invention, the machine learning algorithm is used for replacing a threshold segmentation method, so that the accuracy of the recognition of the placing posture is effectively improved. The scheme of the invention can accurately identify the goods placing state in the packaging and transporting process, and is beneficial to making more reasonable packaging test standards and pertinently improving the product packaging design.
Drawings
In order to more clearly illustrate the technical solution of the present invention, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a flowchart of a method for constructing an initial cargo placement attitude model according to an exemplary embodiment of the present invention.
FIG. 2 is a flow chart of an implementation of the machine learning initial placement pose model construction of the present invention, according to an exemplary embodiment of the present invention.
FIG. 3 is a graphical illustration of typical accelerations in the machine learning initial placement pose model of the present invention, according to an exemplary embodiment of the present invention.
FIG. 4 is a machine-learned initial placement pose model verification flow diagram of the present invention, according to an exemplary embodiment of the present invention.
FIG. 5 is a block diagram of an initial placement gesture recognition platform of the present invention according to an exemplary embodiment of the present invention.
FIG. 6 is a flowchart illustrating operation of the initial placement gesture recognition platform of the present invention according to an illustrative embodiment of the present invention.
DETAILED DESCRIPTION OF EMBODIMENT (S) OF INVENTION
As used herein, the terms "first," "second," and the like may be used to describe elements of exemplary embodiments of the invention. These terms are only used to distinguish one element from another element, and the inherent features or order of the corresponding elements and the like are not limited by the terms. Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Terms, such as those defined in commonly used dictionaries, are to be interpreted as having a meaning that is consistent with their context in the relevant art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.
Those skilled in the art will understand that the devices and methods of the present invention described herein and illustrated in the accompanying drawings are non-limiting exemplary embodiments and that the scope of the present invention is defined solely by the claims. Features illustrated or described in connection with one exemplary embodiment may be combined with features of other embodiments. Such modifications and variations are intended to be included within the scope of the present invention.
Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the drawings, a detailed description of related known functions or configurations is omitted to avoid unnecessarily obscuring the technical points of the present invention. In addition, the same reference numerals refer to the same circuits, modules or units throughout the description, and repeated descriptions of the same circuits, modules or units are omitted for brevity.
Further, it should be understood that one or more of the following methods or aspects thereof may be performed by at least one control unit or controller. The terms "control unit," "controller," "control module," or "master module" may refer to a hardware device that includes a memory and a processor. The memory or computer-readable storage medium is configured to store program instructions, while the processor is specifically configured to execute the program instructions to perform one or more processes that will be described further below. Moreover, it is to be appreciated that the following methods may be performed by including a processor in conjunction with one or more other components, as will be appreciated by one of ordinary skill in the art.
In order to identify the initial placement attitude, the invention is embedded with a transportation environment analysis platform, carries out identification analysis on the data acquired by the acceleration sensor which is transported along with the goods, for example, the probability of the working attitude in the initial part of the goods in the actual transportation process is analyzed, the invention can be used for formulating the laboratory packaging test standard, better investigating the product strength and the packaging buffer performance, and meanwhile, the invention aims at optimizing and improving the product structural strength and the buffer packaging design, such as the heavy protection of the initial placement attitude direction with high probability.
The product package is generally hexahedron, the initial placement posture can be generally divided into three conditions of face, edge and angle, the recognition accuracy is improved by repeatedly training the model through a machine learning method, and compared with a common threshold value method, the recognition accuracy can be improved by the scheme provided by the invention.
Fig. 1 is a flowchart of a method for constructing an initial cargo placement attitude model according to an exemplary embodiment of the present invention. As shown in figure 1 of the drawings, in which,
at the step S1, collecting triaxial acceleration data when the cargo is initially placed, and determining a section when the cargo is stably placed according to the triaxial acceleration data;
at the step S2, obtaining an average value of the triaxial accelerations of the placement of the goods in the section where the placement of the goods is stable, and performing nonlinear transformation on the average value;
at the step S3, the average value of the three-axis acceleration is compared with the initial placement posture of the cargo, the cargo placement posture is labeled according to the comparison result to obtain labeled data,
in the step S4, the average value of the three-axis accelerations after the nonlinear transformation and the labeled data are sent to a neural network for training, and a cargo initial placement posture neural network model is determined according to a preset cargo initial placement posture recognition accuracy requirement.
FIG. 2 is a flow chart of an implementation of the machine learning initial placement pose model construction of the present invention, according to an exemplary embodiment of the present invention. As shown in figure 2 of the drawings, in which,
firstly, acquiring a series of vibration or impact events of known placing postures through a three-axis acceleration sensor, and importing three-axis acceleration data of the vibration or impact events into a platform; then, using the formula
(wherein a denotes a resultant vector acceleration, a)x、axAnd axX, Y and Z direction accelerations acquired by an acceleration sensor respectively) to carry out vector synthesis on the triaxial acceleration data; and finally, putting the acceleration waveform synthesized by the vectors into a Butterworth low-pass filter for filtering high-frequency noise.
Proceeding next to the process of finding stable segments, FIG. 3 is a diagram of typical accelerations in the machine-learned initial placement pose model of the present invention, according to an exemplary embodiment of the present invention.
As shown in FIG. 3, first, a peak of the vector acceleration is found, and its time coordinate is recorded as tmax(ii) a Then, from the acceleration start position t0The step length N (corresponding to the data point serial number 0) is traversed to the right (on the peak side), where N represents the number of sampling points and can be selected according to the sampling frequency of the acceleration data, and the higher the sampling frequency is, the larger N can be obtained, for example, when the sampling frequency is 800Hz, N can be 2. Each time the interval [0, N x N ] is calculated]Mean value of vector acceleration within interval
n is the nth traversal. Normally, when the cargo is placed still, the acceleration sensor only receives the action of gravity, and the vector acceleration is close to the gravity acceleration of 1 g. If it is
Wherein b is a defined threshold value, generally 0.1-0.2 g, recording the intervalTraversing to a vector acceleration peak t for a candidate interval, or else not recording as a candidate areamaxEnd of treatment, tmaxIs the peak time; and finally, if the candidate interval exists, selecting the candidate interval with the longest time length as an initial placement stable interval, and if the candidate interval does not exist, selecting a fixed interval as an initial placement interval. The fixed interval may be selected [0,150 ] during actual operation]. Where 150 represents the data point, which is related to the data sampling rate and the impact pre-trigger time setting, which is the time period during which the data sample should be recorded before the acceleration impact trigger is set. The dimension of 150 sampling points is the number of sampling points, namely the 150 th sampling point from the acceleration can select an empirical value which is related to the data sampling rate, the sampling duration and the pre-trigger duration; for example, when the sampling rate of the detection device is 800Hz, the total sampling point number of one acceleration data is 2880, that is, the sampling time is 3.6s, and the impact pre-trigger time is set to 1.6s, considering that the falling weight loss time does not exceed 1s, the time length of the placement section in one acceleration data may be 0.6s, which corresponds to 480 points, and it is appropriate to take the upper limit of the fixed interval to 150 points.
According to one or more embodiments of the invention, after the initial placement stable interval is determined, the corresponding three-axis acceleration waveform is intercepted according to the initial placement stable interval, and the respective average value of the three-axis acceleration in the stable placement interval is calculated,
and
then, for the convenience of subsequent machine learning, the respective average values of the three-axis acceleration are subjected to nonlinear transformation, and the transformed values are respectively
And
forming neural network input vectors
Then, the three-axis average value is calculated
And
comparing and analyzing the initial placing posture of the vibration or impact test corresponding to the data, clearing error data, and marking the initial placing posture of the vibration or impact event according to the data, wherein the surface is set to 0, the edge is set to 1, and the angle is set to 2, and calculating
And storing the data and the labeled data into a database.
According to one or more embodiments of the invention, a series of non-linear transforms of the respective average of the three-axis accelerations for known initial placement poses are transformed
And the annotation data are sent into the constructed DNN neural network, the network parameter (such as the number of network layers, the number of nodes in each layer, the random inactivation probability and the like) is adjusted according to the loss rate and the accuracy, whether the identification accuracy requirement is met or not is judged, if the identification accuracy requirement is met, the model is successfully constructed, the model meeting the requirements of the client is output, if the requirement is not met, the known initial placement attitude event is added, the network parameter is adjusted, the model is retrained until the requirements are met, and the model meeting the requirements of the client is output.
According to one or more embodiments of the invention, the accuracy of the recognition of the initial placement posture of the goods is set to accurately recognize that the initial placement posture of the goods is more than 90 percent
According to one or more embodiments of the invention, the database stores raw triaxial acceleration data and a nonlinear transformation value of the respective average values of triaxial acceleration of the stabilized interval
And annotations, etc.
According to one or more embodiments of the invention, washing and labeling are used in the model training phase, where washing: screening data used for training the model, and excluding error data: for example, the sensor is not attached to the surface, which causes the error of the recorded data. Wherein the labeling is as follows: the input data of the model is labeled, i.e. what initial placement posture the data corresponds to, e.g. face set to 0, edge set to 1, angle set to 2.
According to one or more embodiments of the invention, the inputs of the neural network model at the time of training are: the average value of the three-axis acceleration after nonlinear transformation and the labeled data; the output of the neural network model during training is as follows: after training, initially placing a posture recognition neural network model, wherein the neural network model at least comprises the number of network layers of the neural network model, the number of nodes on each layer and relevant parameters of the neural network; the input of the neural network model in prediction is as follows: the average value of the three-axis acceleration after the nonlinear transformation in the section when the initial placement attitude of the goods is placed stably; the output of the neural network model in prediction is as follows: initial placement attitude of the cargo. .
According to one or more embodiments of the invention, a three-axis acceleration sensor is built in the transportation environment detection device, and the three-axis acceleration data refers to X, Y, Z acceleration curves in three directions acquired by the three-axis acceleration sensor when an impact event occurs. Synthesizing by using a formula:
wherein a denotes a resultant vector acceleration, ax、axAnd axAcceleration sensor X, Y and the acceleration in the Z direction, respectively.
FIG. 4 is a machine-learned initial placement pose model verification flow diagram of the present invention, according to an exemplary embodiment of the present invention. As shown in fig. 4, the data of the triaxial acceleration of the unknown placing posture event is imported into the platform, the respective average values of the triaxial acceleration in the stable initial placing posture interval and the calculation interval are identified by the same method as the training model, corresponding nonlinear transformation is performed, the transformed data is sent into the trained initial placing posture identification model for identification and judgment, the placing posture is output after the judgment, and the process is ended.
FIG. 5 is a block diagram of an initial placement gesture recognition platform of the present invention according to an exemplary embodiment of the present invention.
As shown in fig. 6, the platform for identifying the initial placement posture of the cargo, the input device, the neural network model and the output device, wherein the input device is used for collecting initial placement data of the cargo; the neural network model calculates the collected initial placement data of the goods; and the output device outputs a cargo initial placement attitude report according to the calculation result of the neural network model, wherein the neural network model is trained according to the method shown in the figure 2.
FIG. 6 is a flowchart illustrating operation of the initial placement gesture recognition platform of the present invention according to an illustrative embodiment of the present invention.
As shown in fig. 6, a transportation environment detection device with built-in three-axis acceleration sensor and data acquisition, storage and transmission components is placed in a product package (usually, the transportation environment detection device may be a black box to the outside), and the transportation environment detection device is started; and then, carrying out impact test (falling, throwing and the like) test or actual transportation on the transportation environment detection device to obtain initial placement attitude data, then uploading the initial placement attitude data to an analysis platform through wireless remote or local connection, running an initial placement attitude analysis algorithm and a model by the platform, carrying out model training or initial placement attitude recognition, then summarizing corresponding information to generate an analysis report, and closing the transportation environment detection device after the test or the transportation is finished.
The transportation environment detection device can be a black box for the outside, and is composed of a built-in three-axis acceleration sensor, a data acquisition component, a data storage component, a data transmission component, a battery component and the like. The transportation environment detection device can perform testing of an initial placement posture, and mainly comprises an initial placement test (performed by using an impact table), a drop test (a drop test machine), and throwing (a throwing machine or hand throwing). The data collected by the transportation environment detection device is triaxial acceleration curve data of an initial placement attitude event.
According to one or more embodiments of the invention, an analysis report is automatically generated by the platform, and the analysis report comprises the initial placement posture result of the goods in the test or the actual transportation process. The cut-off conditions of the platform operation technology are as follows: after the test or the actual transportation is finished, the operation is manually stopped or automatically stopped after all the initial placing posture events of the goods are identified.
According to one or more embodiments of the invention, the platform for identifying the initial placement posture of the goods is a logistics environment analysis platform, mainly comprises a server, a database, client sides (a webpage side, a PC side and a mobile side) and the like, integrates an analysis algorithm and a machine learning model, and can realize functions of online detection, online storage, automatic identification, statistical analysis, information release, multi-user sharing and the like.
According to one or more embodiments of the invention, the scheme of the invention finds an initial placement stable interval section in the cargo vibration or shock acceleration signal through a threshold value. In addition, the scheme of the invention calculates the nonlinear variation value of the three-axis average value in the initial placement stable interval, and sends the nonlinear variation value to the neural network for machine learning to produce the initial posture recognition model. In addition, the scheme of the invention applies the initial attitude model to identify the object state when the goods are stably placed in the actual transportation process.
According to one or more embodiments of the invention, in order to identify the initial placement posture, the transportation environment analysis platform is embedded, and identification analysis is performed on the acceleration sensor data transported along with the goods, for example, the probability of each initial placement posture of the goods in the actual transportation process is analyzed, so that the laboratory packaging test standard can be formulated, the product strength and the packaging buffer performance can be better inspected, and meanwhile, the product structural strength and the buffer packaging design are improved aiming at optimization, for example, the initial placement posture with high probability is mainly protected in the direction.
In accordance with one or more embodiments of the present invention, processing logic in methods of the present invention may implement processing as flows of the above methods of the present invention using encoded instructions (e.g., computer and/or machine readable instructions) stored on a non-transitory computer and/or machine readable medium (e.g., a hard disk drive, a flash memory, a read-only memory, an optical disk, a digital versatile disk, a cache, a random-access memory, and/or any other storage device or storage disk) in which information is stored for any period of time (e.g., for extended periods of time, permanent, transitory instances, a temporary cache, and/or an information cache). As used herein, the term "non-transitory computer-readable medium" is expressly defined to include any type of computer-readable storage device and/or storage disk and to exclude propagating signals and to exclude transmission media.
In accordance with one or more embodiments of the present invention, the method of the present invention may be implemented using control circuitry, (control logic, a master control system or control module), which may include one or more processors, or which may internally include a non-transitory computer-readable medium. In particular, the master control system or control module may comprise a microcontroller MCU. The processor implementing the processes of the present method may be such as, but not limited to, one or more single-core or multi-core processors. The processor(s) may include any combination of general-purpose processors and special-purpose processors (e.g., graphics processors, application processors, etc.). The processor may be coupled thereto and/or may include a memory/storage device and may be configured to execute instructions stored in the memory/storage device to implement various applications and/or operating systems running on the controller in accordance with the present invention.
The drawings referred to above and the detailed description of the invention, which are exemplary of the invention, serve to explain the invention without limiting the meaning or scope of the invention as described in the claims. Accordingly, modifications may be readily made by those skilled in the art from the foregoing description. Further, those skilled in the art may delete some of the constituent elements described herein without deteriorating the performance, or may add other constituent elements to improve the performance. Further, the order of the steps of the methods described herein may be varied by one skilled in the art depending on the environment of the process or apparatus. Therefore, the scope of the present invention should be determined not by the embodiments described above but by the claims and their equivalents.
While the invention has been described in connection with what is presently considered to be practical embodiments, it is to be understood that the invention is not to be limited to the disclosed embodiments, but on the contrary, is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims.
Claims (11)
1. A method for constructing a cargo initial placement attitude model comprises the following steps: the method comprises the following steps:
s1: acquiring triaxial acceleration data when goods are initially placed, and determining a section when the goods are stably placed according to the triaxial acceleration data;
s2: acquiring the average value of the three-axis acceleration of the goods placed in the section when the goods are placed stably, and carrying out nonlinear transformation on the average value;
s3: comparing the average value of the three-axis acceleration with the initial placing posture of the goods, labeling the placing posture of the goods according to the comparison result to obtain labeling data,
s4: and sending the average value of the three-axis acceleration after the nonlinear transformation and the labeled data into a neural network for training, and determining a cargo initial placement posture neural network model according to a preset cargo initial placement posture identification accuracy requirement.
2. The method of claim 1, wherein the S1 includes:
the initial placement attitude of the cargo is acquired through a transportation environment detection device, wherein a three-axis acceleration sensor is included in the transportation environment detection device.
3. The method as claimed in claim 1, wherein in the S1, a peak position of a wave form curve of a resultant vector acceleration a of the three-axis acceleration of the initial placement attitude of the cargo is determined, andrecording the time coordinate of the peak as tmax from the initial position t of a0The step length N is traversed towards the direction of the wave crest side, and the interval [0, N × N ] is calculated in each traversal]Average value of vector acceleration a in interval
If it is not
Determining the interval as a candidate interval when traversing to tmaxWhen the goods are placed, the traversal is finished, and the candidate block section with the longest time length is selected as the block section when the goods are placed stably, wherein
ax、axAnd axThe acceleration in X, Y and Z directions when the triaxial acceleration sensor detects that the goods are initially placed is respectively, n is the number of traversal times, and b is a preset threshold value.
4. The method according to claim 3, wherein in the step S1, when the stable time zone section for determining the cargo placement cannot be obtained at the end of the traversal, the time zone section for the cargo placement stability is set to a preset value, and the preset value is related to the three-axis acceleration data sampling rate, the sampling time length and the pre-trigger time length.
5. The method of claim 1, wherein in the step S2, the average value of the triaxial acceleration is
And
said non-linear transformation being
And
6. the method of claim 1, wherein in S3, the initial placement posture of the cargo is face, edge and angle, and the labeled data is: the plane of the initial placement posture of the load is set to 0, the edge is set to 1, and the angle is set to 2.
7. The method of claim 1, wherein the preset initial placement attitude of the cargo is identified with an accuracy of more than 90% of the accurate identification of the initial placement attitude of the cargo.
8. The method of claim 1, wherein,
the input of the neural network model during training is as follows: the average value of the three-axis acceleration after nonlinear transformation and the labeled data; the output of the neural network model during training is as follows: after training, initially placing a posture recognition neural network model, wherein the neural network model at least comprises the number of network layers of the neural network model, the number of nodes on each layer and relevant parameters of the neural network;
the input of the neural network model in prediction is as follows: the average value of the three-axis acceleration after the nonlinear transformation in the section when the initial placement attitude of the goods is placed stably; the output of the neural network model in prediction is as follows: initial placement attitude of the cargo.
9. A platform for identifying an initial placement attitude of a cargo, comprising: input device, neural network model, output device, wherein,
the input device is used for acquiring initial placement attitude data of cargos;
the neural network model calculates the collected initial placement attitude data of the goods;
the output device outputs a posture report of initial placement of the cargo according to the calculation result of the neural network model,
wherein the neural network model is according to any one of claims 1-8.
10. The method of claim 9, wherein the cargo initial placement attitude report comprises a cargo initial placement attitude.
11. A method of identifying an initial placement attitude for a cargo, comprising:
data related to the initial placement posture of the cargo is sent to the platform for identifying the initial placement posture of the cargo according to any one of claims 9 or 10, and the initial placement posture of the cargo is determined according to a report output by the platform.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111044006.5A CN113836795B (en) | 2021-09-07 | 2021-09-07 | Method and platform for constructing initial goods placement attitude model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111044006.5A CN113836795B (en) | 2021-09-07 | 2021-09-07 | Method and platform for constructing initial goods placement attitude model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113836795A true CN113836795A (en) | 2021-12-24 |
| CN113836795B CN113836795B (en) | 2024-06-07 |
Family
ID=78958563
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111044006.5A Active CN113836795B (en) | 2021-09-07 | 2021-09-07 | Method and platform for constructing initial goods placement attitude model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113836795B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108664122A (en) * | 2018-04-04 | 2018-10-16 | 歌尔股份有限公司 | A kind of attitude prediction method and apparatus |
| CN112097374A (en) * | 2020-09-16 | 2020-12-18 | 珠海格力电器股份有限公司 | Device control method, device and computer readable medium |
| CN113256224A (en) * | 2021-06-22 | 2021-08-13 | 天津知途物联科技有限公司 | Impact inclination electronic tag for cargo transportation monitoring and monitoring method thereof |
-
2021
- 2021-09-07 CN CN202111044006.5A patent/CN113836795B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108664122A (en) * | 2018-04-04 | 2018-10-16 | 歌尔股份有限公司 | A kind of attitude prediction method and apparatus |
| CN112097374A (en) * | 2020-09-16 | 2020-12-18 | 珠海格力电器股份有限公司 | Device control method, device and computer readable medium |
| CN113256224A (en) * | 2021-06-22 | 2021-08-13 | 天津知途物联科技有限公司 | Impact inclination electronic tag for cargo transportation monitoring and monitoring method thereof |
Non-Patent Citations (1)
| Title |
|---|
| 李岳林等: "车用汽油机瞬态空燃比的混沌时序非线性组合辨识模型", 中国公路学报, vol. 28, no. 4, 15 April 2015 (2015-04-15), pages 109 - 115 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113836795B (en) | 2024-06-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110991311B (en) | Target detection method based on dense connection deep network | |
| US20160104058A1 (en) | Generic object detection in images | |
| CN106250838A (en) | vehicle identification method and system | |
| US20200110970A1 (en) | Parameter training method for a convolutional neural network and method for detecting items of interest visible in an image | |
| CN113850838A (en) | Ship voyage intention acquisition method and device, computer equipment and storage medium | |
| CN111091101A (en) | High-precision pedestrian detection method, system and device based on one-step method | |
| Utebayeva et al. | Multi-label UAV sound classification using Stacked Bidirectional LSTM | |
| US9177215B2 (en) | Sparse representation for dynamic sensor networks | |
| Hendrawan et al. | Improvement object detection algorithm based on YoloV5 with BottleneckCSP | |
| CN116946610B (en) | Method and device for picking up goods in intelligent warehousing system | |
| CN113836795A (en) | Method and platform for building initial cargo placement attitude model | |
| Feng et al. | A novel approach for trajectory feature representation and anomalous trajectory detection | |
| CN113792484B (en) | Cargo impact attitude model construction method and platform | |
| CN111968102A (en) | Target equipment detection method, system, medium and electronic terminal | |
| CN115298705A (en) | License plate recognition method and device, electronic equipment and storage medium | |
| CN113793097B (en) | Analysis method and system for cargo transportation risk | |
| Li et al. | Computer vision based conveyor belt congestion recognition in logistics industrial parks | |
| US20210232801A1 (en) | Model-based iterative reconstruction for fingerprint scanner | |
| CN113361811A (en) | Method, system, device and computer readable storage medium for predicting operation state | |
| CN110889453A (en) | Target detection and tracking method, device, system, medium and equipment | |
| CN114255435A (en) | Method and device for detecting abnormality of transport device, electronic apparatus, and storage medium | |
| CN113655240B (en) | Cargo falling posture recognition method and system | |
| Cai et al. | Research on pedestrian detection technology based on improved DPM model | |
| CN114359698B (en) | Sonar image recognition method and system based on bidirectional skip feedback loop structure | |
| CN108614887B (en) | Internet of things information processing method and system with noise in supply chain environment |
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 |