CN117058526A - Automatic cargo identification method and system based on artificial intelligence - Google Patents
Automatic cargo identification method and system based on artificial intelligence Download PDFInfo
- Publication number
- CN117058526A CN117058526A CN202311309146.XA CN202311309146A CN117058526A CN 117058526 A CN117058526 A CN 117058526A CN 202311309146 A CN202311309146 A CN 202311309146A CN 117058526 A CN117058526 A CN 117058526A
- Authority
- CN
- China
- Prior art keywords
- cargo
- goods
- shelf number
- yolov5
- recognition model
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 42
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 30
- 238000012544 monitoring process Methods 0.000 claims abstract description 23
- 238000001514 detection method Methods 0.000 claims abstract description 16
- 230000008569 process Effects 0.000 claims abstract description 10
- 238000012549 training Methods 0.000 claims description 21
- 238000007781 pre-processing Methods 0.000 claims description 13
- 230000004927 fusion Effects 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 6
- 230000002708 enhancing effect Effects 0.000 claims description 5
- 238000011176 pooling Methods 0.000 claims description 5
- 230000005284 excitation Effects 0.000 claims description 4
- 238000001125 extrusion Methods 0.000 claims description 4
- 238000005457 optimization Methods 0.000 claims description 3
- 238000004806 packaging method and process Methods 0.000 claims description 2
- 230000006870 function Effects 0.000 description 21
- 230000008901 benefit Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 4
- 239000000463 material Substances 0.000 description 3
- 230000006835 compression Effects 0.000 description 2
- 238000007906 compression Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 230000003213 activating effect Effects 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000004931 aggregating effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000007670 refining Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000005728 strengthening Methods 0.000 description 1
- 230000001502 supplementing effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000000844 transformation Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/28—Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- 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
Abstract
The invention relates to the field of artificial intelligence, in particular to an automatic cargo identification method and system based on artificial intelligence. Acquiring historical cargo image data to be trained; establishing a target YOLOv5 cargo image recognition model based on a YOLOv target detection algorithm; acquiring real-time cargo image data on a goods shelf through an image acquisition device, inputting the real-time cargo image data into a target YOLOv5 cargo image recognition model for recognition, and obtaining real-time cargo types; judging whether the real-time goods are the same as the adjacent goods, if not, acquiring a first goods shelf number and a second goods shelf number, and sending the first goods shelf number and the second goods shelf number to a server for early warning; acquiring the early warning times of the first shelf number in the server, obtaining the target shelf number with the largest early warning times, and carrying out key monitoring on the target shelf number by using the image acquisition device. The commodity circulation warehouse can solve a large amount of time in the goods placement searching process, improves commodity circulation warehouse's circulation efficiency and goods handling efficiency.
Description
Technical Field
The invention relates to the field of artificial intelligence, in particular to an automatic cargo identification method and system based on artificial intelligence.
Background
The ability of deep convolutional neural networks in the artificial intelligence field to learn generalized visual features has led to tremendous success in the machine vision field. The capacity of the system can be applied to a plurality of fields, plays a role in daily life and logistics industry, in the process of circulation of a logistics warehouse, goods in the logistics warehouse are carried generally along with the situation that certain goods are placed on a goods shelf, for example, the goods are similar in length, but different in goods are placed on the same goods shelf, and because the logistics warehouse is large in area and high in logistics goods shelf position, a large amount of manpower and material resources are consumed when the misplaced goods are found, so that certain logistics efficiency circulation is low and manpower resource is wasted, and therefore, how to utilize an artificial intelligence technology to energize the goods shelf of the logistics warehouse, and the reduction of the placing error rate of the goods is a technical problem to be solved in the current stage.
Disclosure of Invention
The invention aims to solve the problems, and designs an automatic cargo identification method and system based on artificial intelligence.
The technical scheme for achieving the purpose is that in the automatic goods identification method based on artificial intelligence, the automatic goods identification method comprises the following steps:
acquiring historical cargo image data in a database, and performing data preprocessing on the historical cargo image data to obtain historical cargo image data to be trained;
based on a Yolo target detection algorithm, a Yolo v5 cargo image recognition model is established, a GIOU-LOSS LOSS function is replaced by a LOSS function in the Yolo v5 cargo image recognition model, and an initial Yolo v5 cargo image recognition model is obtained;
inputting the historical cargo image data to be trained into the initial Yolov5 cargo image recognition model for training, and adding a CBAM convolution attention module into the initial Yolov5 cargo image recognition model to obtain a target Yolov5 cargo image recognition model;
acquiring real-time cargo image data on a goods shelf through an image acquisition device, and inputting the real-time cargo image data into the target YOLOv5 cargo image recognition model for recognition to obtain a real-time cargo type;
judging whether the real-time goods are the same as the adjacent goods, if not, acquiring a first goods shelf number and a second goods shelf number, and sending the first goods shelf number and the second goods shelf number to a server for early warning;
and acquiring the early warning times of the first shelf number in the server, obtaining a target shelf number with the highest early warning times, and carrying out key monitoring on the target shelf number by using the image acquisition device.
Further, in the automatic cargo identification method, the acquiring the historical cargo image data in the database, and performing data preprocessing on the historical cargo image data to obtain historical cargo image data to be trained includes:
acquiring historical goods image data in a database, wherein the historical goods image data at least comprises goods shape image data, goods packaging image data, goods category image data, goods front image data and goods side image data;
performing data preprocessing on the historical cargo image data, and performing image enhancement processing on the historical cargo image data to obtain enhanced historical cargo image data;
performing binarization processing on the enhanced historical cargo image data to obtain binarized historical cargo image data;
and cutting the binarized historical cargo image data to obtain historical cargo image data to be trained.
Further, in the automatic cargo recognition method, the creating a YOLOv5 cargo image recognition model based on the YOLOv target detection algorithm, replacing a GIOU-LOSS function with the LOSS function in the YOLOv5 cargo image recognition model to obtain an initial YOLOv5 cargo image recognition model, including:
building a YOLO v5 cargo image recognition model based on a YOLO target detection algorithm, wherein the YOLO v5 cargo image recognition model at least comprises an input layer, a convolution layer, a full connection layer, a pooling layer and an output layer;
replacing the LOSS function in the YOLOv5 cargo image recognition model by using the GIOU-LOSS LOSS function;
enhancing an input end set as the YOLOv5 cargo image recognition model by utilizing the mosaics data;
adding a Focus structure for slicing the image data in the YOLOv5 cargo image recognition model;
and carrying out network feature fusion in the YOLOv5 cargo image recognition model by using an FPN+PAN structure to obtain an initial YOLOv5 cargo image recognition model.
Further, in the automatic cargo recognition method, the inputting the historical cargo image data to be trained into the initial YOLOv5 cargo image recognition model for training, adding a CBAM convolution attention module into the initial YOLOv5 cargo image recognition model, and obtaining a target YOLOv5 cargo image recognition model includes:
inputting the historical cargo image data to be trained into the initial YOLOv5 cargo image recognition model for training;
adding a CBAM convolution attention module in the training process of the initial Yolov5 cargo image recognition model;
and adding a SENet extrusion and excitation module in the training process of the initial Yolov5 cargo image recognition model to obtain a target Yolov5 cargo image recognition model.
Further, in the automatic cargo identification method, the acquiring real-time cargo image data on the goods shelf by the image acquisition device, inputting the real-time cargo image data to the target YOLOv5 cargo image identification model for identification, and obtaining a real-time cargo type includes:
acquiring real-time cargo image data on a goods shelf through an image acquisition device, wherein the real-time cargo image data at least comprises cargo front image data and cargo side image data;
inputting the real-time cargo image data into the target YOLOv5 cargo image recognition model for recognition to obtain real-time cargo types;
the real-time goods at least comprise daily goods, food goods, household appliances and toys.
Further, in the automatic cargo identification method, the determining whether the real-time cargo type is the same as the adjacent cargo type, if not, acquiring a first shelf number and a second shelf number, and sending the first shelf number and the second shelf number to a server for early warning, includes:
acquiring adjacent cargo image information adjacent to the real-time cargo type through an image acquisition device, and inputting the adjacent cargo image information into the target YOLOv5 cargo image recognition model for recognition to obtain the adjacent cargo type;
the adjacent goods image information at least comprises 10 adjacent goods images and also comprises adjacent goods shelf images adjacent to the real-time goods;
judging whether the real-time cargo type is the same as the adjacent cargo type, if not, acquiring a first shelf number and a second shelf number;
the first shelf number is the shelf number where the real-time goods are actually located, and the second shelf number is the shelf number where the real-time goods are fixed;
and sending the first shelf number and the second shelf number to a server for early warning, and informing the shelf placement errors of the real-time goods.
Further, in the automatic goods identification method, the acquiring the number of times that the first goods shelf number is pre-warned in the server, to obtain a target goods shelf number with the largest pre-warning number of times, and using the image acquisition device to perform key monitoring on the target goods shelf number includes:
acquiring the early warning times of the first shelf number in the server, and acquiring a target shelf number with the largest early warning times;
the target goods shelf number comprises goods shelves with goods placement error times larger than 10 times in 48 hours, goods shelves with goods placement error times larger than 5 times in 24 hours, and goods shelves with goods placement error times larger than 3 times in 12 hours;
if the target shelf number is judged, the image acquisition device is utilized to carry out key monitoring on the target shelf number;
and acquiring the goods placement error times of the target goods shelf number in real time, and if the goods placement error times are less than 3 times in 48 hours, canceling to perform key monitoring on the target goods shelf number.
Further, in the automatic cargo identification system based on artificial intelligence, the automatic cargo identification system includes:
the data acquisition module is used for acquiring historical goods image data in the database, and carrying out data preprocessing on the historical goods image data to obtain historical goods image data to be trained;
the model building module is used for building a YOLOv5 cargo image recognition model based on a YOLOv target detection algorithm, replacing a LOSS function in the YOLOv5 cargo image recognition model with a GIOU-LOSS LOSS function, and obtaining an initial YOLOv5 cargo image recognition model;
the model optimization module is used for inputting the historical cargo image data to be trained into the initial Yolov5 cargo image recognition model for training, and adding a CBAM convolution attention module into the initial Yolov5 cargo image recognition model to obtain a target Yolov5 cargo image recognition model;
the automatic identification module is used for acquiring real-time cargo image data on a goods shelf through the image acquisition device, inputting the real-time cargo image data into the target YOLOv5 cargo image identification model for identification, and obtaining real-time cargo types;
the identification judging module is used for judging whether the real-time goods are the same as the adjacent goods, if not, acquiring a first goods shelf number and a second goods shelf number, and sending the first goods shelf number and the second goods shelf number to a server for early warning;
and the key monitoring module is used for acquiring the early warning times of the first shelf number in the server, obtaining the target shelf number with the largest early warning times, and carrying out key monitoring on the target shelf number by utilizing the image acquisition device.
Further, in the automatic cargo identification system based on artificial intelligence, the model building module comprises the following submodules:
the building sub-module is used for building a YOLOv5 cargo image recognition model based on a YOLOv target detection algorithm, wherein the YOLOv5 cargo image recognition model at least comprises an input layer, a convolution layer, a full-connection layer, a pooling layer and an output layer;
a replacement sub-module, configured to replace a LOSS function in the YOLOv5 cargo image recognition model with a GIOU-LOSS function;
the setting submodule is used for enhancing the input end set as the YOLOv5 cargo image recognition model by utilizing the Mosaic data;
the slicing sub-module is used for adding a Focus structure to slice the image data in the YOLOv5 cargo image recognition model;
and the fusion sub-module is used for carrying out network feature fusion in the YOLOv5 cargo image recognition model by using the FPN+PAN structure to obtain an initial YOLOv5 cargo image recognition model.
Further, in the automatic cargo identification system based on artificial intelligence, the identification judging module includes the following submodules:
the acquisition sub-module is used for acquiring adjacent cargo image information adjacent to the real-time cargo type through the image acquisition device, inputting the adjacent cargo image information into the target YOLOv5 cargo image recognition model for recognition, and obtaining the adjacent cargo type;
the adjacent sub-module is used for the adjacent goods image information to at least comprise 10 adjacent goods images and also comprises adjacent goods shelf images adjacent to the real-time goods types;
the judging sub-module is used for judging whether the real-time cargo type is the same as the adjacent cargo type, and if the real-time cargo type is different from the adjacent cargo type, a first shelf number and a second shelf number are obtained;
the coding sub-module is used for enabling the first shelf number to be the shelf number where the real-time goods are actually located, and enabling the second shelf number to be the shelf number where the real-time goods are fixed;
and the early warning sub-module is used for sending the first shelf number and the second shelf number to a server for early warning and notifying the shelf placement errors of the real-time goods.
The method has the advantages that the historical goods image data in the database are obtained, and the historical goods image data are subjected to data preprocessing to obtain the historical goods image data to be trained; based on a Yolo target detection algorithm, a Yolo v5 cargo image recognition model is established, a GIOU-LOSS LOSS function is replaced by a LOSS function in the Yolo v5 cargo image recognition model, and an initial Yolo v5 cargo image recognition model is obtained; inputting the historical cargo image data to be trained into the initial Yolov5 cargo image recognition model for training, and adding a CBAM convolution attention module into the initial Yolov5 cargo image recognition model to obtain a target Yolov5 cargo image recognition model; acquiring real-time cargo image data on a goods shelf through an image acquisition device, and inputting the real-time cargo image data into the target YOLOv5 cargo image recognition model for recognition to obtain a real-time cargo type; judging whether the real-time goods are the same as the adjacent goods, if not, acquiring a first goods shelf number and a second goods shelf number, and sending the first goods shelf number and the second goods shelf number to a server for early warning; and acquiring the early warning times of the first shelf number in the server, obtaining a target shelf number with the highest early warning times, and carrying out key monitoring on the target shelf number by using the image acquisition device. The goods placement and searching device can solve a great amount of time in the goods placement and searching process, improves the circulation efficiency and the goods carrying efficiency of a logistics warehouse, improves the overall operation efficiency, reduces the use of manpower and material resources, monitors the goods shelf with the largest placement error times in a key manner, and reduces the goods placement error rate of the goods shelf.
Drawings
Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention.
FIG. 1 is a schematic diagram of a first embodiment of an automatic cargo identification method based on artificial intelligence according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a second embodiment of an artificial intelligence based automatic cargo identification method according to an embodiment of the invention;
FIG. 3 is a schematic diagram of a third embodiment of an automatic cargo identification method based on artificial intelligence according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of a first embodiment of an artificial intelligence based automatic cargo identification system in accordance with an embodiment of the invention.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
As used herein, the singular forms "a", "an", "the" and "the" are intended to include the plural forms as well, unless expressly stated otherwise, as understood by those skilled in the art. It will be further understood that the terms "comprises" and/or "comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
The invention is specifically described below with reference to the accompanying drawings, as shown in fig. 1, an automatic cargo identification method based on artificial intelligence, the automatic cargo identification method comprising the following steps:
step 101, acquiring historical cargo image data in a database, and performing data preprocessing on the historical cargo image data to obtain historical cargo image data to be trained;
specifically, in this embodiment, historical cargo image data in a database is obtained, where the historical cargo image data at least includes cargo shape image data, cargo package image data, cargo type image data, cargo front image data, and cargo side image data; carrying out data preprocessing on the historical cargo image data, and carrying out image enhancement processing on the historical cargo image data to obtain enhanced historical cargo image data; performing binarization processing on the enhanced historical cargo image data to obtain binarized historical cargo image data; and cutting the binarized historical cargo image data to obtain historical cargo image data to be trained.
102, building a YOLOv5 cargo image recognition model based on a YOLOv target detection algorithm, and replacing a LOSS function in the YOLOv5 cargo image recognition model with a GIOU-LOSS LOSS function to obtain an initial YOLOv5 cargo image recognition model;
specifically, in this embodiment, a YOLO 5 cargo image recognition model is established based on a YOLO target detection algorithm, where the YOLO 5 cargo image recognition model includes at least an input layer, a convolution layer, a full connection layer, a pooling layer, and an output layer; replacing the LOSS function in the YOLOv5 cargo image recognition model by the GIOU-LOSS LOSS function; enhancing an input end set as a YOLOv5 cargo image recognition model by utilizing the mosaics data; adding a Focus structure for slicing image data in the YOLOv5 cargo image recognition model; and carrying out network feature fusion in the YOLOv5 cargo image recognition model by using the FPN+PAN structure to obtain an initial YOLOv5 cargo image recognition model.
Specifically, the method also comprises the step of calculating the positioning LOSS for the complete cross ratio LOSS (Complete IOU Loss, CIOU LOSS) by using the GIOU-LOSS LOSS function;
SENet is an attention module that performs feature recognition in two ways, extrusion activation, and enhances the learning of convolution features by explicitly modeling channel interdependencies so that the network can enhance its sensitivity to informative features that can be exploited by subsequent transformations. SENet recalibrates the filter response by compression and excitation prior to conversion, compresses features found by previous convolution learning, and finally excites the features in the network. Therefore, SENet is added in the last layer of the network, and the learning and recognition of the whole network to the detail part are enhanced. In practical training, the characteristic of the detail of the cargo image such as particle size, color and the like learned in partial convolution before compression is characterized in that the last layer can be identified by activating the characteristic enhancement, so that missing of the characteristic in the last layer of training is avoided, the network learning capacity is further enhanced, and the network identification accuracy is improved.
CBAM is a module that multiplies the attention map by the input feature map for adaptive feature refinement [9 ]. CBAM consists of two sequential sub-modules: channel attention module (Channel Attention Module, CAM) and spatial attention module (Spatial Attention Module, SAM). On each convolution block of the depth network, adaptively refining the middle feature mapping through CBAM, using the average pool characteristic in the channel and the maximum pool characteristic at the same time, and further calculating and aggregating the space information of the feature mapping through a space attention module; and generating a space attention pattern in the space aspect, supplementing the attention of a channel, performing high-precision learning on the easily mixed goods images, strengthening the display detail characteristics of the tiled goods images, and better distinguishing the goods images with different grades.
Step 103, inputting historical cargo image data to be trained into an initial Yolov5 cargo image recognition model for training, and adding a CBAM convolution attention module into the initial Yolov5 cargo image recognition model to obtain a target Yolov5 cargo image recognition model;
specifically, in this embodiment, historical cargo image data to be trained is input to an initial YOLOv5 cargo image recognition model for training; adding a CBAM convolution attention module in the training process of the initial Yolov5 cargo image recognition model; and adding a SENet extrusion and excitation module in the training process of the initial Yolov5 cargo image recognition model to obtain the target Yolov5 cargo image recognition model.
104, acquiring real-time cargo image data on a goods shelf through an image acquisition device, and inputting the real-time cargo image data into a target YOLOv5 cargo image recognition model for recognition to obtain a real-time cargo type;
specifically, in this embodiment, the image acquisition device acquires real-time cargo image data on the shelf, where the real-time cargo image data at least includes cargo front image data and cargo side image data; inputting the real-time cargo image data into a target YOLOv5 cargo image recognition model for recognition to obtain real-time cargo types; the real-time goods at least comprise daily goods, food goods, household appliances and toys.
Step 105, judging whether the real-time goods are the same as the adjacent goods, if not, acquiring a first goods shelf number and a second goods shelf number, and sending the first goods shelf number and the second goods shelf number to a server for early warning;
specifically, in this embodiment, the image acquisition device acquires the image information of the adjacent cargo adjacent to the real-time cargo category, and inputs the image information of the adjacent cargo to the target YOLOv5 cargo image recognition model for recognition to obtain the adjacent cargo category; the adjacent goods image information at least comprises 10 adjacent goods images and also comprises adjacent goods shelf images adjacent to the real-time goods types; judging whether the real-time cargo type is the same as the adjacent cargo type, if not, acquiring a first shelf number and a second shelf number; the first shelf number is the shelf number where the real-time cargo type is actually located, and the second shelf number is the shelf number where the real-time cargo type is fixed; and sending the first shelf number and the second shelf number to a server for early warning, and informing the shelf placement errors of the real-time goods.
And 106, acquiring the early warning times of the first shelf number in the server, obtaining the target shelf number with the largest early warning times, and carrying out key monitoring on the target shelf number by using the image acquisition device.
Specifically, in this embodiment, the number of times of early warning of the first shelf number in the server is obtained, and the target shelf number with the largest number of early warning times is obtained; the target goods shelf number comprises goods shelves with goods placement error times more than 10 times in 48 hours, goods shelves with goods placement error times more than 5 times in 24 hours, and goods shelves with goods placement error times more than 3 times in 12 hours;
if the target goods shelf number is judged, the image acquisition device is utilized to carry out key monitoring on the target goods shelf number; and acquiring the goods placement error times of the target goods shelf number in real time, and if the goods placement error times are less than 3 times in 48 hours, canceling to perform key monitoring on the target goods shelf number.
The method has the advantages that the historical goods image data in the database are obtained, and the historical goods image data are subjected to data preprocessing to obtain the historical goods image data to be trained; based on a Yolo target detection algorithm, a Yolo v5 cargo image recognition model is established, a GIOU-LOSS LOSS function is replaced by a LOSS function in the Yolo v5 cargo image recognition model, and an initial Yolo v5 cargo image recognition model is obtained; inputting historical cargo image data to be trained into an initial Yolov5 cargo image recognition model for training, and adding a CBAM convolution attention module into the initial Yolov5 cargo image recognition model to obtain a target Yolov5 cargo image recognition model; acquiring real-time cargo image data on a goods shelf through an image acquisition device, inputting the real-time cargo image data into a target YOLOv5 cargo image recognition model for recognition, and obtaining real-time cargo types; judging whether the real-time goods are the same as the adjacent goods, if not, acquiring a first goods shelf number and a second goods shelf number, and sending the first goods shelf number and the second goods shelf number to a server for early warning; acquiring the early warning times of the first shelf number in the server, obtaining the target shelf number with the largest early warning times, and carrying out key monitoring on the target shelf number by using the image acquisition device. The goods placement and searching device can solve a great amount of time in the goods placement and searching process, improves the circulation efficiency and the goods carrying efficiency of a logistics warehouse, improves the overall operation efficiency, reduces the use of manpower and material resources, monitors the goods shelf with the largest placement error times in a key manner, and reduces the goods placement error rate of the goods shelf.
In this embodiment, referring to fig. 2, in a second embodiment of an automatic cargo identification method based on artificial intelligence in the embodiment of the present invention, whether a real-time cargo type is the same as an adjacent cargo type is determined, if not, a first shelf number and a second shelf number are obtained, and the first shelf number and the second shelf number are sent to a server for early warning, where the early warning includes the following steps:
step 201, acquiring adjacent cargo image information adjacent to the real-time cargo type through an image acquisition device, and inputting the adjacent cargo image information into a target YOLOv5 cargo image recognition model for recognition to obtain the adjacent cargo type;
step 202, the adjacent goods image information at least comprises 10 adjacent goods images and also comprises adjacent goods shelf images adjacent to the real-time goods types;
step 203, judging whether the real-time cargo type is the same as the adjacent cargo type, if not, acquiring a first shelf number and a second shelf number;
204, the first shelf number is the shelf number where the real-time cargo category is actually located, and the second shelf number is the shelf number where the real-time cargo category is fixed;
and 205, sending the first shelf number and the second shelf number to a server for early warning, and notifying the shelf placement errors of the real-time goods.
In this embodiment, referring to fig. 3, in a third embodiment of an automatic cargo identification method based on artificial intelligence in the embodiment of the present invention, the method for obtaining the number of times of early warning of a first shelf number in a server, obtaining a target shelf number with the largest number of times of early warning, and performing key monitoring on the target shelf number by using an image acquisition device includes the following steps:
step 301, acquiring the early warning times of a first shelf number in a server, and acquiring a target shelf number with the largest early warning times;
302, the target goods shelf number comprises goods shelves with goods placement error times larger than 10 times in 48 hours, goods shelves with goods placement error times larger than 5 times in 24 hours, and goods shelves with goods placement error times larger than 3 times in 12 hours;
step 303, if the target shelf number is judged, the image acquisition device is utilized to carry out key monitoring on the target shelf number;
and 304, acquiring the number of goods placement errors of the target goods shelf number in real time, and if the number of goods placement errors is less than 3 times in 48 hours, canceling to perform key monitoring on the target goods shelf number.
The description of the automatic cargo identification method based on artificial intelligence provided by the embodiment of the invention is given above, and the description of the automatic cargo identification system based on artificial intelligence provided by the embodiment of the invention is given below, referring to fig. 4, an embodiment of the automatic cargo identification system in the embodiment of the invention includes:
the data acquisition module is used for acquiring historical goods image data in the database, and carrying out data preprocessing on the historical goods image data to obtain historical goods image data to be trained;
the model building module is used for building a YOLOv5 cargo image recognition model based on a YOLOv target detection algorithm, replacing a LOSS function in the YOLOv5 cargo image recognition model with a GIOU-LOSS LOSS function, and obtaining an initial YOLOv5 cargo image recognition model;
the model optimization module is used for inputting historical cargo image data to be trained into an initial Yolov5 cargo image recognition model for training, and adding a CBAM convolution attention module into the initial Yolov5 cargo image recognition model to obtain a target Yolov5 cargo image recognition model;
the automatic identification module is used for acquiring real-time cargo image data on the goods shelf through the image acquisition device, inputting the real-time cargo image data into the target YOLOv5 cargo image identification model for identification, and obtaining real-time cargo types;
the identification judging module is used for judging whether the real-time goods are the same as the adjacent goods, if not, acquiring a first goods shelf number and a second goods shelf number, and sending the first goods shelf number and the second goods shelf number to the server for early warning;
and the key monitoring module is used for acquiring the early warning times of the first shelf number in the server, obtaining the target shelf number with the largest early warning times, and carrying out key monitoring on the target shelf number by utilizing the image acquisition device.
The foregoing has shown and described the basic principles, principal features and advantages of the invention. It will be understood by those skilled in the art that the present invention is not limited to the above-described embodiments, and that the above-described embodiments and descriptions are only preferred embodiments of the present invention, and are not intended to limit the invention, and that various changes and modifications may be made therein without departing from the spirit and scope of the invention as claimed. The scope of the invention is defined by the appended claims and equivalents thereof.
Claims (10)
1. An automatic cargo identification method based on artificial intelligence is characterized by comprising the following steps:
acquiring historical cargo image data in a database, and performing data preprocessing on the historical cargo image data to obtain historical cargo image data to be trained;
based on a Yolo target detection algorithm, a Yolo v5 cargo image recognition model is established, a GIOU-LOSS LOSS function is replaced by a LOSS function in the Yolo v5 cargo image recognition model, and an initial Yolo v5 cargo image recognition model is obtained;
inputting the historical cargo image data to be trained into the initial Yolov5 cargo image recognition model for training, and adding a CBAM convolution attention module into the initial Yolov5 cargo image recognition model to obtain a target Yolov5 cargo image recognition model;
acquiring real-time cargo image data on a goods shelf through an image acquisition device, and inputting the real-time cargo image data into the target YOLOv5 cargo image recognition model for recognition to obtain a real-time cargo type;
judging whether the real-time goods are the same as the adjacent goods, if not, acquiring a first goods shelf number and a second goods shelf number, and sending the first goods shelf number and the second goods shelf number to a server for early warning;
and acquiring the early warning times of the first shelf number in the server, obtaining a target shelf number with the highest early warning times, and carrying out key monitoring on the target shelf number by using the image acquisition device.
2. The automatic cargo identification method based on artificial intelligence as claimed in claim 1, wherein the acquiring the historical cargo image data in the database, and performing data preprocessing on the historical cargo image data to obtain the historical cargo image data to be trained, includes:
acquiring historical goods image data in a database, wherein the historical goods image data at least comprises goods shape image data, goods packaging image data, goods category image data, goods front image data and goods side image data;
performing data preprocessing on the historical cargo image data, and performing image enhancement processing on the historical cargo image data to obtain enhanced historical cargo image data;
performing binarization processing on the enhanced historical cargo image data to obtain binarized historical cargo image data;
and cutting the binarized historical cargo image data to obtain historical cargo image data to be trained.
3. The automatic cargo recognition method based on artificial intelligence of claim 1, wherein the building a YOLOv5 cargo image recognition model based on a YOLOv target detection algorithm, replacing a LOSS function in the YOLOv5 cargo image recognition model with a GIOU-LOSS function to obtain an initial YOLOv5 cargo image recognition model, comprises:
building a YOLO v5 cargo image recognition model based on a YOLO target detection algorithm, wherein the YOLO v5 cargo image recognition model at least comprises an input layer, a convolution layer, a full connection layer, a pooling layer and an output layer;
replacing the LOSS function in the YOLOv5 cargo image recognition model by using the GIOU-LOSS LOSS function;
enhancing an input end set as the YOLOv5 cargo image recognition model by utilizing the mosaics data;
adding a Focus structure for slicing the image data in the YOLOv5 cargo image recognition model;
and carrying out network feature fusion in the YOLOv5 cargo image recognition model by using an FPN+PAN structure to obtain an initial YOLOv5 cargo image recognition model.
4. The automatic cargo recognition method based on artificial intelligence according to claim 1, wherein the inputting the historical cargo image data to be trained into the initial YOLOv5 cargo image recognition model for training, adding a CBAM convolution attention module into the initial YOLOv5 cargo image recognition model, and obtaining a target YOLOv5 cargo image recognition model, comprises:
inputting the historical cargo image data to be trained into the initial YOLOv5 cargo image recognition model for training;
adding a CBAM convolution attention module in the training process of the initial Yolov5 cargo image recognition model;
and adding a SENet extrusion and excitation module in the training process of the initial Yolov5 cargo image recognition model to obtain a target Yolov5 cargo image recognition model.
5. The automatic cargo identification method based on artificial intelligence as claimed in claim 1, wherein the acquiring real-time cargo image data on a shelf by the image acquisition device, inputting the real-time cargo image data to the target YOLOv5 cargo image identification model for identification, and obtaining real-time cargo types, comprises:
acquiring real-time cargo image data on a goods shelf through an image acquisition device, wherein the real-time cargo image data at least comprises cargo front image data and cargo side image data;
inputting the real-time cargo image data into the target YOLOv5 cargo image recognition model for recognition to obtain real-time cargo types;
the real-time goods at least comprise daily goods, food goods, household appliances and toys.
6. The automatic cargo identification method based on artificial intelligence according to claim 1, wherein the determining whether the real-time cargo type is the same as the neighboring cargo type, if not, acquiring a first shelf number and a second shelf number, and sending the first shelf number and the second shelf number to a server for early warning comprises:
acquiring adjacent cargo image information adjacent to the real-time cargo type through an image acquisition device, and inputting the adjacent cargo image information into the target YOLOv5 cargo image recognition model for recognition to obtain the adjacent cargo type;
the adjacent goods image information at least comprises 10 adjacent goods images and also comprises adjacent goods shelf images adjacent to the real-time goods;
judging whether the real-time cargo type is the same as the adjacent cargo type, if not, acquiring a first shelf number and a second shelf number;
the first shelf number is the shelf number where the real-time goods are actually located, and the second shelf number is the shelf number where the real-time goods are fixed;
and sending the first shelf number and the second shelf number to a server for early warning, and informing the shelf placement errors of the real-time goods.
7. The automatic cargo identification method based on artificial intelligence according to claim 1, wherein the obtaining the number of times of early warning of the first shelf number in the server to obtain a target shelf number with the largest number of times of early warning, and the using the image acquisition device to perform key monitoring on the target shelf number comprises:
acquiring the early warning times of the first shelf number in the server, and acquiring a target shelf number with the largest early warning times;
the target goods shelf number comprises goods shelves with goods placement error times larger than 10 times in 48 hours, goods shelves with goods placement error times larger than 5 times in 24 hours, and goods shelves with goods placement error times larger than 3 times in 12 hours;
if the target shelf number is judged, the image acquisition device is utilized to carry out key monitoring on the target shelf number;
and acquiring the goods placement error times of the target goods shelf number in real time, and if the goods placement error times are less than 3 times in 48 hours, canceling to perform key monitoring on the target goods shelf number.
8. An automatic cargo identification system based on artificial intelligence, characterized in that the automatic cargo identification system comprises the following modules:
the data acquisition module is used for acquiring historical goods image data in the database, and carrying out data preprocessing on the historical goods image data to obtain historical goods image data to be trained;
the model building module is used for building a YOLOv5 cargo image recognition model based on a YOLOv target detection algorithm, replacing a LOSS function in the YOLOv5 cargo image recognition model with a GIOU-LOSS LOSS function, and obtaining an initial YOLOv5 cargo image recognition model;
the model optimization module is used for inputting the historical cargo image data to be trained into the initial Yolov5 cargo image recognition model for training, and adding a CBAM convolution attention module into the initial Yolov5 cargo image recognition model to obtain a target Yolov5 cargo image recognition model;
the automatic identification module is used for acquiring real-time cargo image data on a goods shelf through the image acquisition device, inputting the real-time cargo image data into the target YOLOv5 cargo image identification model for identification, and obtaining real-time cargo types;
the identification judging module is used for judging whether the real-time goods are the same as the adjacent goods, if not, acquiring a first goods shelf number and a second goods shelf number, and sending the first goods shelf number and the second goods shelf number to a server for early warning;
and the key monitoring module is used for acquiring the early warning times of the first shelf number in the server, obtaining the target shelf number with the largest early warning times, and carrying out key monitoring on the target shelf number by utilizing the image acquisition device.
9. The automated cargo identification system of claim 8 wherein the modeling module comprises the following sub-modules:
the building sub-module is used for building a YOLOv5 cargo image recognition model based on a YOLOv target detection algorithm, wherein the YOLOv5 cargo image recognition model at least comprises an input layer, a convolution layer, a full-connection layer, a pooling layer and an output layer;
a replacement sub-module, configured to replace a LOSS function in the YOLOv5 cargo image recognition model with a GIOU-LOSS function;
the setting submodule is used for enhancing the input end set as the YOLOv5 cargo image recognition model by utilizing the Mosaic data;
the slicing sub-module is used for adding a Focus structure to slice the image data in the YOLOv5 cargo image recognition model;
and the fusion sub-module is used for carrying out network feature fusion in the YOLOv5 cargo image recognition model by using the FPN+PAN structure to obtain an initial YOLOv5 cargo image recognition model.
10. The automated cargo identification system of claim 8 wherein the identification decision module comprises the following sub-modules:
the acquisition sub-module is used for acquiring adjacent cargo image information adjacent to the real-time cargo type through the image acquisition device, inputting the adjacent cargo image information into the target YOLOv5 cargo image recognition model for recognition, and obtaining the adjacent cargo type;
the adjacent sub-module is used for the adjacent goods image information to at least comprise 10 adjacent goods images and also comprises adjacent goods shelf images adjacent to the real-time goods types;
the judging sub-module is used for judging whether the real-time cargo type is the same as the adjacent cargo type, and if the real-time cargo type is different from the adjacent cargo type, a first shelf number and a second shelf number are obtained;
the coding sub-module is used for enabling the first shelf number to be the shelf number where the real-time goods are actually located, and enabling the second shelf number to be the shelf number where the real-time goods are fixed;
and the early warning sub-module is used for sending the first shelf number and the second shelf number to a server for early warning and notifying the shelf placement errors of the real-time goods.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311309146.XA CN117058526A (en) | 2023-10-11 | 2023-10-11 | Automatic cargo identification method and system based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311309146.XA CN117058526A (en) | 2023-10-11 | 2023-10-11 | Automatic cargo identification method and system based on artificial intelligence |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117058526A true CN117058526A (en) | 2023-11-14 |
Family
ID=88655707
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311309146.XA Pending CN117058526A (en) | 2023-10-11 | 2023-10-11 | Automatic cargo identification method and system based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117058526A (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190236531A1 (en) * | 2018-01-10 | 2019-08-01 | Trax Technologies Solutions Pte Ltd. | Comparing planogram compliance to checkout data |
| CN111428621A (en) * | 2020-03-20 | 2020-07-17 | 京东方科技集团股份有限公司 | Shelf interaction method and device and shelf |
| WO2022205329A1 (en) * | 2021-04-01 | 2022-10-06 | 京东方科技集团股份有限公司 | Object detection method, object detection apparatus, and object detection system |
| CN115273017A (en) * | 2022-04-29 | 2022-11-01 | 桂林电子科技大学 | Traffic sign detection recognition model training method and system based on Yolov5 |
| WO2022262116A1 (en) * | 2021-06-17 | 2022-12-22 | 熊丹 | Oral health management system for adjusting electric toothbrush on basis of artificial intelligence image recognition |
| CN115631460A (en) * | 2022-11-01 | 2023-01-20 | 上海师范大学 | Commodity state identification method based on improved YOLOv5 target detection |
| CN115731533A (en) * | 2022-11-29 | 2023-03-03 | 淮阴工学院 | Vehicle-mounted target detection method based on improved YOLOv5 |
| CN115829462A (en) * | 2022-11-18 | 2023-03-21 | 未来机器人（深圳）有限公司 | Goods management method, goods management system and computer readable storage medium |
| CN116630737A (en) * | 2023-04-26 | 2023-08-22 | 吉林化工学院 | Safety equipment wearing detection method based on improved YOLOv5 |
-
2023
- 2023-10-11 CN CN202311309146.XA patent/CN117058526A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190236531A1 (en) * | 2018-01-10 | 2019-08-01 | Trax Technologies Solutions Pte Ltd. | Comparing planogram compliance to checkout data |
| CN111428621A (en) * | 2020-03-20 | 2020-07-17 | 京东方科技集团股份有限公司 | Shelf interaction method and device and shelf |
| WO2022205329A1 (en) * | 2021-04-01 | 2022-10-06 | 京东方科技集团股份有限公司 | Object detection method, object detection apparatus, and object detection system |
| WO2022262116A1 (en) * | 2021-06-17 | 2022-12-22 | 熊丹 | Oral health management system for adjusting electric toothbrush on basis of artificial intelligence image recognition |
| CN115273017A (en) * | 2022-04-29 | 2022-11-01 | 桂林电子科技大学 | Traffic sign detection recognition model training method and system based on Yolov5 |
| CN115631460A (en) * | 2022-11-01 | 2023-01-20 | 上海师范大学 | Commodity state identification method based on improved YOLOv5 target detection |
| CN115829462A (en) * | 2022-11-18 | 2023-03-21 | 未来机器人（深圳）有限公司 | Goods management method, goods management system and computer readable storage medium |
| CN115731533A (en) * | 2022-11-29 | 2023-03-03 | 淮阴工学院 | Vehicle-mounted target detection method based on improved YOLOv5 |
| CN116630737A (en) * | 2023-04-26 | 2023-08-22 | 吉林化工学院 | Safety equipment wearing detection method based on improved YOLOv5 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107617573B (en) | Logistics code identification and sorting method based on multitask deep learning | |
| CN101281597B (en) | Apparatus and method for on-line real time capturing and recognizing product package pattern identification information | |
| CN109342456B (en) | Welding spot defect detection method, device and equipment and readable storage medium | |
| CN103593670B (en) | A kind of copper plate/strip detection method of surface flaw based on online limit of sequence learning machine | |
| CN102521558B (en) | Bar code identifying method and device | |
| CN107145810A (en) | A kind of comprehensive bar code identifying device and method | |
| CN109741551B (en) | Commodity identification settlement method, device and system | |
| CN109902563A (en) | A kind of multi-angle model recognizing method and system | |
| CN108508844A (en) | Transferred product system and transfer method based on plane secondary light source and machine vision | |
| US20230222643A1 (en) | Semantic deep learning and rule optimization for surface corrosion detection and evaluation | |
| CN117253024B (en) | Industrial salt quality inspection control method and system based on machine vision | |
| CN117058526A (en) | Automatic cargo identification method and system based on artificial intelligence | |
| CN117372853A (en) | Underwater target detection algorithm based on image enhancement and attention mechanism | |
| CN110555219B (en) | Three-dimensional CAD model similarity retrieval system and method based on image recognition | |
| CN116245882A (en) | Circuit board electronic element detection method and device and computer equipment | |
| CN115169375B (en) | AR and gun ball linkage-based high-level material visualization method | |
| CN113361430A (en) | Vehicle state information acquisition monitoring method and system | |
| CN116187718B (en) | Intelligent goods identification and sorting method and system based on computer vision | |
| CN111612481A (en) | Production and sales feedback management system and method | |
| CN113902803B (en) | Method for acquiring double-label welding positions of end faces of bundled bars based on image processing | |
| CN110334228A (en) | A kind of Internet Problems map screening method based on deep learning | |
| CN115100642B (en) | 3D library position state detection method and system based on deep learning | |
| CN116993165B (en) | Safety evaluation and risk prediction method and system for fruit and vegetable juice of children | |
| CN117475240A (en) | Vegetable checking method and system based on image recognition | |
| CN113011437B (en) | Product traceability system based on industrial Internet |
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 |