CN116012715A - Method and system for synchronizing library position states based on monitoring camera - Google Patents
Method and system for synchronizing library position states based on monitoring camera Download PDFInfo
- Publication number
- CN116012715A CN116012715A CN202310042096.7A CN202310042096A CN116012715A CN 116012715 A CN116012715 A CN 116012715A CN 202310042096 A CN202310042096 A CN 202310042096A CN 116012715 A CN116012715 A CN 116012715A
- Authority
- CN
- China
- Prior art keywords
- library
- image
- model
- monitoring
- module
- 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
- 238000012544 monitoring process Methods 0.000 title claims abstract description 99
- 238000000034 method Methods 0.000 title claims abstract description 29
- 238000012549 training Methods 0.000 claims abstract description 44
- 230000001360 synchronised effect Effects 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000002457 bidirectional effect Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
Images
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a method and a system for synchronizing a library position state based on a monitoring camera, wherein the method comprises the following steps: acquiring a monitoring image, and extracting a plurality of library-position ROI images from the monitoring image; respectively adjusting brightness and contrast parameters of a plurality of library-position ROI images to a first preset value and a second preset value as initial samples; dividing an initial sample into a floor item library position ROI image and a non-floor item library position ROI image, and training the initial sample by using a depth model to obtain a model for completing training; inputting the monitoring image to be tested into the model which is trained by utilizing the model which is trained, and obtaining a classification result of the library position ROI image in the monitoring image; and obtaining the library state of the library ROI of the current monitoring area according to the classification result. The invention solves the problems that the prior manual synchronization of the state of the library is complicated, the goods are put in and put out, and the error is easy to occur in the manual synchronization of the library.
Description
Technical Field
The invention relates to the technical field of monitoring camera synchronous bin position states, in particular to a method and a system based on the monitoring camera synchronous bin position states.
Background
At present, most intelligent warehouse storage areas have the state that an AGV scheduling system and field personnel jointly maintain the warehouse positions, the state of the warehouse positions is classified as whether the warehouse positions are in goods or not, particularly when the field personnel go out and put in the warehouse, a PDA or a tablet is needed, the step of manually updating the state of the warehouse positions is tedious, at least 10 steps of operation are needed, and the manual out and put in the warehouse can be completed. Especially, when using a PDA or a tablet to select a library bit that needs to be operated, a special tag for identifying the current library bit, such as a two-dimension code, is needed to achieve the purpose of library bit selection. In long-time work, the problem that the warehouse position is inconsistent with the goods can not be avoided, and the normal operation of the AGV dispatching system is affected.
Disclosure of Invention
Aiming at the defects, the invention provides a method and a system for synchronizing the state of a library based on a monitoring camera, which aim to solve the problems that the steps of the existing manual synchronization of the state of the library are complicated, the goods are lowered in the warehouse-in and warehouse-out process, and the manual synchronization of the state of the library is easy to make mistakes.
To achieve the purpose, the invention adopts the following technical scheme:
a method for synchronizing a library position state based on a monitoring camera comprises the following steps:
step S1: acquiring a monitoring image, and extracting a plurality of library-position ROI images from the monitoring image;
s2, respectively adjusting brightness and contrast parameters of a plurality of library position ROI images to a first preset value and a second preset value as initial samples;
step S3: dividing the initial sample into a floor item library position ROI image and a non-floor item library position ROI image, and training the initial sample by using a depth model to obtain a model for completing training;
step S4: inputting a monitoring image to be tested into the training-completed model by utilizing the training-completed model to obtain a classification result of the library position ROI image in the monitoring image;
step S5: and obtaining the library state of the library ROI of the current monitoring area according to the classification result.
Preferably, in step S1, before the monitoring image is acquired, the bin ROI of the monitoring area is calibrated and the corresponding bin name is configured.
Preferably, in step S1, the monitoring image is acquired in real time through the RTSP protocol.
Preferably, in step S3, the following steps are specifically included:
step S31: inputting the depth model, training the initial sample, and obtaining the image classification accuracy of the depth model through calculation;
step S32: judging whether the image classification accuracy of the depth model is larger than a third preset value, and if so, outputting the depth model.
Another aspect of the present application provides a system for synchronizing a library position state based on a monitoring camera, the system comprising:
the acquisition module is used for acquiring a monitoring image and extracting a plurality of library position ROI images from the monitoring image;
the adjusting module is used for respectively adjusting the brightness and contrast parameters of the plurality of library position ROI images into a first preset value and a second preset value as initial samples;
the model training module is used for dividing the initial sample into a floor item library position ROI image and a non-floor item library position ROI image, and training the initial sample by using a depth model to obtain a model which is completed to be trained;
the model application module is used for inputting the monitoring image to be tested into the model after training by utilizing the model after training to obtain a classification result of the library position ROI image in the monitoring image;
and the library state acquisition module is used for acquiring the library state of the library ROI of the current monitoring area according to the classification result.
Preferably, the system further comprises a marking module, wherein the marking module is used for marking the library position ROI of the monitoring area and configuring the corresponding library position name.
Preferably, the acquiring module comprises an acquiring sub-module, and the acquiring sub-module is used for acquiring the monitoring image in real time through an RTSP protocol.
Preferably, the model training module comprises
An input sub-module for inputting the depth model;
the training sub-module is used for training the initial sample;
the computing sub-module is used for obtaining the image classification accuracy of the depth model through computing;
the judging sub-module is used for judging whether the image classification accuracy of the depth model is larger than a third preset value or not;
and the output sub-module is used for outputting the depth model if the image classification accuracy of the depth model is larger than a third preset value.
The technical scheme provided by the embodiment of the application can comprise the following beneficial effects:
according to the scheme, a monitoring camera is used for shooting a monitoring image, a Farser (bin monitoring system) is used for collecting a plurality of Zhang Kuwei ROI images in the monitoring image, the adjusted plurality of bin ROI images are used as initial samples, a depth model is used for training the initial samples, a training model is obtained, the training model is applied, the bin ROI images in the monitoring image are subjected to image classification in real time, and classification results are synchronized to Lothar (bin management and AGV scheduling system) in real time. The traditional method for manually synchronizing the state of the warehouse has complicated steps, and the method for automatically synchronizing the state of the warehouse is adopted, so that the complicated steps can be reduced, the efficiency of delivering and delivering cargoes into the warehouse can be effectively improved, the problem that the manual synchronization of the warehouse is easy to make mistakes is avoided, and the AGV scheduling system is influenced.
Drawings
FIG. 1 is a flow chart of steps of a method for synchronizing a library bit state based on a surveillance camera.
Detailed Description
Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to like or similar elements or elements having like or similar functions throughout. The embodiments described below by referring to the drawings are exemplary only for explaining the present invention and are not to be construed as limiting the present invention.
A method for synchronizing a library position state based on a monitoring camera comprises the following steps:
step S1: acquiring a monitoring image, and extracting a plurality of library-position ROI images from the monitoring image;
s2, respectively adjusting brightness and contrast parameters of a plurality of library position ROI images to a first preset value and a second preset value as initial samples;
step S3: dividing the initial sample into a floor item library position ROI image and a non-floor item library position ROI image, and training the initial sample by using a depth model to obtain a model for completing training;
step S4: inputting a monitoring image to be tested into the training-completed model by utilizing the training-completed model to obtain a classification result of the library position ROI image in the monitoring image;
step S5: and obtaining the library state of the library ROI of the current monitoring area according to the classification result.
According to the method for synchronizing the library position states based on the monitoring camera, as shown in fig. 1, the method is jointly realized by the monitoring camera, a Farser (library position monitoring system) and a Lothar (library position management and AGV scheduling system). The first step is to acquire a monitoring image, extract a plurality of library position ROI images from the monitoring image as initial samples, and in this embodiment, the monitoring image is shot by a monitoring camera, and the library position monitoring system randomly samples 150 Zhang Kuwei ROI images from the monitoring image every hour for 24 hours. The second step is to adjust the brightness and contrast parameters of the plurality of library ROI images to a first preset value and a second preset value respectively, and as an initial sample, the first preset value is 0.5 and the second preset value is 0.1 in this embodiment, so that the purpose of enhancing the library ROI images can be achieved. The third step is to divide the initial sample into a floor item library ROI image and a non-floor item library ROI image, train the initial sample with a depth model to obtain a trained model, and the depth model used in this embodiment is an open-source depth model google net (parallel connected network), which is better for training a new initial sample than the former convolutional neural network structure, except that the depth is extended, the width of the network is extended. And the fourth step is to input the monitoring image to be tested into the training-completed model by utilizing the training-completed model to obtain the classification result of the library ROI image in the monitoring image, wherein the classification result of the library ROI image in the monitoring image is divided into AGV, FLOOR and GOODS (GOODS), and the Farser (library monitoring system) synchronizes the classification result to Lothar (library management and AGV scheduling system) in real time. And fifthly, obtaining the bin state of the bin ROI of the current monitoring area according to the classification result, and generating a picking task of the current bin ROI by Lothar (bin management and AGV dispatching system) when the classification result is GOODS (GOODS).
According to the scheme, a monitoring camera is used for shooting a monitoring image, a Farser (bin monitoring system) is used for collecting a plurality of Zhang Kuwei ROI images in the monitoring image, the adjusted plurality of bin ROI images are used as initial samples, a depth model is used for training the initial samples, a training model is obtained, the training model is applied, the bin ROI images in the monitoring image are subjected to image classification in real time, and classification results are synchronized to Lothar (bin management and AGV scheduling system) in real time. The traditional method for manually synchronizing the state of the warehouse has complicated steps, and the method for automatically synchronizing the state of the warehouse is adopted, so that the complicated steps can be reduced, the efficiency of delivering and delivering cargoes into the warehouse can be effectively improved, the problem that the manual synchronization of the warehouse is easy to make mistakes is avoided, and the AGV scheduling system is influenced.
Preferably, in step S1, before the monitoring image is acquired, the library ROI of the monitoring area is calibrated and the corresponding library name is configured. In this embodiment, the calibration of the bin ROI of the monitoring area is advantageous to enable the farser (bin monitoring system) to identify and classify the bin ROI image in the monitored image more quickly, and the configuration of the corresponding bin names is advantageous to better distinguish the category of each bin.
Preferably, in step S1, the monitoring image is acquired in real time through the RTSP protocol. In this embodiment, the RTSP protocol is a real-time streaming protocol, which is a relatively more protocol for security devices, and can transmit audio and video streams in one-to-many manner, support bidirectional transmission, and have high tolerance to network delay.
Preferably, in step S3, the method specifically includes the following steps:
step S31: inputting the depth model, training the initial sample, and obtaining the image classification accuracy of the depth model through calculation;
step S32: judging whether the image classification accuracy of the depth model is larger than a third preset value, and if so, outputting the depth model.
In this embodiment, the output depth model is a model after training, and the third preset value is 0.95, and when the image classification accuracy of the depth model is greater than 0.95, the reliability of the depth model is ensured, so that the accuracy of classification of the library ROI image in the monitoring image can be effectively ensured.
Another aspect of the present application provides a system for synchronizing a library position state based on a monitoring camera, the system comprising:
the acquisition module is used for acquiring a monitoring image and extracting a plurality of library position ROI images from the monitoring image;
the adjusting module is used for respectively adjusting the brightness and contrast parameters of the plurality of library position ROI images into a first preset value and a second preset value as initial samples;
the model training module is used for dividing the initial sample into a floor item library position ROI image and a non-floor item library position ROI image, and training the initial sample by using a depth model to obtain a model which is completed to be trained;
the model application module is used for inputting the monitoring image to be tested into the model after training by utilizing the model after training to obtain a classification result of the library position ROI image in the monitoring image;
and the library state acquisition module is used for acquiring the library state of the library ROI of the current monitoring area according to the classification result.
According to the system based on the monitoring camera synchronous library position state, the automatic synchronization of the library position is realized under the combined action of the acquisition module, the adjustment module, the model training module, the model application module and the library position state acquisition module. According to the scheme, a monitoring camera is used for shooting a monitoring image, a Farser (library monitoring system) is used for collecting a plurality of Zhang Kuwei ROI images in the monitoring image, brightness and contrast parameters of the plurality of Zhang Kuwei ROI images are respectively adjusted to be a first preset value and a second preset value, the first preset value and the second preset value in the embodiment are respectively 0.5 and 0.1, the adjusted plurality of Zhang Kuwei ROI images are used as initial samples, a depth model is used for training the initial samples and obtaining a training completed model, the training completed model is used for carrying out image classification on the library ROI images in the monitoring image in real time, and classification results are synchronized to Lothar (library management and AGV scheduling system) in real time. The traditional method for manually synchronizing the state of the warehouse has complicated steps, and the method for automatically synchronizing the state of the warehouse is adopted, so that the complicated steps can be reduced, the efficiency of delivering and delivering cargoes into the warehouse can be effectively improved, the problem that the manual synchronization of the warehouse is easy to make mistakes is avoided, and the AGV scheduling system is influenced.
Preferably, the system further comprises a marking module, wherein the marking module is used for marking the library position ROI of the monitoring area and configuring the corresponding library position name. In this embodiment, the calibration of the bin ROI of the monitoring area is advantageous to enable the farser (bin monitoring system) to identify and classify the bin ROI image in the monitored image more quickly, and the configuration of the corresponding bin names is advantageous to better distinguish the category of each bin.
Preferably, the obtaining module comprises an obtaining sub-module, and the obtaining sub-module is used for obtaining the monitoring image in real time through an RTSP protocol. In this embodiment, the monitoring image is obtained in real time by the RTSP protocol in the obtaining sub-module, where the RTSP protocol is a real-time streaming protocol, is a relatively more protocol for security devices, and can transmit audio and video streams one to many, support bidirectional transmission, and have high tolerance to network delay.
Preferably, the model training module comprises
An input sub-module for inputting the depth model;
the training sub-module is used for training the initial sample;
the computing sub-module is used for obtaining the image classification accuracy of the depth model through computing;
the judging sub-module is used for judging whether the image classification accuracy of the depth model is larger than a third preset value or not;
and the output sub-module is used for outputting the depth model if the image classification accuracy of the depth model is larger than a third preset value.
In this embodiment, the output depth model is a trained model, and the third preset value is 0.95, and when the accuracy of image classification of the depth model is greater than 0.95, the reliability of the depth model is ensured, so that the accuracy of classification of the library ROI image in the monitored image can be effectively ensured.
Furthermore, functional units in various embodiments of the present invention may be integrated into one processing module, or each unit may exist alone physically, or two or more units may be integrated into one module. The integrated modules may be implemented in hardware or in software functional modules. The integrated modules may also be stored in a computer readable storage medium if implemented in the form of software functional modules and sold or used as a stand-alone product.
While embodiments of the present invention have been shown and described above, it will be understood that the above embodiments are illustrative and not to be construed as limiting the invention, and that variations, modifications, alternatives and variations of the above embodiments may be made by those skilled in the art within the scope of the invention.
Claims (8)
1. A method for synchronizing a library position state based on a monitoring camera is characterized by comprising the following steps of: the method comprises the following steps:
step S1: acquiring a monitoring image, and extracting a plurality of library-position ROI images from the monitoring image;
s2, respectively adjusting brightness and contrast parameters of a plurality of library position ROI images to a first preset value and a second preset value as initial samples;
step S3: dividing the initial sample into a floor item library position ROI image and a non-floor item library position ROI image, and training the initial sample by using a depth model to obtain a model for completing training;
step S4: inputting a monitoring image to be tested into the training-completed model by utilizing the training-completed model to obtain a classification result of the library position ROI image in the monitoring image;
step S5: and obtaining the library state of the library ROI of the current monitoring area according to the classification result.
2. The method for synchronizing the library bit states based on the monitoring camera according to claim 1, wherein the method comprises the following steps: in step S1, before the monitor image is acquired, the library ROI of the monitor area is calibrated and the corresponding library name is configured.
3. The method for synchronizing the library bit states based on the monitoring camera according to claim 1, wherein the method comprises the following steps: in step S1, the monitoring image is acquired in real time through the RTSP protocol.
4. The method for synchronizing the library bit states based on the monitoring camera according to claim 1, wherein the method comprises the following steps: in step S3, the method specifically includes the following steps:
step S31: inputting the depth model, training the initial sample, and obtaining the image classification accuracy of the depth model through calculation;
step S32: judging whether the image classification accuracy of the depth model is larger than a third preset value, and if so, outputting the depth model.
5. A system based on synchronous bin position state of surveillance camera head, characterized by: a method for synchronizing a library state based on a monitoring camera according to any one of claims 1-4, the system comprising:
the acquisition module is used for acquiring a monitoring image and extracting a plurality of library position ROI images from the monitoring image;
the adjusting module is used for respectively adjusting the brightness and contrast parameters of the plurality of library position ROI images into a first preset value and a second preset value as initial samples;
the model training module is used for dividing the initial sample into a floor item library position ROI image and a non-floor item library position ROI image, and training the initial sample by using a depth model to obtain a model which is completed to be trained;
the model application module is used for inputting the monitoring image to be tested into the model after training by utilizing the model after training to obtain a classification result of the library position ROI image in the monitoring image;
and the library state acquisition module is used for acquiring the library state of the library ROI of the current monitoring area according to the classification result.
6. The system for synchronizing the status of a library based on a surveillance camera of claim 5, wherein: the system also comprises a marking module, wherein the marking module is used for marking the library position ROI of the monitoring area and configuring the corresponding library position name.
7. The system for synchronizing the status of a library based on a surveillance camera of claim 5, wherein: the acquisition module comprises an acquisition sub-module, and the acquisition sub-module is used for acquiring the monitoring image in real time through an RTSP protocol.
8. The system for synchronizing the status of a library based on a surveillance camera of claim 5, wherein: the model training module comprises
An input sub-module for inputting the depth model;
the training sub-module is used for training the initial sample;
the computing sub-module is used for obtaining the image classification accuracy of the depth model through computing;
the judging sub-module is used for judging whether the image classification accuracy of the depth model is larger than a third preset value or not;
and the output sub-module is used for outputting the depth model if the image classification accuracy of the depth model is larger than a third preset value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310042096.7A CN116012715A (en) | 2023-01-12 | 2023-01-12 | Method and system for synchronizing library position states based on monitoring camera |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310042096.7A CN116012715A (en) | 2023-01-12 | 2023-01-12 | Method and system for synchronizing library position states based on monitoring camera |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116012715A true CN116012715A (en) | 2023-04-25 |
Family
ID=86037760
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310042096.7A Pending CN116012715A (en) | 2023-01-12 | 2023-01-12 | Method and system for synchronizing library position states based on monitoring camera |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116012715A (en) |
-
2023
- 2023-01-12 CN CN202310042096.7A patent/CN116012715A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11113512B2 (en) | Attendance monitoring method, system and apparatus for teacher during class | |
| CN109308460B (en) | Article detection method, system and computer readable storage medium | |
| US8326043B2 (en) | Video detection system and methods | |
| CN107483887A (en) | The early-warning detection method of emergency case in a kind of smart city video monitoring | |
| US10971192B2 (en) | Methods and systems for detection of anomalous motion in a video stream and for creating a video summary | |
| CN110290346B (en) | Bidding video acquisition method based on intelligent video analysis | |
| CN110096945B (en) | Indoor monitoring video key frame real-time extraction method based on machine learning | |
| CN110569804B (en) | Failure scene judging method and system based on text data, server and medium | |
| DE112019004488T5 (en) | DOUBLE-MODE DATA ACQUISITION SYSTEM FOR COLLISION DETECTION AND OBJECT DIMENSIONING | |
| CN105867326A (en) | Forming machine process quality parameter tracking system and tracking method thereof | |
| CN113516102A (en) | Deep learning parabolic behavior detection method based on video | |
| CN105407321A (en) | Monitoring data processing method and system | |
| CN114943923B (en) | Method and system for recognizing explosion flare smoke of cannonball based on video of deep learning | |
| CN111435435B (en) | Companion identification method, device, server and system | |
| CN106781167B (en) | Method and device for monitoring motion state of object | |
| CN113190700B (en) | Face snapshot, screening and storage method and system for real estate transaction | |
| JP7502570B2 (en) | Liquor product positioning method, liquor product information management method, and apparatus, device, and storage medium thereof | |
| CN116012715A (en) | Method and system for synchronizing library position states based on monitoring camera | |
| CN110019953A (en) | A kind of real-time quick look system of payload image data | |
| CN116778357A (en) | Power line unmanned aerial vehicle inspection method and system utilizing visible light defect identification | |
| WO2023280117A1 (en) | Indication signal recognition method and device, and computer storage medium | |
| CN115526859A (en) | Method for identifying production defects, distributed processing platform, equipment and storage medium | |
| CN114241190A (en) | Intelligent control system and method for chip production | |
| CN111666439A (en) | Working method for rapidly extracting and dividing medical image big data aiming at cloud environment | |
| CN112036214A (en) | Method for identifying small target in low-image-quality video in real time |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20231206 Address after: 310000, 1st Floor, Main Building, No. 2 Tushan 1st Lane, Shangcheng District, Hangzhou City, Zhejiang Province Applicant after: Hangzhou Tusk Robotics Co.,Ltd. Address before: 528000 Floor 2, A1 Building, Xingzhongchuang Xiabei Fan Industrial Center, Guicheng Street, Nanhai District, Foshan City, Guangdong Province Applicant before: Guangdong tusk robot Co.,Ltd. |
|
| TA01 | Transfer of patent application right |