CN114708618A - Intelligent work clothes identification method and system for intelligent park based on classification - Google Patents

Intelligent work clothes identification method and system for intelligent park based on classification Download PDF

Info

Publication number
CN114708618A
CN114708618A CN202210423745.3A CN202210423745A CN114708618A CN 114708618 A CN114708618 A CN 114708618A CN 202210423745 A CN202210423745 A CN 202210423745A CN 114708618 A CN114708618 A CN 114708618A
Authority
CN
China
Prior art keywords
detection
data
worker
identification
intelligent
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
Application number
CN202210423745.3A
Other languages
Chinese (zh)
Inventor
邓国军
游靖
何晓明
刘宏源
吴浩
陈琳钰
杨�远
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Henan Sincerity Information Technology Co ltd
Original Assignee
Henan Sincerity Information Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Henan Sincerity Information Technology Co ltd filed Critical Henan Sincerity Information Technology Co ltd
Priority to CN202210423745.3A priority Critical patent/CN114708618A/en
Publication of CN114708618A publication Critical patent/CN114708618A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • 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
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • 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
    • 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
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Computational Linguistics (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Image Analysis (AREA)

Abstract

The invention discloses a classification-based intelligent work clothes identification method for an intelligent park, which specifically comprises the following steps: (1) collecting data; (2) training a detection/recognition model; (3) pedestrian detection, work clothes and work hat classification and identification; and discloses wisdom garden intelligence worker's clothes identification system based on classification, the system includes camera, detection module, database, detection/recognition model, management center and alarm. The invention relates to the technical field of image processing, and particularly provides a classification-based intelligent work clothes identification method and system for an intelligent park.

Description

Intelligent work clothes identification method and system for intelligent park based on classification
Technical Field
The invention relates to the technical field of image processing, in particular to a classification-based intelligent work service identification method and system for an intelligent park.
Background
The regular work clothes are worn, thereby being beneficial to training the discipline of enterprises, strengthening the culture cohesion of the enterprises, enhancing the enterprise attribution sense of employees, showing the fair and fair concept of the enterprises and building good enterprise order. The work clothes, with the standards and specifications of enterprises being gathered, the team spirit of group coordination and harmony is transmitted to the outside, and the enterprise dignity and enterprise information are transmitted. The worker clothes are an important external form of enterprise culture output, and are also an important measure for guaranteeing the production environment, protecting the personal safety of workers, distinguishing identities and improving the working efficiency. The requirement that workers correctly wear the clothing in the fields of food processing, mechanical manufacturing, financial services, public services, buildings and medical treatment is severer at present.
Currently, the main means for employee service supervision and management are mainly divided into two types: manual management and intelligent management. The manual management cost is high, the efficiency is low, and the long-time continuous efficient work is difficult. The intelligent management is mainly realized by means of the current gradually mature artificial intelligence technology, such as deep learning, kmeans and other algorithms through the retrieval technology, but the method has two main disadvantages, one is that the accuracy of the identification of the work clothes based on the retrieval needs to be improved, and on the other hand, the speed of the identification of the work clothes based on the retrieval is influenced by the retrieval library, and when the retrieval library is larger, the identification speed is reduced. In a conventional deep learning model, the model is usually trained using SoftMax loss as an objective function, which is mainly to constrain the model to learn the depth features related to the current task by optimizing the difference between the predicted probability distribution of the beam model output samples and the true sample label distribution. However, the depth features learned through the SoftMax function are relatively poor in interpretability and robustness and sensitive to outliers, so that requirements for feature engineering are more severe, which is contrary to the original intention of deep learning. Therefore, a fast, efficient, stable and low-cost method for identifying work clothes is needed.
Disclosure of Invention
Aiming at the situation and overcoming the existing defects, the invention provides an intelligent worker-uniform identification method and system for a smart park based on classification, which realize an automatic worker-uniform identification model by utilizing a deep neural network technology, carry out real-time and efficient worker-uniform identification on workers entering an operation area by combining external camera equipment, synchronize an identification result to a rear-end management platform and realize high integration of identification, alarm and management; the invention overcomes the problems of high labor consumption, poor timeliness, high management cost and the like of the traditional manual inspection of the dress of the worker clothes, and improves the level of automatic and intelligent management of enterprises.
The invention provides the following technical scheme: the invention provides a classification-based intelligent work clothes identification method for an intelligent park, which specifically comprises the following steps:
(1) data collection: acquiring original data of pedestrians through a camera, wherein the original data comprises positive sample data of normally worn work clothes and work caps and negative sample data of not worn work clothes and work caps;
(2) training a detection/recognition model:
a) the method comprises the steps that data preprocessing is carried out on original data collected by a camera through an image processing algorithm of a detection module and is stored in a database, and then a worker clothes detection/identification model is trained through the data;
b) in the training process, distance constraint between a feature space and a class center is performed by utilizing a regularized center loss increase sample after network mapping, so that intra-class similarity is enhanced, and meanwhile, the function also strictly constrains the difference between classes, i.e. heterogeneous samples are pushed away from each other, so that the initialization dependency of the loss function on the class center is reduced;
(3) pedestrian detection, worker's clothes, worker's cap classification discernment: and deploying the trained detection/identification model to a management center, processing the image frame transmitted by the camera in real time through the detection/identification model, identifying the detection/identification result, and giving an alarm for the detection/identification result which does not meet the requirement.
The invention provides a classification-based intelligent work clothes identification system for an intelligent park, which comprises a camera, a detection module, a database, a detection/identification model, a management center and an alarm, wherein the camera is used for collecting original data of pedestrians, and the original data comprises positive samples of normal work clothes and work hats and negative sample data of work clothes and work hats which are not worn; the detection module is responsible for carrying out data preprocessing on the original data acquired by the camera through an image processing algorithm; the database is used for storing data preprocessed by the detection module and performing detection/recognition model training by using the data; the detection/identification model detects the pedestrian and obtains a detection/identification result; and the management center is used for collecting the detection/identification result and controlling the alarm to give an alarm.
The invention with the structure has the following beneficial effects: the invention provides a classification-based intelligent work clothes identification method and system for an intelligent park, which is characterized in that algorithm model customization is carried out based on artificial intelligence deep learning, and large-scale work clothes data identification training is carried out by utilizing a deep learning neural network technology; the data of the field camera is matched, the dress condition of the worker in the field operation personnel is automatically identified in an end-to-end mode, real-time early warning is realized, external personnel can be effectively prevented from entering a workshop, the supervision efficiency is improved, and the supervision cost is reduced; the invention utilizes normalized Center Loss (normalized Center Loss) as a Loss function of a classification model, takes the distance from a sample to a class Center corresponding to the sample as a monitoring signal to optimize network parameters, so that the similar samples gather to the class Center in a learned feature space, and in addition, the normalized Center Loss also takes the distance between the class centers as a monitoring signal, so that the class centers are continuously pulled apart in the network training process to ensure the difference between the classes.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention. In the drawings:
fig. 1 is a block diagram of a system flow structure of an intelligent work service identification method and system for an intelligent park based on classification according to the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments; all other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It should be noted that the terms "front," "back," "left," "right," "upper" and "lower" used in the following description refer to directions in the drawings, and the terms "inner" and "outer" refer to directions toward and away from, respectively, the geometric center of a particular component.
As shown in fig. 1, the technical solution adopted by the present invention is as follows: the invention provides a classification-based intelligent work service identification method and system for an intelligent park, which specifically comprise the following steps:
(1) data collection: acquiring original data of pedestrians through a camera, wherein the original data comprises positive sample data of normally worn work clothes and work caps and negative sample data of not worn work clothes and work caps;
(2) training a detection/recognition model:
a) the method comprises the steps that data preprocessing is carried out on original data collected by a camera through an image processing algorithm of a detection module and is stored in a database, and then a worker clothes detection/identification model is trained through the data;
b) in the training process, distance constraint between a feature space and a class center is carried out by utilizing a regularized center loss increase sample after network mapping, so that intra-class similarity is enhanced, and meanwhile, the function also strictly constrains the difference between classes, namely, heterogeneous samples are pushed away from each other, so that the initialization dependency of the loss function on the class center is reduced;
(3) pedestrian detection, worker's clothes, worker's cap classification discernment: and deploying the trained detection/identification model to a management center, processing the image frame transmitted by the camera in real time through the detection/identification model, identifying the detection/identification result, and giving an alarm for the detection/identification result which does not meet the requirement.
The invention provides a classification-based intelligent work clothes identification system for an intelligent park, which comprises a camera, a detection module, a database, a detection/identification model, a management center and an alarm, wherein the camera is used for collecting original data of pedestrians, and the original data comprises positive samples of normal work clothes and work caps and negative sample data of work clothes and work caps which are not worn; the detection module is responsible for carrying out data preprocessing on the original data acquired by the camera through an image processing algorithm; the database is used for storing data preprocessed by the detection module and performing detection/recognition model training by using the data; the detection/identification model detects the pedestrian and obtains a detection/identification result; and the management center is used for collecting the detection/identification result and controlling the alarm to give an alarm.
Example (b):
1. regularized Center Loss (regulated Center Loss):
Figure BDA0003607581570000031
2. the pedestrian detection and helmet detection model is the YOLO v5 used, wherein the helmet detection is that the number of detection categories of the YOLO v5 model is modified (from 80 categories to 1 category) and trained.
3. The service identification is a ResNet50 model trained by utilizing self-owned data, and the model training is to jointly use SoftMax classification loss and regularized center loss as target loss optimization network parameters.
Figure BDA0003607581570000041
m represents the total number of samples in a training subset (Mini-batch), f (x)i)∈RdIs a d-dimensional feature vector learned by the ith sample in the training subset (Mini-batch),
Figure BDA0003607581570000042
is xiCorresponding class center, α1Is a constraint threshold for the similarity between the sample to the class center. The second term in the above equation is to constrain the spacing between different class centers to be greater than a set threshold α2Xi is a constant, λ1Is a Hyper Parameter (Hyper Parameter) that controls the contribution proportion of the inter-class constraint in the loss function.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.

Claims (2)

1. The intelligent work clothes identification method of the intelligent park based on classification is characterized by comprising the following steps:
(1) data collection: acquiring original data of pedestrians through a camera, wherein the original data comprises positive sample data of normally worn work clothes and work caps and negative sample data of not worn work clothes and work caps;
(2) training a detection/recognition model:
a) the method comprises the steps that data preprocessing is carried out on original data collected by a camera through an image processing algorithm of a detection module and is stored in a database, and then a worker clothes detection/identification model is trained through the data;
b) in the training process, distance constraint between a feature space and a class center is carried out by utilizing a regularized center loss increase sample after network mapping, so that intra-class similarity is enhanced, and meanwhile, the function also strictly constrains the difference between classes, namely, heterogeneous samples are pushed away from each other, so that the initialization dependency of the loss function on the class center is reduced;
(3) pedestrian detection, worker's clothes, worker's cap classification discernment: and deploying the trained detection/identification model to a management center, processing the image frame transmitted by the camera in real time through the detection/identification model, identifying the detection/identification result, and giving an alarm for the detection/identification result which does not meet the requirement.
2. The intelligent classification-based intelligent worker garment identification system of claim 1 comprises a camera, a detection module, a database, a detection/identification model, a management center and an alarm, wherein the camera is used for collecting raw data of pedestrians, and the raw data comprises positive sample data of normal wearing of worker garments and worker caps and negative sample data of no wearing of worker garments and worker caps; the detection module is responsible for carrying out data preprocessing on the original data acquired by the camera through an image processing algorithm; the database is used for storing data preprocessed by the detection module and performing detection/recognition model training by using the data; the detection/identification model detects the pedestrian and obtains a detection/identification result; and the management center is used for collecting the detection/identification result and controlling the alarm to give an alarm.
CN202210423745.3A 2022-04-21 2022-04-21 Intelligent work clothes identification method and system for intelligent park based on classification Pending CN114708618A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210423745.3A CN114708618A (en) 2022-04-21 2022-04-21 Intelligent work clothes identification method and system for intelligent park based on classification

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210423745.3A CN114708618A (en) 2022-04-21 2022-04-21 Intelligent work clothes identification method and system for intelligent park based on classification

Publications (1)

Publication Number Publication Date
CN114708618A true CN114708618A (en) 2022-07-05

Family

ID=82175871

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210423745.3A Pending CN114708618A (en) 2022-04-21 2022-04-21 Intelligent work clothes identification method and system for intelligent park based on classification

Country Status (1)

Country Link
CN (1) CN114708618A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116824511A (en) * 2023-08-03 2023-09-29 行为科技(北京)有限公司 Tool identification method and device based on deep learning and color space

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116824511A (en) * 2023-08-03 2023-09-29 行为科技(北京)有限公司 Tool identification method and device based on deep learning and color space

Similar Documents

Publication Publication Date Title
CN111291699B (en) Substation personnel behavior identification method based on monitoring video time sequence action positioning and abnormity detection
CN111898514B (en) Multi-target visual supervision method based on target detection and action recognition
CN109657592B (en) Face recognition method of intelligent excavator
CN110414400B (en) Automatic detection method and system for wearing of safety helmet on construction site
CN109255298A (en) Safety helmet detection method and system in dynamic background
Cao et al. CNN-based intelligent safety surveillance in green IoT applications
CN105426820A (en) Multi-person abnormal behavior detection method based on security monitoring video data
CN109241847A (en) The Oilfield Operation District safety monitoring system of view-based access control model image
CN110162968A (en) A kind of Network Intrusion Detection System based on machine learning
CN114708618A (en) Intelligent work clothes identification method and system for intelligent park based on classification
CN112435414A (en) Security monitoring system based on face recognition and monitoring method thereof
CN106991400A (en) A kind of fire hazard smoke detecting method and device
CN113191273A (en) Oil field well site video target detection and identification method and system based on neural network
CN117197726B (en) Important personnel accurate management and control system and method
Wu et al. Computer vision-based intelligent elevator information system for efficient demand-based operation and optimization
Peng et al. [Retracted] Helmet Wearing Recognition of Construction Workers Using Convolutional Neural Network
CN117014193A (en) Unknown Web attack detection method based on behavior baseline
CN116665305A (en) Method and system for detecting worker behaviors based on computer vision and knowledge graph
CN109669403B (en) Dangerous agricultural input material intelligent monitoring system based on DBN-SOFTMAX
CN115249341A (en) Intelligent security monitoring system
CN115860144A (en) Machine learning system for anti-electricity-stealing site
CN116362936A (en) Potential risk personnel identification method
CN115391151A (en) Method for intelligently discovering alarm tag based on object relationship
CN115169673A (en) Intelligent campus epidemic risk monitoring and early warning system and method
CN114937308A (en) Intelligent security system based on human body dynamic behavior feature recognition and implementation method thereof

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