CN116959117A - Monitoring system for garbage throwing behavior analysis and illegal behavior multi-angle snapshot - Google Patents
Monitoring system for garbage throwing behavior analysis and illegal behavior multi-angle snapshot Download PDFInfo
- Publication number
- CN116959117A CN116959117A CN202310970135.XA CN202310970135A CN116959117A CN 116959117 A CN116959117 A CN 116959117A CN 202310970135 A CN202310970135 A CN 202310970135A CN 116959117 A CN116959117 A CN 116959117A
- Authority
- CN
- China
- Prior art keywords
- behavior
- data
- module
- face
- analysis
- 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
- 239000010813 municipal solid waste Substances 0.000 title claims abstract description 70
- 238000012544 monitoring process Methods 0.000 title claims abstract description 66
- 238000004458 analytical method Methods 0.000 title claims abstract description 43
- 230000006399 behavior Effects 0.000 claims abstract description 228
- 206010000117 Abnormal behaviour Diseases 0.000 claims abstract description 41
- 238000012545 processing Methods 0.000 claims abstract description 22
- 238000007405 data analysis Methods 0.000 claims abstract description 10
- 238000011156 evaluation Methods 0.000 claims description 28
- 238000012549 training Methods 0.000 claims description 21
- 239000011159 matrix material Substances 0.000 claims description 19
- 238000004422 calculation algorithm Methods 0.000 claims description 16
- 238000012795 verification Methods 0.000 claims description 12
- 238000005457 optimization Methods 0.000 claims description 11
- 230000009471 action Effects 0.000 claims description 9
- 238000000034 method Methods 0.000 claims description 9
- 230000008569 process Effects 0.000 claims description 8
- 230000001815 facial effect Effects 0.000 claims description 7
- 238000002372 labelling Methods 0.000 claims description 6
- 238000007477 logistic regression Methods 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 5
- 230000002159 abnormal effect Effects 0.000 claims description 4
- 238000005516 engineering process Methods 0.000 claims description 4
- 238000010801 machine learning Methods 0.000 claims description 4
- 238000013527 convolutional neural network Methods 0.000 claims description 3
- 238000010219 correlation analysis Methods 0.000 claims description 3
- 238000013075 data extraction Methods 0.000 claims description 3
- 238000013500 data storage Methods 0.000 claims description 3
- 238000004806 packaging method and process Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000012216 screening Methods 0.000 claims description 3
- 238000004088 simulation Methods 0.000 claims description 3
- 238000003860 storage Methods 0.000 claims description 3
- 238000001514 detection method Methods 0.000 abstract description 6
- 230000006870 function Effects 0.000 abstract description 4
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 238000007726 management method Methods 0.000 description 6
- 238000007635 classification algorithm Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 239000010806 kitchen waste Substances 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/20—Movements or behaviour, e.g. gesture recognition
-
- 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
- G06V10/40—Extraction of image or video features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
- G06V10/765—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects using rules for classification or partitioning the feature space
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/44—Event detection
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
- G06V40/173—Classification, e.g. identification face re-identification, e.g. recognising unknown faces across different face tracks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Multimedia (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Human Computer Interaction (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Biomedical Technology (AREA)
- Medical Informatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Databases & Information Systems (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Social Psychology (AREA)
- Psychiatry (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a monitoring system for garbage throwing behavior analysis and illegal behavior multi-angle snapshot, and relates to the technical field of artificial intelligence Internet of things. In order to solve the problems that in the prior art, the snapshot angle is locally limited and the face cannot be accurately positioned; the monitoring system for the garbage throwing behavior analysis and the illegal behavior multi-angle snapshot comprises a data acquisition unit, a data processing unit, a data analysis unit and an intelligent snapshot unit; the automatic snapshot information is automatically identified through the camera, the functions of automatic snapshot, identification, warehousing, inquiring and comparison are realized by combining the face information and the time information, the automatic detection of people, objects and the like is realized by setting different detection areas and rules, so that the monitoring efficiency and accuracy are improved, the deduction and analysis of the behaviors of the target object are realized by carrying out three-dimensional modeling on the collected video data, the intelligent snapshot is carried out on the abnormal behaviors, and the intelligent snapshot device has the advantages of high precision, high robustness and high-level identification accuracy.
Description
Technical Field
The invention relates to the technical field of artificial intelligence Internet of things, in particular to a monitoring system for garbage throwing behavior analysis and illegal behavior multi-angle snapshot.
Background
The daily production of urban community garbage is large, garbage mixing and garbage disposal are extremely difficult, part of cities already adopt a manual persuasion mode to guide people to carry out classified garbage throwing, the manual persuasion mode has the problems of high cost, low efficiency and the like, throwing behaviors which do not meet the classified throwing requirements are automatically dissuaded, and throwing people refused to carry out uncorrectable legal disposal. An intelligent system for garbage throwing is disclosed, which contributes a force for building a clean and tidy urban environment, and related patents exist; for example, chinese patent publication No. CN114715562a discloses a method for identifying kitchen waste illegal throwing behavior, which comprises the following steps: detecting pedestrians from video frames; detecting that a garbage bag is arranged on the hand of a pedestrian, and tracking the pedestrian until the pedestrian reaches a kitchen garbage can area; the pedestrian is identified by dumping action; after detecting effective dumping action recognition, continuing to track pedestrians until the pedestrians leave the kitchen garbage can area; judging whether a garbage bag exists on the hands of the pedestrians again; if the pedestrian has a garbage bag on the hand, judging that the kitchen garbage throwing behavior of the pedestrian is legal, otherwise, judging that the kitchen garbage throwing behavior of the pedestrian is illegal. The garbage bag is arranged on the hands of the pedestrians, the pedestrians are tracked until reaching the kitchen garbage bin area to capture the object for throwing kitchen garbage, the illegal throwing behavior of the kitchen garbage throwing of the pedestrians is automatically recognized, and the environmental awareness and the resource utilization rate of community residents are improved.
The above patent, although automatically recognizing the delivery behavior, still has the following problems:
in the prior art, the snapshot angle is locally limited, the face can not be accurately positioned, the related data and the behavior data corresponding to the person can not be obtained based on the face recognition, the abnormal behavior can be estimated based on the behavior habit of each person, the conditions of low accuracy of the recognition result and large error are caused, and the abnormal behaviors of the person can not be classified.
Disclosure of Invention
The invention aims to provide a monitoring system for garbage throwing behavior analysis and illegal behavior multi-angle snapshot, which is used for analyzing target actions and behavior characteristics by processing time sequence images and realizing behavior recognition and analysis of targets based on model training and recognition so as to solve the problems in the background technology.
In order to achieve the above purpose, the present invention provides the following technical solutions:
monitoring system to rubbish input action analysis and violation action multi-angle snapshot includes:
the data acquisition unit is used for acquiring video data of the garbage dispensing point actively uploaded by the data acquisition terminal, and acquiring sub-data carried by the video data, including the coordinate data of the garbage dispensing point and the basic data of the data acquisition terminal;
packaging the video data and the sub-data of the garbage dispensing point to generate a monitoring data set, and transmitting the monitoring data set to a data processing unit for data processing based on the Internet of things;
the data processing unit is used for storing the monitoring data set, identifying the face characteristics in the video data based on the face recognition technology, inputting the identified face into a pre-stored face database for information acquisition and comparison, and determining the face recognition result, wherein the pre-stored face database comprises: face image data, character information corresponding to the face image, historical behavior information and evaluation data;
the data analysis unit is used for acquiring a character behavior video data segment of the target object based on the face recognition result, extracting behavior characteristics in the video data segment, modeling and analyzing the behavior characteristics through a machine learning algorithm, and determining a behavior analysis result of the target object;
and the intelligent snapshot unit is used for predicting the analysis result of the target object based on the historical behavior information and the evaluation data, judging the abnormal behavior in the analysis result, performing intelligent snapshot, inputting an intelligent snapshot image into the pre-stored face database for storage, and establishing an association relation with corresponding face image data based on the face recognition result.
Further, the data acquisition terminal comprises an automatic rotation driving base and a camera, a rotation shaft is embedded in the automatic rotation driving base, an automatic reset module is installed on one side of the rotation shaft, the camera is electrically connected with an algorithm box through a wire, the algorithm box comprises an intelligent terminal analysis module and a cloud decision module, and the camera is provided with a shooting assembly, a position adjusting group and an angle adjusting assembly.
Further, the data processing unit includes:
the data storage module is used for acquiring the monitoring data set uploaded by the data processing unit, extracting the garbage dispensing point coordinate data in the monitoring data set, identifying a data vector space corresponding to the garbage dispensing point coordinate data, and storing video data in the monitoring data set into the data vector space;
the video preprocessing module is used for extracting face sample data characteristics from video data of the monitoring data set, determining a target monitoring data segment with the face data characteristics based on the face sample data characteristics, and integrating the target monitoring data segment to generate a target monitoring data packet;
the face recognition module is used for extracting the characteristics of face data in the target monitoring data packet based on the convolutional neural network algorithm training model, inputting the face characteristics into a pre-stored face database for one-to-one matching, and obtaining a recognition result;
the face recognition module is also used for determining the face direction of the human body and the corresponding human body position data based on the obtained face characteristics, and matching the corresponding control instructions to control the data acquisition terminal to adjust the shooting angle to the target position;
and the information matching module is used for acquiring the character information, the historical behavior information and the evaluation data corresponding to the face image based on the identification result.
Further, the face recognition module inputs face features into a pre-stored face database for one-to-one matching, and the face recognition module comprises;
carrying out standardization processing on the image corresponding to the facial features of the face to obtain a standardized image;
acquiring a characteristic value of each pixel point in a standardized image, and constructing a first matrix based on the characteristic value of each pixel point;
each pre-stored face image in the pre-stored face database corresponds to a second matrix, wherein the second matrix corresponds to the pre-stored face image one by one;
and respectively calculating the similarity of the first matrix and each second matrix in a pre-stored face database, extracting a corresponding pre-stored face image after the similarity is higher than a preset threshold value, and taking the corresponding pre-stored face image as a recognition result.
Further, the data analysis unit includes:
the feature extraction module is used for extracting behavior feature data of the same person from the target monitoring data packet based on the identification result and acquiring a behavior feature data list of the same person based on the time sequence;
and the three-dimensional modeling module is used for constructing a three-dimensional model corresponding to each person based on the video data, and carrying out three-dimensional simulation on the behavior characteristic data list based on the three-dimensional model.
Further, the feature extraction module includes:
establishing a time axis according to the data length of the target monitoring data packet, inputting the behavior characteristic data list into the time axis, generating dynamic behavior data, and acquiring behavior characteristic data parameters based on the dynamic behavior data;
and acquiring a plurality of historical data of the person and historical behavior characteristics corresponding to each historical data, performing correlation analysis on the behavior characteristic data parameters and the historical behavior characteristics, acquiring an analysis result, and determining the behavior habit of the person and the corresponding target behavior characteristics according to the analysis result.
Further, the intelligent snapshot unit includes:
the behavior prejudging module is used for determining the behavior habit of the same person based on the behavior characteristic data list of the same person, carrying out behavior deduction on the behavior habit of the same person based on the three-dimensional model, and judging whether abnormal behaviors occur in a deduction result;
the snapshot module is used for acquiring a target monitoring data segment corresponding to the person if the person is judged to have abnormal behaviors, and matching corresponding instruction data to control the data acquisition terminal to snapshot the target monitoring data segment corresponding to the person;
the evaluation module is used for acquiring the classification result of the abnormal behavior, evaluating the abnormal behavior based on the classification result, calculating an average value by combining the historical evaluation scores corresponding to the characters, and determining the average value as the evaluation value of the characters.
Further, judging whether the deduction result shows abnormal behavior or not includes calculating an abnormal coefficient of each person in the garbage throwing process according to the behavior habit of the person and the identification result of each parameter, specifically:
acquiring a standard behavior threshold of each parameter of the garbage throwing behavior, and acquiring a current behavior threshold of each character for each parameter according to an identification result of each parameter of each character;
comparing the standard behavior threshold value with the current behavior threshold value to determine the behavior deviation degree of each person for each parameter;
according to the behavior habit of each character and the behavior deviation degree of the character for each parameter, adjusting the abnormal behavior coefficient of the character in the garbage throwing process;
and comparing the abnormal behavior coefficient with a preset threshold, and judging the behavior as abnormal behavior if the abnormal behavior coefficient is higher than the preset threshold.
Further, the data analysis unit includes a behavior prediction module, the behavior prediction module including:
the data extraction labeling sub-module is used for extracting behavior characteristic data of the corresponding identified person from a video data segment of the past stored historical video data, labeling the behavior characteristic data, and screening and grouping to obtain a training set and a plurality of verification sets;
the model training sub-module is used for carrying out data training by adopting a training set, and then carrying out parameter adjustment optimization by adopting one verification set to obtain a behavior prediction model;
the model evaluation sub-module is used for evaluating the behavior prediction model by adopting a logistic regression model, and the logistic regression model calculates the accuracy index of the behavior prediction model by adopting the following formula:
in the above formula, ω represents an accuracy index of the behavior prediction model; x is x i Ith behavior feature data representing a person extracted from a video data segment; p (x) i ) Representing behavior prediction model versus behavior feature data x i Is predicted to conform to expected behavior data y i Probability of (2); y is i Expected behavior data representing predictions of ith behavior feature data; 1-p (x) i ) Representing behavior prediction model versus behavior feature data x i Is not in accordance with the predicted undesirable behavior data y i Probability of (2);
the judging and optimizing sub-module is used for judging whether the behavior prediction model is required to be optimized according to the accuracy index obtained by the model evaluation sub-module; if the accuracy index is smaller than the set index threshold value, indicating that the behavior prediction model still needs to be optimized, randomly selecting an unused verification set to perform parameter adjustment optimization on the behavior prediction model again, and evaluating the behavior prediction model by the model evaluation sub-module again after parameter adjustment optimization until the accuracy index of the behavior prediction model is not smaller than the index threshold value; the behavior prediction model is then used to predict behavior of the person identified in the segment of real-time video data.
Compared with the prior art, the invention has the beneficial effects that:
the automatic snapshot information is automatically identified through the camera, the functions of automatic snapshot, identification, warehousing, inquiring and comparing are realized by combining the information of face information, time and the like, the automatic detection of people, objects and the like is realized through setting different detection areas and rules, the intelligent snapshot is identified, the monitoring efficiency and the accuracy are improved, the three-dimensional modeling is carried out on the collected video data, the deduction and analysis of the behavior of a target object are realized, the intelligent snapshot is carried out on abnormal behaviors, and the intelligent snapshot device has the advantages of high precision, high robustness and high-level identification accuracy.
Drawings
Fig. 1 is a block diagram of a monitoring system for garbage throwing behavior analysis and illegal behavior multi-angle snapshot.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
In order to solve the technical problem that in the prior art, the snapping angle is locally limited and the face cannot be accurately positioned, please refer to fig. 1, the present embodiment provides the following technical scheme:
monitoring system to rubbish input action analysis and violation action multi-angle snapshot includes:
the data acquisition unit is used for acquiring video data of the garbage dispensing point actively uploaded by the data acquisition terminal, and acquiring sub-data carried by the video data, including the coordinate data of the garbage dispensing point and the basic data of the data acquisition terminal;
packaging the video data and the sub-data of the garbage dispensing point to generate a monitoring data set, and transmitting the monitoring data set to a data processing unit for data processing based on the Internet of things;
specifically, the basic data of the data acquisition terminal comprises camera operation data and control terminal data thereof, the coordinate data of the corresponding garbage dispensing points are uploaded, the data acquisition terminals installed on the garbage dispensing points are convenient to maintain and manage, meanwhile, the dispensing behaviors of the garbage dispensing points are convenient to analyze, important observation is carried out on the corresponding garbage dispensing points with more illegal dispensing behaviors, and management staff can dispatch staff on site according to actual needs, so that the management efficiency and management quality are improved, and management staff are simplified;
the data acquisition terminal comprises an automatic rotation driving base and a camera, a rotating shaft is embedded in the automatic rotation driving base, an automatic reset module is arranged on one side of the rotating shaft, the camera is electrically connected with an algorithm box through a wire, the algorithm box comprises an intelligent terminal analysis module and a cloud decision module, and the camera is provided with a shooting assembly, a position adjusting group and an angle adjusting assembly;
specifically, the algorithm box comprises intelligent terminal analysis and cloud decision making, so that integrated linkage of deep intelligence, target positioning, identification snapshot, multiple behavior analysis and autonomous tracking is realized; the algorithm box receives the control instruction, and simultaneously controls the shooting angle and position in real time through the position adjusting assembly and the angle adjusting assembly, so that the effect of accurate tracking, positioning and shooting is achieved, and the terminal is used for controlling operation, so that the automatic rotation driving base is driven to achieve the lifting effect; meanwhile, in order to ensure stability, the rotation shaft drives the base turntable to rotate so as to realize angle conversion of the lens, so that the shooting angles are diversified, and clearer illegal photos can be shot;
the data processing unit is used for storing the monitoring data set, identifying the face characteristics in the video data based on the face recognition technology, inputting the identified face into a pre-stored face database for information acquisition and comparison, and determining the face recognition result, wherein the pre-stored face database comprises: face image data, character information corresponding to the face image, historical behavior information and evaluation data;
the data analysis unit is used for acquiring a character behavior video data segment of the target object based on the face recognition result, extracting behavior characteristics in the video data segment, modeling and analyzing the behavior characteristics through a machine learning algorithm, and determining a behavior analysis result of the target object;
and the intelligent snapshot unit is used for predicting the analysis result of the target object based on the historical behavior information and the evaluation data, judging the abnormal behavior in the analysis result, performing intelligent snapshot, inputting an intelligent snapshot image into the pre-stored face database for storage, and establishing an association relation with corresponding face image data based on the face recognition result.
Specifically, the automatic snapshot information is automatically identified through the camera, the functions of automatic snapshot, identification, warehousing, inquiring and comparison and the like are realized by combining the information such as face information and time, the automatic detection of people, objects and the like is realized through setting different detection areas and rules, intelligent snapshot is identified, so that the monitoring efficiency and accuracy are improved, the collected video data are subjected to three-dimensional modeling, the behavior deduction and analysis of a target object are realized, the abnormal behavior is intelligently snapshot, and the intelligent snapshot device has high accuracy, high robustness and high-level identification accuracy.
In order to solve the technical problems in the prior art that the recognition result is low in accuracy and large in error due to the fact that the related data and the behavior data corresponding to the person cannot be recognized based on the face recognition, referring to fig. 1, the present embodiment provides the following technical scheme:
a data processing unit comprising:
the data storage module is used for acquiring the monitoring data set uploaded by the data processing unit, extracting the garbage dispensing point coordinate data in the monitoring data set, identifying a data vector space corresponding to the garbage dispensing point coordinate data, and storing video data in the monitoring data set into the data vector space;
the video preprocessing module is used for extracting face sample data characteristics from video data of the monitoring data set, determining a target monitoring data segment with the face data characteristics based on the face sample data characteristics, and integrating the target monitoring data segment to generate a target monitoring data packet;
the face recognition module is used for extracting the characteristics of face data in the target monitoring data packet based on the convolutional neural network algorithm training model, inputting the face characteristics into a pre-stored face database for one-to-one matching, and obtaining a recognition result;
the face recognition module is also used for determining the face direction of the human body and the corresponding human body position data based on the obtained face characteristics, and matching the corresponding control instructions to control the data acquisition terminal to adjust the shooting angle to the target position;
carrying out standardization processing on the image corresponding to the facial features of the face to obtain a standardized image;
acquiring a characteristic value of each pixel point in a standardized image, and constructing a first matrix based on the characteristic value of each pixel point;
each pre-stored face image in the pre-stored face database corresponds to a second matrix, wherein the second matrix corresponds to the pre-stored face image one by one;
respectively calculating the similarity of the first matrix and each second matrix in a pre-stored face database, extracting a corresponding pre-stored face image after the similarity is higher than a preset threshold value, and taking the corresponding pre-stored face image as a recognition result;
and the information matching module is used for acquiring the character information, the historical behavior information and the evaluation data corresponding to the face image based on the identification result.
Specifically, through carrying out standardized processing on the image corresponding to the facial features of the face, the consistency of the size of the pre-stored facial image in a pre-stored facial database is realized, the correspondence between the first matrix and the second matrix is conveniently ensured, the pre-stored facial image with high similarity between the first matrix and each second matrix is used as a recognition result, and further, the accuracy of judging the similarity and the size of a preset threshold value is improved, and the accuracy of the obtained recognition result is improved.
In order to solve the technical problem that in the prior art, abnormal behaviors cannot be evaluated based on behavior habits of each person, and classification is performed according to the abnormal behaviors of the person, referring to fig. 1, the present embodiment provides the following technical scheme:
a data analysis unit comprising:
the feature extraction module is used for extracting behavior feature data of the same person from the target monitoring data packet based on the identification result and acquiring a behavior feature data list of the same person based on the time sequence;
establishing a time axis according to the data length of the target monitoring data packet, inputting the behavior characteristic data list into the time axis, generating dynamic behavior data, and acquiring behavior characteristic data parameters based on the dynamic behavior data;
acquiring a plurality of historical data of the person and historical behavior characteristics corresponding to each historical data, performing correlation analysis on the behavior characteristic data parameters and the historical behavior characteristics, acquiring analysis results, and determining behavior habits of the person and corresponding target behavior characteristics according to the analysis results;
and the three-dimensional modeling module is used for constructing a three-dimensional model corresponding to each person based on the video data, and carrying out three-dimensional simulation on the behavior characteristic data list based on the three-dimensional model.
Intelligent snapshot unit includes:
the behavior prejudging module is used for determining the behavior habit of the same person based on the behavior characteristic data list of the same person, carrying out behavior deduction on the behavior habit of the same person based on the three-dimensional model, and judging whether abnormal behaviors occur in a deduction result;
judging whether the deduction result shows abnormal behavior or not, wherein the abnormal coefficient of each character in the garbage throwing process is calculated according to the behavior habit of the character and the identification result of each parameter, and specifically comprises the following steps:
acquiring a standard behavior threshold of each parameter of the garbage throwing behavior, and acquiring a current behavior threshold of each character for each parameter according to an identification result of each parameter of each character;
comparing the standard behavior threshold value with the current behavior threshold value to determine the behavior deviation degree of each person for each parameter;
according to the behavior habit of each character and the behavior deviation degree of the character for each parameter, adjusting the abnormal behavior coefficient of the character in the garbage throwing process;
comparing the abnormal behavior coefficient with a preset threshold, and if the abnormal behavior coefficient is higher than the preset threshold, judging the behavior as abnormal behavior;
the snapshot module is used for acquiring a target monitoring data segment corresponding to the person if the person is judged to have abnormal behaviors, and matching corresponding instruction data to control the data acquisition terminal to snapshot the target monitoring data segment corresponding to the person;
specifically, the function expansion of the data acquisition terminal can be carried out, snapshot data are reported to a related big data platform, and measures such as warning, rejection, law enforcement and punishment are adopted by an upper big data platform for garbage generation units with unqualified garbage classification or property units thereof, so that a long-acting management mechanism is formed, a garbage classification responsibility main body is guided and supervised to improve thought understanding, the accurate classification of household garbage sources, the standard collection and transportation of the terminal and the scientific disposal of the terminal are facilitated, the quality improvement and the efficiency improvement of garbage classification are ensured, and the living environment of residents is further improved;
the evaluation module is used for acquiring the classification result of the abnormal behavior, evaluating the abnormal behavior based on the classification result, calculating an average value by combining the historical evaluation scores corresponding to the characters, and determining the average value as the evaluation value of the characters;
specifically, the abnormal behavior classification support vector machine classification algorithm model, the naive Bayesian algorithm model and the application of the neural network algorithm model are adopted; identifying and judging the behaviors of the targets and classifying the targets through a machine learning classification algorithm; the classification algorithm includes counting the number of documents that appear for each category in the training dataset, and the number of times each sample appears in each category, calculating a conditional probability that each sample appears in each category, and processing the types that do not appear in the training dataset.
Specifically, whether the behaviors of each person in the garbage throwing process have abnormal coefficients can be determined by calculating the behavior habits of each person in the garbage throwing process, so that deduction can be performed on the basis of the behavior habits for each person, the deduction and analysis results are more objective and reliable, abnormal behavior snapshot is performed according to the judgment results of the abnormal behaviors, meanwhile, the abnormal behaviors are classified on the basis of a classification algorithm model, including illegal behaviors such as mixed throwing, behaviors and packet loss, and the like, further, the person is evaluated, a data basis is further provided for the next throwing behaviors of the person, centralized management and key education are facilitated for key people, and the normalization of garbage classification is effectively improved.
On the basis of the foregoing embodiment, the data analysis unit includes a behavior prediction module, which includes:
the data extraction labeling sub-module is used for extracting behavior characteristic data of the corresponding identified person from a video data segment of the past stored historical video data, labeling the behavior characteristic data, and screening and grouping to obtain a training set and a plurality of verification sets;
the model training sub-module is used for carrying out data training by adopting a training set, and then carrying out parameter adjustment optimization by adopting one verification set to obtain a behavior prediction model;
the model evaluation sub-module is used for evaluating the behavior prediction model by adopting a logistic regression model, and the logistic regression model calculates the accuracy index of the behavior prediction model by adopting the following formula:
in the above formula, ω represents an accuracy index of the behavior prediction model; x is x i Ith behavior feature data representing a person extracted from a video data segment; p (x) i ) Representing behavior prediction model versus behavior feature data x i Is predicted to conform to expected behavior data y i Probability of (2); y is i Expected behavior data representing predictions of ith behavior feature data; 1-p (x) i ) Representing behavior prediction model versus behavior feature data x i Is not in accordance with the predicted undesirable behavior data y i Probability of (2);
the judging and optimizing sub-module is used for judging whether the behavior prediction model is required to be optimized according to the accuracy index obtained by the model evaluation sub-module; if the accuracy index is smaller than the set index threshold value, indicating that the behavior prediction model still needs to be optimized, randomly selecting an unused verification set to perform parameter adjustment optimization on the behavior prediction model again, and evaluating the behavior prediction model by the model evaluation sub-module again after parameter adjustment optimization until the accuracy index of the behavior prediction model is not smaller than the index threshold value; the behavior prediction model is then used to predict behavior of the person identified in the segment of real-time video data.
The embodiment adopts the video data segment of the historical video data to extract the behavior characteristic data of the corresponding identified characters, constructs a training set and a plurality of verification sets, uses the training set and part of the verification sets for data training and parameter adjustment optimization of the behavior prediction model, adopts a model evaluation submodule to evaluate the behavior prediction model, calculates the accuracy index of the behavior prediction model and compares the accuracy index with an index threshold value, judges the accuracy degree of the behavior prediction model, and additionally selects the verification sets to perform parameter adjustment optimization again if the accuracy degree can not meet the index threshold value requirement; after the accuracy degree reaches the index threshold requirement, the behavior prediction model is used for predicting the behavior of the person identified in the real-time video data segment; by adopting the scheme, the behavior prediction accuracy of the behavior prediction model can be effectively ensured, and the prediction error probability is reduced.
The foregoing is only a preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art, who is within the scope of the present invention, should be covered by the protection scope of the present invention by making equivalents and modifications to the technical solution and the inventive concept thereof.
Claims (9)
1. Monitoring system to rubbish input action analysis and violation action multi-angle snapshot, its characterized in that: comprising the following steps:
the data acquisition unit is used for acquiring video data of the garbage dispensing point actively uploaded by the data acquisition terminal, and acquiring sub-data carried by the video data, including the coordinate data of the garbage dispensing point and the basic data of the data acquisition terminal;
packaging the video data and the sub-data of the garbage dispensing point to generate a monitoring data set, and transmitting the monitoring data set to a data processing unit for data processing based on the Internet of things;
the data processing unit is used for storing the monitoring data set, identifying the face characteristics in the video data based on the face recognition technology, inputting the identified face into a pre-stored face database for information acquisition and comparison, and determining the face recognition result, wherein the pre-stored face database comprises: face image data, character information corresponding to the face image, historical behavior information and evaluation data;
the data analysis unit is used for acquiring a character behavior video data segment of the target object based on the face recognition result, extracting behavior characteristics in the video data segment, modeling and analyzing the behavior characteristics through a machine learning algorithm, and determining a behavior analysis result of the target object;
and the intelligent snapshot unit is used for predicting the analysis result of the target object based on the historical behavior information and the evaluation data, judging the abnormal behavior in the analysis result, performing intelligent snapshot, inputting an intelligent snapshot image into the pre-stored face database for storage, and establishing an association relation with corresponding face image data based on the face recognition result.
2. The monitoring system for garbage placement behavior analysis and multi-angle snap shots of violations according to claim 1, wherein: the data acquisition terminal comprises an automatic rotation driving base and a camera, wherein a rotation shaft is embedded in the automatic rotation driving base, an automatic reset module is installed on one side of the rotation shaft, the camera is electrically connected with an algorithm box through a wire, the algorithm box comprises an intelligent terminal analysis module and a cloud decision module, and the camera is provided with a shooting assembly, a position adjusting group and an angle adjusting assembly.
3. The monitoring system for garbage placement behavior analysis and multi-angle snapshot of offensive behavior of claim 2, wherein: the data processing unit includes:
the data storage module is used for acquiring the monitoring data set uploaded by the data processing unit, extracting the garbage dispensing point coordinate data in the monitoring data set, identifying a data vector space corresponding to the garbage dispensing point coordinate data, and storing video data in the monitoring data set into the data vector space;
the video preprocessing module is used for extracting face sample data characteristics from video data of the monitoring data set, determining a target monitoring data segment with the face data characteristics based on the face sample data characteristics, and integrating the target monitoring data segment to generate a target monitoring data packet;
the face recognition module is used for extracting the characteristics of face data in the target monitoring data packet based on the convolutional neural network algorithm training model, inputting the face characteristics into a pre-stored face database for one-to-one matching, and obtaining a recognition result;
the face recognition module is also used for determining the face direction of the human body and the corresponding human body position data based on the obtained face characteristics, and matching the corresponding control instructions to control the data acquisition terminal to adjust the shooting angle to the target position;
and the information matching module is used for acquiring the character information, the historical behavior information and the evaluation data corresponding to the face image based on the identification result.
4. A monitoring system for garbage placement analysis and multi-angle snap shots of offensive behavior as in claim 3, wherein: the face recognition module inputs face features into a pre-stored face database for one-to-one matching, and comprises the following steps of;
carrying out standardization processing on the image corresponding to the facial features of the face to obtain a standardized image;
acquiring a characteristic value of each pixel point in a standardized image, and constructing a first matrix based on the characteristic value of each pixel point;
each pre-stored face image in the pre-stored face database corresponds to a second matrix, wherein the second matrix corresponds to the pre-stored face image one by one;
and respectively calculating the similarity of the first matrix and each second matrix in a pre-stored face database, extracting a corresponding pre-stored face image after the similarity is higher than a preset threshold value, and taking the corresponding pre-stored face image as a recognition result.
5. The monitoring system for garbage placement behavior analysis and multi-angle snap shots of violations of claim 4, wherein: the data analysis unit includes:
the feature extraction module is used for extracting behavior feature data of the same person from the target monitoring data packet based on the identification result and acquiring a behavior feature data list of the same person based on the time sequence;
and the three-dimensional modeling module is used for constructing a three-dimensional model corresponding to each person based on the video data, and carrying out three-dimensional simulation on the behavior characteristic data list based on the three-dimensional model.
6. The monitoring system for garbage placement behavior analysis and multi-angle snap shots of violations of claim 5, wherein: the feature extraction module comprises:
establishing a time axis according to the data length of the target monitoring data packet, inputting the behavior characteristic data list into the time axis, generating dynamic behavior data, and acquiring behavior characteristic data parameters based on the dynamic behavior data;
and acquiring a plurality of historical data of the person and historical behavior characteristics corresponding to each historical data, performing correlation analysis on the behavior characteristic data parameters and the historical behavior characteristics, acquiring an analysis result, and determining the behavior habit of the person and the corresponding target behavior characteristics according to the analysis result.
7. The monitoring system for garbage placement behavior analysis and multi-angle snap shots of violations of claim 6, wherein: the intelligent snapshot unit comprises:
the behavior prejudging module is used for determining the behavior habit of the same person based on the behavior characteristic data list of the same person, carrying out behavior deduction on the behavior habit of the same person based on the three-dimensional model, and judging whether abnormal behaviors occur in a deduction result;
the snapshot module is used for acquiring a target monitoring data segment corresponding to the person if the person is judged to have abnormal behaviors, and matching corresponding instruction data to control the data acquisition terminal to snapshot the target monitoring data segment corresponding to the person;
the evaluation module is used for acquiring the classification result of the abnormal behavior, evaluating the abnormal behavior based on the classification result, calculating an average value by combining the historical evaluation scores corresponding to the characters, and determining the average value as the evaluation value of the characters.
8. The monitoring system for garbage placement behavior analysis and multi-angle snap shots of violations of claim 7, wherein: judging whether the deduction result shows abnormal behavior or not, wherein the abnormal coefficient of each character in the garbage throwing process is calculated according to the behavior habit of the character and the identification result of each parameter, and specifically comprises the following steps:
acquiring a standard behavior threshold of each parameter of the garbage throwing behavior, and acquiring a current behavior threshold of each character for each parameter according to an identification result of each parameter of each character;
comparing the standard behavior threshold value with the current behavior threshold value to determine the behavior deviation degree of each person for each parameter;
according to the behavior habit of each character and the behavior deviation degree of the character for each parameter, adjusting the abnormal behavior coefficient of the character in the garbage throwing process;
and comparing the abnormal behavior coefficient with a preset threshold, and judging the behavior as abnormal behavior if the abnormal behavior coefficient is higher than the preset threshold.
9. The monitoring system for garbage placement behavior analysis and multi-angle snap shots of violations according to claim 1, wherein: the data analysis unit includes a behavior prediction module, the behavior prediction module including:
the data extraction labeling sub-module is used for extracting behavior characteristic data of the corresponding identified person from a video data segment of the past stored historical video data, labeling the behavior characteristic data, and screening and grouping to obtain a training set and a plurality of verification sets;
the model training sub-module is used for carrying out data training by adopting a training set, and then carrying out parameter adjustment optimization by adopting one verification set to obtain a behavior prediction model;
the model evaluation sub-module is used for evaluating the behavior prediction model by adopting a logistic regression model, and the logistic regression model calculates the accuracy index of the behavior prediction model by adopting the following formula:
in the above formula, ω represents an accuracy index of the behavior prediction model; x is x i Ith behavior feature data representing a person extracted from a video data segment; p (x) i ) Representing behavior prediction model versus behavior feature data x i Is predicted to conform to expected behavior data y i Probability of (2); y is i Expected behavior data representing predictions of ith behavior feature data; 1-p (x) i ) Representing behavior prediction model versus behavior feature data x i Is not in accordance with the predicted undesirable behavior data y i Probability of (2);
the judging and optimizing sub-module is used for judging whether the behavior prediction model is required to be optimized according to the accuracy index obtained by the model evaluation sub-module; if the accuracy index is smaller than the set index threshold value, indicating that the behavior prediction model still needs to be optimized, randomly selecting an unused verification set to perform parameter adjustment optimization on the behavior prediction model again, and evaluating the behavior prediction model by the model evaluation sub-module again after parameter adjustment optimization until the accuracy index of the behavior prediction model is not smaller than the index threshold value; the behavior prediction model is then used to predict behavior of the person identified in the segment of real-time video data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310970135.XA CN116959117A (en) | 2023-08-03 | 2023-08-03 | Monitoring system for garbage throwing behavior analysis and illegal behavior multi-angle snapshot |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310970135.XA CN116959117A (en) | 2023-08-03 | 2023-08-03 | Monitoring system for garbage throwing behavior analysis and illegal behavior multi-angle snapshot |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116959117A true CN116959117A (en) | 2023-10-27 |
Family
ID=88452854
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310970135.XA Pending CN116959117A (en) | 2023-08-03 | 2023-08-03 | Monitoring system for garbage throwing behavior analysis and illegal behavior multi-angle snapshot |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116959117A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117541988A (en) * | 2023-11-21 | 2024-02-09 | 苏州纳故环保科技有限公司 | Object detection method of renewable resource recycling system |
-
2023
- 2023-08-03 CN CN202310970135.XA patent/CN116959117A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117541988A (en) * | 2023-11-21 | 2024-02-09 | 苏州纳故环保科技有限公司 | Object detection method of renewable resource recycling system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110738127B (en) | Helmet identification method based on unsupervised deep learning neural network algorithm | |
| CN110059581A (en) | People counting method based on depth information of scene | |
| CN108319926A (en) | A kind of the safety cap wearing detecting system and detection method of building-site | |
| CN113435546B (en) | Migratable image recognition method and system based on differentiation confidence level | |
| CN115527203B (en) | Cereal drying remote control method and system based on Internet of things | |
| CN105512640A (en) | Method for acquiring people flow on the basis of video sequence | |
| CN109829382B (en) | Abnormal target early warning tracking system and method based on intelligent behavior characteristic analysis | |
| CN102163290A (en) | Method for modeling abnormal events in multi-visual angle video monitoring based on temporal-spatial correlation information | |
| Cao et al. | CNN-based intelligent safety surveillance in green IoT applications | |
| CN110728252B (en) | Face detection method applied to regional personnel motion trail monitoring | |
| CN116959117A (en) | Monitoring system for garbage throwing behavior analysis and illegal behavior multi-angle snapshot | |
| CN112298844B (en) | Garbage classification supervision method and device | |
| CN117611015B (en) | Real-time monitoring system for quality of building engineering | |
| CN113191273A (en) | Oil field well site video target detection and identification method and system based on neural network | |
| CN116935108A (en) | Method, device, equipment and medium for monitoring abnormal garbage throwing behavior | |
| Lu et al. | An efficient network for multi-scale and overlapped wildlife detection | |
| CN114229279A (en) | Intelligent garbage room capable of realizing garbage classification behavior supervision | |
| CN116959099B (en) | Abnormal behavior identification method based on space-time diagram convolutional neural network | |
| CN117541054A (en) | Community security monitoring method and system based on intelligent property | |
| CN115188031A (en) | Fingerprint identification method, computer program product, storage medium and electronic device | |
| CN113128452A (en) | Greening satisfaction acquisition method and system based on image recognition | |
| CN117392616B (en) | Method and device for identifying supervision behaviors of garbage throwing, electronic equipment and medium | |
| US20230386327A1 (en) | Methods and internet of things systems for managing traffic road cleaning in smart city | |
| CN118097198B (en) | Automatic dressing compliance management and control system and method based on artificial intelligence | |
| CN118172711A (en) | AI big data intelligent management method and system |
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 |