CN105354830A - Method, apparatus and system for controller fatigue detection based on multiple regression model - Google Patents
Method, apparatus and system for controller fatigue detection based on multiple regression model Download PDFInfo
- Publication number
- CN105354830A CN105354830A CN201510645187.5A CN201510645187A CN105354830A CN 105354830 A CN105354830 A CN 105354830A CN 201510645187 A CN201510645187 A CN 201510645187A CN 105354830 A CN105354830 A CN 105354830A
- Authority
- CN
- China
- Prior art keywords
- data
- controller
- fatigue
- index
- land
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- 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
- G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
-
- 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
- G06V40/174—Facial expression recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30232—Surveillance
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Human Computer Interaction (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
- Emergency Alarm Devices (AREA)
Abstract
The present invention provides a method, apparatus and system for air traffic controller fatigue detection. The method comprises: receiving video monitoring data of a controller position and air-ground communication data obtained by a real-time audio and video all-data collecting apparatus; detecting and analyzing the video monitoring data of the controller position and the air-ground communication data to obtain controller fatigue indicator detecting data; and with reference to a preset sample of the controller fatigue indictor detecting data, establishing a multiple linear regression model and a multiple non-linear regression model, and introducing the controller fatigue indicator detecting data after standardized conversion into the models to obtain an aggregated index of the current controller fatigue indicators. If the aggregated index of the controller fatigue indicators exceeds an alarm standard, then an alarm is triggered. In addition, the present invention provides an apparatus corresponding to the method, as well as a physical system. The method, apparatus and system can be applied to the field of air traffic control, and have relatively high practicability and operability.
Description
Technical field
The invention provides air traffic controller's fatigue detection method, the Apparatus and system of a kind of integrated face and voice, belong to image processing and pattern recognition field.
Technical background
At present, the correlative study for controller's fatigue is mainly carried out from two aspects, namely based on controller's fatigue detecting of face recognition, and based on controller's fatigue detecting of workload assessment.The former is based on the parsing to controller's facial characteristics, obtain the data of PERCLOS value, average three fatigue detecting indexs such as eye-closing period, yawn frequency, threshold value corresponding to three indexs judged by coupling controller fatigue, controller's fatigue is detected, if application number is that CN103400471A gives a kind of Driver Fatigue Detection, it adopts least square method to carry out analyzing and processing to eyes, face characteristic, judges whether driver is in fatigue state.And the latter is based on ATC controller workload evaluation studies, developed successively since 20 century 70s and three class controller fatigue detection methods, that is: (1) is according to controller's physiology, behavioural characteristic analysis, draws control workload intensity; (2) the subjective assessment method of observation and questionnaire form is taked; (3) controller's work is segmented, observable work is surveyed to the time of counting and consuming, temporal consumption is converted into for invisible work, realizes the qualitative assessment to ATC controller workload in time measure mode.
The existing research for controller's fatigue detecting, mainly has the following disadvantages: (1) research method aspect, and qualitative examination is more, and quantitative examination is less, causes objectivity not enough; (2) Testing index aspect, index dimension is comparatively single, comprehensive, comprehensive deficiency; (3) application aspect, existing research still rests on the laboratory study stage, serves primarily in strategic decision, and few towards the practical engineering application of air traffic control unit.
Summary of the invention
The invention provides a kind of air traffic controller's fatigue detection method, Apparatus and system, wherein method comprises the following steps:
Step 1, receives the empty communicating data of control seat video monitoring data and land that real-time audio and video all data harvester obtains; Step 2, carries out detection to the empty communicating data of described control seat video monitoring data and land and analyzes, obtain controller's fatigue index and detect data; Step 3, data sample is detected with reference to preset controller's fatigue index, set up Multivariate regression model and multiple nonlinear regression model (NLRM) respectively, controller's fatigue index in step 2 is detected after data carry out standardization conversion and bring model into, obtain current controller's fatigue index and assemble index, if controller's fatigue index is assembled index and is exceeded alarm standard, then trigger alerts.
Preferably, the process receiving described control seat video monitoring data is: Real-time Collection video information, after carrying out noise reduction, filtering process, extracts eigenwert.
Preferably, after extracting eigenwert, mate with the template data preserved in video template storehouse, thus the identity of current controller is identified.
Preferably, the process receiving the empty communicating data in described land is: adopt cable to be drawn from distributing frame by voice signal and be connected to speech processor, when voice signal being detected, start recording, there is no sound when detecting and continue 3 seconds, then stopping recording, namely complete and once record, often complete and once record, acoustic information is preserved with audio file formats.
Preferably, also comprise after step 1 and the empty communicating data of described control seat video monitoring data and land is carried out synchronous step, be specially: with the timestamp of video monitoring data for time reference, the process sound intermediate frequency play plays the current play time data that thread real-time reception video playback thread transmits, and synchronous adjustment audio frequency played data, the empty communicating data of control seat video monitoring data and land is synchronously play, realizes playing in real time and history playback.
Preferably, carry out detection to described control seat video monitoring data in step 2 to analyze, the controller's fatigue index obtained detects data and comprises: PERCLOS (percentageofeyelidclosure, percent eye-closure) value, average eye-closing period and frequency of yawning.
Preferably, carry out detection to the empty communicating data in described land in step 2 and analyze, the controller's fatigue index obtained detects data and comprises: channel seizure ratio, talk times, average audio, average loudness of a sound and land, sector sky call keyword occurrence rate.
A kind of air traffic controller's fatigue detection device, comprises with lower module: receiver module, for receiving the empty communicating data of control seat video monitoring data and land that real-time audio and video all data harvester obtains; Computing module, analyzing for carrying out detection to the empty communicating data of described control seat video monitoring data and land, obtaining controller's fatigue index and detecting data; Alarm module, for detecting data sample with reference to preset controller's fatigue index, set up Multivariate regression model and multiple nonlinear regression model (NLRM) respectively, the controller's fatigue index exported by computing module detects after data carry out standardization conversion and brings model into, obtain current controller's fatigue index and assemble index, if controller's fatigue index is assembled index and is exceeded alarm standard, then trigger alerts.
A kind of air traffic controller's fatigue detecting system, comprising: control seat real-time audio and video all data acquisition platform: for carrying out the acquisition and processing of control seat video monitoring data, the empty communicating data in land; Controller's fatigue detecting database that the empty call voice of integrated face recognition and land is resolved: for the data of control seat real-time audio and video all data acquisition platform acquisition process are carried out sorting out and preserving, store every controller's fatigue index detection data that controller's fatigue index detection module calculates simultaneously, and the calculating of controller's fatigue strength alarm module and alarm result, thus history of forming statistics; Controller's fatigue index detection module: carry out detection analyze for the video monitoring data to control seat real-time audio and video all data acquisition platform acquisition and processing, the empty communicating data in land, obtain controller's fatigue index and detect data; Controller's fatigue strength alarm module: based on the historical statistics result in controller's fatigue detecting database that the empty call voice of integrated face recognition and land is resolved, set up Multivariate regression model and multiple nonlinear regression model (NLRM) respectively, controller's fatigue index is detected after data carry out standardization conversion and bring model into, obtain current controller's fatigue index and assemble index, if controller's fatigue index is assembled index and is exceeded alarm standard, then trigger alerts.
Compared with prior art, fatigue detection method, device, the system of integrated face provided by the invention and voice, by each index such as facial characteristics, working load of mark controller fatigue, comprehensively, synthetically consider that there is stronger practicality and operability.
Accompanying drawing explanation
Fig. 1 is system construction drawing of the present invention;
Fig. 2 is audio data collecting process flow diagram of the present invention;
Fig. 3 is the synchronous process flow diagram of audio, video data of the present invention;
Fig. 4 is system network architecture figure of the present invention;
Fig. 5 is that the controller's fatigue strength that the present invention is based on multiple regression detects and response alarm processing figure.
Embodiment
The present invention is described in detail below in conjunction with accompanying drawing.
Embodiment 1:
Fig. 1 is system construction drawing of the present invention, the fatigue detecting system of integrated face provided by the invention and voice comprises: control seat real-time audio and video all data acquisition platform, controller's fatigue detecting database that the empty call voice of integrated face recognition and land is resolved, controller's fatigue index detection module, controller's fatigue strength alarm module, based on the historical statistics result in controller's fatigue detecting database that the empty call voice of integrated face recognition and land is resolved, set up Multivariate regression model and multiple nonlinear regression model (NLRM) respectively, controller's fatigue index is detected after data carry out standardization conversion and bring model into, obtain current controller's fatigue index and assemble index, if controller's fatigue index is assembled index and is exceeded alarm standard, then trigger alerts.
Embodiment 2:
Air traffic controller's fatigue detection method of integrated face and voice, is characterized in that, comprise the following steps:
Step 1, receive the empty communicating data of control seat video monitoring data and land that real-time audio and video all data harvester obtains, the process receiving described control seat video monitoring data is, Real-time Collection video information, after carrying out noise reduction, filtering process, extract eigenwert; The process receiving the empty communicating data in described land is adopt cable to be drawn from distributing frame by voice signal and be connected to speech processor, when detecting that voice signal appears in system audio passage, starts recording; Sound go do not had 3 seconds when detecting in voice-grade channel, stop recording, namely complete and once record, voice signal detected next time, again record, in voice-grade channel, there is no voice signal, then do not record, often complete and once record, preserved by acoustic information with audio file formats, file is named with seat title and markers.To record stored in the audio file information table of database " seat name, audio file name, start time, end time ", gatherer process as shown in Figure 2;
Step 2, comprise and the empty communicating data of described control seat video monitoring data and land is carried out synchronous step, be specially: with the timestamp of video data for time reference, the process sound intermediate frequency play plays the current play time data that thread real-time reception video playback thread transmits, and synchronous adjustment audio frequency played data, synchronously play by the empty communicating data of control seat video monitoring data and land, realize playing in real time and history playback, idiographic flow as shown in Figure 3;
Step 3, carries out detection to the empty communicating data of described control seat video monitoring data and described land and analyzes, obtain controller's fatigue index and detect data;
Step 4, data sample is detected with reference to preset controller's fatigue index, set up Multivariate regression model and multiple nonlinear regression model (NLRM) respectively, controller's fatigue index in step 3 is detected after data carry out standardization conversion and bring model into, obtain current controller's fatigue index and assemble index, if controller's fatigue index is assembled index and is exceeded alarm standard, then trigger alerts.
Embodiment 3:
Air traffic controller's fatigue detection method of a kind of integrated face and voice, by 8 fatigue indexes, controller's degree of fatigue is detected, specifically comprise: the counting statistics by control seat video monitoring data, controller being carried out to facial expression feature, thus analysis draws PERCLOS value, the indicator-specific statistics data of average eye-closing period and frequency of yawning, carry out detection analysis to the empty communicating data in the land of control seat real-time audio and video all data acquisition platform acquisition and processing to obtain: channel seizure ratio, talk times, average audio, average loudness of a sound, land, sector sky call keyword occurrence rate.
Wherein, PERCLOS value detects the ratio adopting eyes closed time and 60 seconds.P80 metering system (eyelid lid crosses the area of pupil more than the time scale shared by 80%) is adopted to calculate PERCLOS value.
Average eye-closing period detection method is: set the number of times of controller's eye closing in the unit interval as n, each duration of closing one's eyes is te
i, in the unit interval, total eye-closing period of controller is te
total, the average eye-closing period of controller is te
average, then
te
average=te
total/ n.
Frequency detecting method of yawning is: according to the analysis to controller's facial characteristics, can count the yawning number of times of controller in the unit interval, and then draw yawning frequency.
The empty talk channel occupancy in land detects: set the total length of unit interval as T
total, total air time length of unit interval sector is T
sec, the empty talk channel occupancy in land is T
rate, then T
rate=T
sec/ T
total.
Land empty talk times detection: the empty call voice data in system acquisition VHF land, analyzes it, and an empty talk times in land is counted in each call, and carrying out adding up to number of times draws achievement data.
Land sky call average audio detects: system carries out wave form analysis to the empty call voice data in the VHF land gathered, and obtains sound frequency data during call, is then averaged to drawn result, obtains land sky call average audio index.
The land average loudness of a sound of sky call detects: system carries out wave form analysis to the empty call voice data in the VHF land gathered, and obtains sound loudness of a sound data during call, is then averaged to drawn result, obtains the average loudness of a sound index of land sky call.
Land, sector sky call keyword occurrence rate detects: system carries out wave form analysis and speech recognition to the empty call voice data in the VHF land gathered, when being resolved to " please repeat (sayagain) " of setting in system, key words such as " corrigendums (correction) " time, the time occur key word and number of times are added up, and obtain land, sector sky call keyword occurrence rate index.
Embodiment 4:
Air traffic controller's fatigue detection method of integrated face and voice, it realizes alarm step based on Multivariate regression model and multiple nonlinear regression model (NLRM).Mainly comprise four parts, be i.e. the ratio choosing of the structure of regression model, regression model, controller's fatigue strength detect, the alarm of controller's fatigue strength.Specific algorithm step is:
Step 1: choose variable: this research assembles index for dependent variable with controller's fatigue index, is designated as Y.Fatigue detecting index amounts to 8, and note independent variable X is: X={X
i, i=1,2 ..., 8}, wherein, the index based on face recognition comprises { X
1, X
2, X
3, represent PERCLOS value respectively, average eye-closing period, frequency of yawning; Based on land sky call load index be { X
4, X
5, X
5, X
7, X
8, represent the empty talk channel occupancy in land, the empty talk times in land, land sky call average audio, the average loudness of a sound of land sky call, land, sector sky call keyword occurrence rate respectively.According to fatigue detecting index of correlation, with 1 minute for duration, generate fatigue detecting sample data, obtain the input value X of above-mentioned 8 indexs.On this basis, by senior control expert, sample controller fatigue strength is classified, as dependent variable input value Y.
Step 2: data processing: consider to there is dimension difference and magnitude differences between different index, for convenience of the regretional analysis of model, need to carry out standardization conversion to achievement data.Make x
ij, x '
ijrepresent the raw data of sample set and the data after standardization conversion respectively,
s
jrepresent respectively jth (j=1,2 ..., 8) average of individual achievement data and variance, then:
data x ' after standardization is changed
ij, as the input data of regretional analysis.
Step 3: build regression model: based on the historical statistics result of above-mentioned 8 controller's fatigue detecting index sample datas, set up Multivariate regression model and multiple nonlinear regression model (NLRM) respectively, and solve coefficient b
i, the Multivariate regression model of foundation is:
Y＝XB+U
Wherein,
The multiple nonlinear regression model (NLRM) set up is: y=f [(b
1, b
2..., b
k); X
1, X
2..., X
n],
Wherein dependent variable Y is controller's fatigue index assembly index, and independent variable X is the indicator-specific statistics data of eight fatigue detecting indexs, X
i(i=1,2 ..., n), n=8, m represent m group statistical data, and μ is except n independent variable is on the stochastic error except the impact of dependent variable Y, Normal Distribution.Wherein nonlinear solshing according to sample data feature, can adopt the forms such as power function, negative exponent, hyperbolic function.The generation of X, Y data can be taked but be not limited to following method: according to fatigue detecting index of correlation, with 1 minute for duration, generates fatigue detecting sample data, obtains the input value X of above-mentioned 8 indexs.On this basis, by senior control expert, sample controller fatigue strength is classified, as dependent variable input value Y.
Step 4: regression model is than choosing: the parameter b of multiple regression model
iafter estimating, after namely obtaining sample regression function, also need to carry out statistical test to this sample regression function further, comprise degree of fitting inspection (coefficient of determination R
2), significance test (p value), and the Estimating Confidence Interval etc. of parameter.According to the coefficient of determination R that model returns
2value, F inspection, t inspection, verify respectively and compare degree of fitting, the conspicuousness of 2 kinds of regression models, on the obvious basis of, conspicuousness higher in model-fitting degree, calculating the metrical error of two kinds of regression models, and choose the minimum a kind of model of error, as last detection model.
Step 5: regression model result exports: according to the average of 8 indexs that m group sample data obtains
variance s
j, to the real time input data t of current collection
j(j=1,2 ..., 8) carry out standardization conversion,
for the input data after process.After carrying out standardization, by t
j' import in regression model, obtain current controller's fatigue index and assemble index, the classification results that corresponding controller's fatigue strength detects.
Step 6: controller's fatigue strength alarm: assemble index results according to controller's fatigue index, with reference to the alarm criteria triggers response alarm of setting.
Embodiment 5:
System network architecture as shown in Figure 4.The network platform of whole system will rely on existing management information net, and acquisition platform and blank pipe are produced network and carried out physical isolation, ensure the unidirectional delivery of data, stop network attack, to ensure related data security and production run system reliability.System provides patterned operation-interface, so that system manager configures parameters.As: the interval time of store video, the interval time etc. of tired diagnosis.There is provided system user to manage and arrange interface, the essential information of system manager to dissimilar user manages, as increased, deleting, revise the operations such as user profile.System manager also can be the authority that all types of users gives different stage.
Native system is by the empty call voice analytic technique of integrated face recognition and land, by each index such as facial characteristics, working load of mark controller fatigue, comprehensively, synthetically consider, and adopt the data statistical approach of Multivariate regression model and multiple nonlinear regression model (NLRM), give quantitative index.System is easy to set up, and identifies that degree is high, and algorithm is easy to realize.Utilize this fatigue detecting system to carry out the fatigue detecting of working site to air traffic controller, replace the controller on duty being in fatigue state in time, guarantee flight safety is had important practical significance.
Below by reference to the accompanying drawings, System's composition of the present invention and principle of work is described in detail.But those of ordinary skill in the art it should be understood that instructions is only for explaining claims.But protection scope of the present invention is not limited to instructions.Any those skilled in the art of being familiar with are in the technical scope that the present invention discloses, and the change that can expect easily or replacement, all should be encompassed within protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion with the protection domain of claims.
Claims (11)
1. air traffic controller's fatigue detection method, is characterized in that, comprises the following steps:
Step 1, receives the empty communicating data of control seat video monitoring data and land that real-time audio and video all data harvester obtains;
Step 2, carries out detection to the empty communicating data of described control seat video monitoring data and land and analyzes, obtain controller's fatigue index and detect data;
Step 3, detects data sample with reference to preset controller's fatigue index, sets up multiple regression model;
Step 4: the controller's fatigue index in step 2 is detected after data carry out standardization conversion and bring multiple regression model into, obtain current controller's fatigue index and assemble index;
Step 5: if controller's fatigue index assembles index exceed alarm standard, then trigger alerts.
2. method according to claim 1, is characterized in that: the process receiving described control seat video monitoring data in described step 1 is: Real-time Collection video information, after carrying out noise reduction, filtering process, extracts eigenwert.
3. method according to claim 2, is characterized in that: after extracting eigenwert, mate, thus identify the identity of current controller with the template data preserved in video template storehouse.
4. method according to claim 1, it is characterized in that: the process receiving the empty communicating data in described land in described step 1 is: adopt cable to be drawn from distributing frame by voice signal and be connected to speech processor, when voice signal being detected, start recording, there is no sound when detecting and continue 3 seconds, then stopping recording, namely complete and once record, often complete and once record, acoustic information is preserved with audio file formats.
5. method according to claim 1, it is characterized in that: also comprise after step 1 and the empty communicating data of described control seat video monitoring data and land is carried out synchronous step, be specially: with the timestamp of video monitoring data for time reference, the process sound intermediate frequency play plays the current play time data that thread real-time reception video playback thread transmits, and synchronous adjustment audio frequency played data, the empty communicating data of control seat video monitoring data and land is synchronously play, realizes playing in real time and history playback.
6. method according to claim 1, is characterized in that: carry out detection to described control seat video monitoring data in step 2 and analyze, and the controller's fatigue index obtained detects data and comprises: PERCLOS value, average eye-closing period and frequency of yawning.
7. method according to claim 1, it is characterized in that: carry out detection to the empty communicating data in described land in step 2 and analyze, the controller's fatigue index obtained detects data and comprises: channel seizure ratio, talk times, average audio, average loudness of a sound and land, sector sky call keyword occurrence rate.
8. method according to claim 1, is characterized in that: specifically comprise the steps: in step 3
Step 3.1 detects data sample with reference to preset controller's fatigue index, sets up Multivariate regression model and multiple nonlinear regression model (NLRM) respectively, and solves coefficient b
i,
Wherein Multivariate regression model is:
Y＝XB+U
Wherein,
Multiple nonlinear regression model (NLRM) is:
Y＝f[(b
1,b
2,…,b
k)；X
1,X
2,…,X
n]
Wherein dependent variable Y is controller's fatigue index assembly index, and independent variable X is that 8 controller's fatigue indexes detect data samples, x
mn(n=1,2 ... 8), m represents m group statistical data, and U is except n independent variable is on the stochastic error except the impact of dependent variable Y, Normal Distribution, and f represents nonlinear solshing;
The coefficient of determination R that step 3.2 returns according to each model
2value, F inspection, t inspection, verify respectively and compare degree of fitting, the conspicuousness of two kinds of regression models, on the obvious basis of, conspicuousness higher in model-fitting degree, calculate the metrical error of two kinds of regression models, and choose the minimum a kind of model of error, as the multiple regression model that controller's fatigue index detects.
9. method according to claim 8, is characterized in that: the average detecting 8 indexs that data statistics obtains according to m group controller fatigue index
variance s
j, data t is detected to controller's fatigue index
j(j=1,2 ..., 8) carry out standardization conversion:
by the data t after conversion
j' import in regression model, obtain current controller's fatigue index and assemble index.
10. air traffic controller's fatigue detection device, is characterized in that, comprises with lower module:
Receiver module, for receiving the empty communicating data of control seat video monitoring data and land that real-time audio and video all data harvester obtains;
Computing module, analyzing for carrying out detection to the empty communicating data of described control seat video monitoring data and land, obtaining controller's fatigue index and detecting data;
Alarm module, for detecting data sample with reference to preset controller's fatigue index, set up Multivariate regression model and multiple nonlinear regression model (NLRM) respectively, the controller's fatigue index exported by computing module detects after data carry out standardization conversion and brings model into, obtain current controller's fatigue index and assemble index, if controller's fatigue index is assembled index and is exceeded alarm standard, then trigger alerts.
11. 1 kinds of air traffic controller's fatigue detecting system, is characterized in that, this system comprises:
Control seat real-time audio and video all data acquisition platform: for carrying out the acquisition and processing of control seat video monitoring data, the empty communicating data in land;
Controller's fatigue detecting database that the empty call voice of integrated face recognition and land is resolved: for the data of control seat real-time audio and video all data acquisition platform acquisition process are carried out sorting out and preserving, store every controller's fatigue index detection data that controller's fatigue index detection module calculates simultaneously, and the calculating of controller's fatigue strength alarm module and alarm result, thus history of forming statistics;
Controller's fatigue index detection module: carry out detection analyze for the video monitoring data to control seat real-time audio and video all data acquisition platform acquisition and processing, the empty communicating data in land, obtain controller's fatigue index and detect data;
Controller's fatigue strength alarm module: based on the historical statistics result in controller's fatigue detecting database that the empty call voice of integrated face recognition and land is resolved, set up Multivariate regression model and multiple nonlinear regression model (NLRM) respectively, controller's fatigue index is detected after data carry out standardization conversion and bring model into, obtain current controller's fatigue index and assemble index, if controller's fatigue index is assembled index and is exceeded alarm standard, then trigger alerts.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510645187.5A CN105354830B (en) | 2015-09-30 | 2015-09-30 | Controller's fatigue detection method, apparatus and system based on multiple regression model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510645187.5A CN105354830B (en) | 2015-09-30 | 2015-09-30 | Controller's fatigue detection method, apparatus and system based on multiple regression model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105354830A true CN105354830A (en) | 2016-02-24 |
| CN105354830B CN105354830B (en) | 2019-04-09 |
Family
ID=55330797
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510645187.5A Active CN105354830B (en) | 2015-09-30 | 2015-09-30 | Controller's fatigue detection method, apparatus and system based on multiple regression model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105354830B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109214693A (en) * | 2018-02-13 | 2019-01-15 | 中国民用航空总局第二研究所 | Appraisal procedure, system, medium and the equipment of air traffic controller's fatigue risk |
| CN110322371A (en) * | 2019-07-08 | 2019-10-11 | 长沙理工大学 | The area Gao Suntai multiplexing electric abnormality user based on multiple linear regression analysis detects localization method |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102014278A (en) * | 2010-12-21 | 2011-04-13 | 四川大学 | Intelligent video monitoring method based on voice recognition technology |
| CN104636890A (en) * | 2015-03-13 | 2015-05-20 | 中国民航大学 | Measurement method for workload of air traffic controller |
-
2015
- 2015-09-30 CN CN201510645187.5A patent/CN105354830B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102014278A (en) * | 2010-12-21 | 2011-04-13 | 四川大学 | Intelligent video monitoring method based on voice recognition technology |
| CN104636890A (en) * | 2015-03-13 | 2015-05-20 | 中国民航大学 | Measurement method for workload of air traffic controller |
Non-Patent Citations (3)
| Title |
|---|
| 张建平等: "终端区空中交通管制运行品质综合评价", 《西南交通大学学报》 * |
| 汪磊 等: "基于面部特征识别的管制员疲劳监测方法研究", 《中国安全科学学报》 * |
| 裴成功等: "管制员工作负荷评估的回归分析方法", 《南京航空航天大学学报》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109214693A (en) * | 2018-02-13 | 2019-01-15 | 中国民用航空总局第二研究所 | Appraisal procedure, system, medium and the equipment of air traffic controller's fatigue risk |
| CN110322371A (en) * | 2019-07-08 | 2019-10-11 | 长沙理工大学 | The area Gao Suntai multiplexing electric abnormality user based on multiple linear regression analysis detects localization method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105354830B (en) | 2019-04-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105261152B (en) | Air traffic controller's fatigue detection method based on cluster analysis, device and system | |
| CN105225420B (en) | Air traffic controller's fatigue detection method based on principal component analysis, device and system | |
| CN105184315B (en) | A kind of quality inspection processing method and system | |
| CN115272037B (en) | Smart city based on Internet of things regional public security management early warning method and system | |
| CN105595990A (en) | Intelligent terminal device for evaluating and distinguishing quality of electrocardiosignal | |
| CN107742127A (en) | A kind of improved anti-electricity-theft intelligent early-warning system and method | |
| CN109171707A (en) | A kind of intelligent cardiac figure classification method | |
| CN105679313A (en) | Audio recognition alarm system and method | |
| CN106255116A (en) | A kind of recognition methods harassing number | |
| CN109034036A (en) | A kind of video analysis method, Method of Teaching Quality Evaluation and system, computer readable storage medium | |
| CN107197331A (en) | A kind of method and device of real-time monitoring live content | |
| CN106847293A (en) | Facility cultivation sheep stress behavior acoustical signal monitoring method | |
| CN106650605B (en) | A kind of Morse automatic signal detection interpretation method based on machine learning | |
| CN102623009A (en) | Abnormal emotion automatic detection and extraction method and system on basis of short-time analysis | |
| CN104246713B (en) | User sensory quality estimation device, main quality dropping reason determination device, similar operation extraction device, and methods therefor | |
| CN109714324A (en) | User network abnormal behaviour based on machine learning algorithm finds method and system | |
| CN112134871A (en) | Abnormal flow detection device and method for energy internet information support network | |
| CN105139029A (en) | Activity recognition method and activity recognition device for persons serving sentences | |
| CN102929887A (en) | Quick video retrieval method and system based on sound feature identification | |
| CN109512390A (en) | Sleep stage method and wearable device based on EEG time domain various dimensions feature and M-WSVM | |
| CN103258039A (en) | Method for detecting false information on microblog | |
| CN112330095A (en) | Quality management method based on decision tree algorithm | |
| CN105354830A (en) | Method, apparatus and system for controller fatigue detection based on multiple regression model | |
| CN113707175B (en) | Acoustic event detection system based on feature decomposition classifier and adaptive post-processing | |
| CN105282502B (en) | Air traffic controller's fatigue detection method based on confidence interval, device and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |