CN118298588A - Safety belt feedback monitoring system and method based on attitude sensor - Google Patents

Abstract

The invention relates to the technical field of construction site application systems, and solves the technical problems that when a construction worker has dangerous behaviors and a safety belt fails, early warning cannot be well carried out and corresponding measures can be taken in real time, in particular to a safety belt feedback monitoring system and a safety belt feedback monitoring method based on an attitude sensor, wherein the safety belt feedback monitoring system is used for sensing the attitude information of a detected target and carrying out reasonable feedback according to the attitude information to output, and the safety belt feedback monitoring system comprises: the information generation module is used for generating first posture information of the first detected target and transmitting the first posture information to the next module; the gesture sensor module is used for sensing the first gesture information generated by the information generation module and processing the first gesture information to obtain gesture data. The invention reduces the waste of manpower and material resources caused by repeated monitoring, can early warn in real time, and has great significance for reducing the incidence rate of site accidents.

Description

Safety belt feedback monitoring system and method based on attitude sensor

Technical Field

The invention relates to the technical field of construction site application systems, in particular to a safety belt feedback monitoring system and method based on an attitude sensor.

Background

The intelligent construction site is used for accurately designing and constructing simulation of engineering projects through a three-dimensional design platform by using an informatization means, and simultaneously, the construction project informatization ecological circle of interconnection cooperation, intelligent production and scientific management is established around construction process management, so that the visual intelligent management of engineering construction is realized, the informatization level of engineering management is improved, and green construction and ecological construction are realized.

The construction site is used as a place where accidents are high, and the safety protection of the construction site is paid attention to. However, the use of existing security precautions is not optimistic. It is counted that many operators who have high falling accidents in the construction operation area are caused by not wearing safety belts. The intelligent safety belts on the market are also mostly lack of safety belt monitoring systems based on actual scenes, when construction workers have dangerous behaviors and the safety belts themselves have faults, early warning cannot be well carried out, corresponding measures are adopted in real time, great trouble is caused to construction companies and related responsible persons, and the progress of various works on the construction sites is seriously affected.

Disclosure of Invention

Aiming at the defects of the prior art, the invention innovatively provides a safety belt feedback monitoring system and a safety belt feedback monitoring method based on an attitude sensor, which solve the technical problems that early warning cannot be well carried out and corresponding measures can be taken in real time when a constructor has dangerous behaviors and the safety belt fails.

In order to solve the technical problems, the invention innovatively provides the following technical scheme: a seatbelt feedback monitoring system based on an attitude sensor for sensing attitude information of a detected object and outputting according to the attitude information making rational feedback, the seatbelt feedback monitoring system comprising:

The information generation module is used for generating first posture information of the detected target for the first time and transmitting the first posture information to the next module;

The gesture sensor module is used for sensing the first gesture information generated by the information generation module and processing the first gesture information to obtain gesture data;

The operation discrimination module is connected with the attitude sensor module and acts on the attitude sensor module to classify the attitude data and perform binarization processing on the attitude data;

if the gesture data is higher than the preset threshold value, outputting the gesture data through the display module;

If the position data is lower than the preset threshold value, the position data is re-uploaded to the information generation module as a feedback signal, and the information generation module re-sends second position information of the detected target for the second time after receiving the feedback signal and transmits the second position information to the position sensor module until the binarization processing result is higher than the preset threshold value, and the position data is output through the display module.

Further, the attitude sensor module is connected with the information generation module in a wireless communication mode, and the attitude sensor is adopted as the attitude sensor module to be assembled on the intelligent safety belt.

Further, the attitude sensor module obtains the aligned attitude information in an oriented manner so as to obtain the generated numerical information, wherein the numerical information comprises Euler angles of a space where the detected target is located.

Further, the operation discrimination module comprises an analysis unit for estimating and reasoning the gesture data, a classification unit for classifying the gesture data according to the result output by the analysis unit, and a discrimination unit for deciding the rationalization of the gesture data so as to generate display data and feedback data.

Further, the estimating and reasoning of the gesture data comprises key point detection, gesture regression and deep learning, and the rationalization decision comprises cross verification of the gesture data and binarization reasoning assignment of the gesture data.

Further, the classification unit classifies the attitude data by using a support vector machine model SVM algorithm.

Further, the gesture data is re-uploaded to the information generation module as a feedback signal, and the feedback signal is filtered by adopting a Kalman filtering algorithm EKF in the re-uploading process.

The technical scheme also innovatively provides a method for realizing the safety belt feedback monitoring system, which comprises the following steps:

S1, acquiring first posture information of a first detected target generated by an information generation module;

s2, sensing the first gesture information through a gesture sensor module to obtain information to be distinguished;

s3, classifying the information to be judged by adopting a Support Vector Machine (SVM) model algorithm, and re-uploading unreasonable information to be judged to an information generation module as a feedback signal;

S4, improving the signal-to-noise ratio of the feedback signal by adopting a Kalman filtering algorithm EKF, and resending second posture information of the detected target for the second time by utilizing the information generating module and returning to the step S2 until the information to be distinguished is reasonable;

and S5, sending the reasonable judging information to the display module.

By means of the technical scheme, the invention innovatively provides a safety belt feedback monitoring system and method based on an attitude sensor, and the safety belt feedback monitoring system and method at least have the following beneficial effects:

1. According to the invention, the attitude sensor module is used for processing information, so that each index of a target to be detected can be acquired more accurately, the information conversion efficiency is greatly improved, the stability and accuracy of attitude information acquisition are improved, the waste of manpower and material resources caused by repeated monitoring for many times is reduced, real-time early warning can be realized, and the method has great significance in reducing the occurrence rate of site accidents.

2. The invention adopts unreasonable gesture information as feedback signals, marks unreasonable dangerous gesture information by using a data binarization method and prevents the unreasonable dangerous gesture information from being output, and continuous information feedback discrimination achieves the effect of filtering gesture information. The method has the advantages of greatly improving the monitoring efficiency, greatly reducing the possibility of danger occurrence, dynamically and real-timely monitoring and recording the gesture data information generated by the object, along with good practicability, universality and wide application prospect.

3. According to the invention, by continuously carrying out feedback judgment on the gesture data, unreasonable dangerous gesture information is screened out, so that monitoring personnel can take measures in advance on possible abnormal conditions, the monitoring efficiency is greatly improved, and the cost of manpower, material resources and the like consumed in the monitoring process is reduced.

Drawings

The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the application and do not constitute a limitation on the application. In the drawings:

FIG. 1 is a schematic block diagram of a seat belt feedback monitoring system of the present invention;

FIG. 2 is a schematic diagram of the structure of the attitude sensor module of the present invention;

FIG. 3 is a schematic diagram of a support vector machine of the present invention;

FIG. 4 is a schematic diagram of a support vector machine of the present invention;

FIG. 5 is a diagram of an example of the belt feedback monitoring system of the present invention as applied to a worksite scenario;

Fig. 6 is a flowchart of a seat belt feedback monitoring method of the present invention.

Detailed Description

In order that the above-recited objects, features and advantages of the present application will become more readily apparent, a more particular description of the application will be rendered by reference to the appended drawings and appended detailed description. Therefore, the realization process of how to apply the technical means to solve the technical problems and achieve the technical effects can be fully understood and implemented.

In the prior art, a plurality of operators who have high falling accidents in a building construction operation area are caused by not wearing safety belts. Most of the intelligent safety belts on the market lack safety belt monitoring systems based on actual scenes, when a construction worker fails with dangerous behaviors and the safety belt itself, early warning cannot be well carried out and corresponding measures are adopted in real time, please refer to fig. 1-6, the embodiment provides a safety belt feedback monitoring system based on an attitude sensor, the monitoring system is based on hardware facilities such as a sensor and various algorithms, various indexes of a target to be detected can be acquired more accurately through processing of information by an attitude sensor module, the efficiency of information conversion is greatly improved through analysis, estimation and feedback of an operation judging module, the stability and accuracy of attitude information acquisition are improved, waste of manpower and material resources caused by repeated monitoring is reduced, and the safety belt feedback monitoring system is capable of carrying out early warning in real time and has great significance in reducing the incidence of site accidents. The safety belt feedback monitoring system is used for sensing the gesture information of the detected target and outputting according to the gesture information by rationalized feedback, please refer to fig. 1, and is composed of an information generating module, a gesture sensor module, an operation discriminating module and a display module.

The information generating module is used for generating first gesture information of the detected target for the first time and transmitting the first gesture information to the next module, and in the embodiment, the gesture sensor module is loaded on the intelligent safety belt and worn by construction personnel on the construction site. For example, a construction staff wears an intelligent safety belt provided with an attitude sensor module, and after the intelligent safety belt enters a working state, the attitude sensor module can monitor the information of the attitude of the construction staff in real time, generate corresponding numerical value information and immediately upload the numerical value information to an operation judging module for further processing.

The attitude sensor module is connected with the information generation module in a wireless communication mode, the attitude sensor module which is used as the attitude sensor module is assembled on the intelligent safety belt, the attitude sensor module is used for sensing first attitude information generated by the information generation module and processing the first attitude information to obtain attitude data, the attitude sensor module obtains the tidied attitude information in an oriented mode so as to obtain generated numerical information, and the numerical information comprises Euler angles of a space where a detected target is located.

In this embodiment, referring to fig. 2, the posture sensor is an MPU6050, and belongs to a six-axis posture sensor, and the accelerometer is composed of a three-axis accelerometer and a three-axis gyroscope, and is used for measuring acceleration of constructors, and the gyroscope is used for measuring angular velocity of constructors. The measurement and calculation of the accelerometer have static stability, and the measurement and calculation of the gyroscope have dynamic stability, so that the two complementary actions perform attitude settlement. Because the site area is complex and uncertain factors are more, the method greatly improves the efficiency of information conversion, improves the stability and accuracy of gesture information acquisition, and has extremely strong operability.

In this embodiment, the acceleration and angular velocity of the target to be measured may be calculated by respectively solving the sensitivities of the triaxial accelerometer and the triaxial gyroscope. The result of the gesture calculation of (2) comprises Euler angles of the space where the target to be measured is located. The information to be distinguished obtained by the gesture sensor through gesture calculation is uploaded to an operation distinguishing module for gesture estimation and gesture analysis.

The operation discrimination module comprises an analysis unit for estimating and reasoning the gesture data, a classification unit for classifying the gesture data according to the output result of the analysis unit, and a discrimination unit for deciding the rationalization of the gesture data so as to generate display data and feedback data, wherein the estimation and reasoning of the gesture data comprises key point detection, gesture regression and deep learning, the rationalization decision comprises cross validation of the gesture data and binary reasoning assignment of the gesture data, the classification unit classifies the gesture data by adopting a support vector machine model SVM algorithm, and the operation discrimination module comprises a plurality of integrated operation tools which are deployed on the wearable equipment and are used for real-time transmission communication with a gesture sensor module and can reasonably judge and infer information to be discriminated.

The support vector machine (Support Vector Machine, SVM) is a commonly used machine learning algorithm, and SVM can be used for linear and nonlinear classification problems, regression and outlier detection. The basic principle is to separate samples of different classes by finding one hyperplane in the feature space and to maximize the distance from the closest sample point to the hyperplane.

Both fig. 3 and fig. 4 are schematic diagrams of support vector machines, taking a two-dimensional plane as an example, the decision boundary is a hyperplane, in fig. 4 it is actually a line, but it can be imagined as a mapping of a plane or even higher dimensional form in the two-dimensional plane, which is determined by support vectors, which are the closest sample points to the decision boundary, which determine the position of the decision boundary. The middle of the interval is the judgment boundary, and the interval distance represents the difference of two types of data. And specifies strictly that all sample points are not in the "buffer" and are correctly on both sides, called hard interval classification.

In complex scenarios at the actual worksite, in order to avoid the contribution of linearly separable pose data and the disturbance of abnormal noise values, soft interval classification may be used. A good balance is found between keeping the "buffer" as large as possible and avoiding interval violations, and the SVM class in sklearn can use the hyper-parameter C (penalty factor) to control the complexity and fault tolerance of the model. A larger C value results in a lower fault tolerance and may result in less misclassification. Therefore, the operation discrimination module C can make a more accurate decision, and unreasonable information is fed back in time.

According to the embodiment, a Support Vector Machine (SVM) algorithm based on a machine learning method is applied to classification of system attitude data, and because the prior system mostly adopts a neural network classifier, the system has the defects that partial classification functions are lost and local optimality can only be guaranteed due to the fact that fitting phenomenon occurs in large sample learning and operation processes, the SVM can separate samples of different types by finding a hyperplane in a feature space, and punishment parameters C can be set to avoid fitting, so that the system is an algorithm specially aiming at small samples and high-dimension mode problems, and the defects of the neural network classifier are effectively solved.

The operation discriminating module is connected with the attitude sensor module, wherein the operation discriminating module is arranged at the front end personnel, interaction between the two modules is formed by adopting a wireless communication mode, the operation discriminating module acts on the attitude sensor module to classify the attitude data and carry out binarization processing on the attitude data, and then rationalization judgment is carried out on the binarized attitude data, namely:

If the gesture data is higher than the preset threshold, marking the gesture data as 1, and outputting the gesture data through a display module;

If the position data is lower than the preset threshold value, the position data is marked as 0, the position data is re-uploaded to the information generation module as a feedback signal, the feedback signal is subjected to filtering processing by adopting a Kalman filtering algorithm EKF in the re-uploading process, the information generation module receives the feedback signal and then re-sends second position information of the detected target for the second time and transmits the second position information to the position sensor module until the binarization processing result is higher than the preset threshold value, and the position data is output through the display module.

In this embodiment, because the signal may be distorted due to interference of environmental noise in a complex scenario, the algorithm is mainly used for measuring noise in a complex environmental scenario, and the method is characterized in that the method can be used for calculation by a recursive method, and the required data storage amount is smaller, so that real-time processing is convenient, and the quality of a feedback signal is improved. Therefore, in order to prevent the feedback signal from distortion under nonlinear environment interference, the feedback signal is processed by using a kalman filter algorithm EKF, so that noise in a complex environment situation is measured, so as to weaken and even eliminate the noise, and improve the signal to noise ratio of the feedback signal. And after receiving the feedback signal, the information generating module resends the second detected gesture information and transmits the second detected gesture information to the gesture sensor module, and the process is repeated until the binarization processing result is higher than a preset threshold value, and the gesture data is output through the display module, so that the whole monitoring process is completed.

The kalman filter (Extended KALMAN FILTER) algorithm means that the best estimate of the current moment is linearly corrected according to the observed value of the current moment based on the best estimate of the previous moment. Kalman filtering is a mathematically linear minimum variance statistical estimation method that obtains the best estimate of a physical parameter by processing a series of actual measured data with errors. The essential problem to be solved is to find the estimate under minimum mean square error. The method is characterized in that the method can be calculated by a recursive method, the required data storage amount is small, and the real-time processing is convenient.

In particular, the core idea of the kalman filter algorithm is to give different weights to the information. These weights reflect the reliability and relevance of the data to refine more accurate, more reliable data. This weighting process can be considered a "filtering" operation. The kalman filter algorithm processes data in two basic steps: prediction and updating.

The predicting step comprises the following steps: the current state is predicted based on the previous state and model of the system. At this step, the filter generates a state estimate and a prediction of its uncertainty (covariance).

The prediction process comprises the following steps:

；

；

In the method, in the process of the invention, Is the actual value of the current period of the signal; Is a coefficient constant; To handle noise; The prediction error of the period is; Is covariance.

The updating step comprises the following steps: when new observations are available, the filter updates its state estimate. This step involves calculating the kalman gain (a weight factor) that determines the degree of confidence in the final estimate in the predicted state and the new observed data. If the uncertainty of the observed data is small (i.e., the signal noise is low), the filter will be more dependent on the new data. Conversely, if the accuracy of the prediction is higher, the filter will rely more on its prediction to update the data.

The updating process comprises the following steps:

；

；

；

In the method, in the process of the invention, Is Kalman gain; Is a scaling factor; Is the average value of the measured noise; Is a signal measurement.

The feedback structure added with the attitude information creatively in the safety belt feedback monitoring system provided by the embodiment classifies and processes the attitude data through the operation discrimination module, the unreasonable dangerous attitude information is marked and prevented from being output by a data binarization method by using the feedback structure, and the effect of filtering the attitude information is achieved by continuous information feedback discrimination. The method has the advantages of greatly improving the monitoring efficiency, greatly reducing the possibility of danger occurrence, dynamically and real-timely monitoring and recording the gesture data information generated by the object, along with good practicability, universality and wide application prospect.

Referring to fig. 6, the present embodiment provides a safety belt feedback monitoring method based on an attitude sensor, which includes the following steps:

S1, acquiring first posture information of a first detected target generated by an information generation module;

s2, sensing the first gesture information through a gesture sensor module to obtain information to be distinguished;

s3, classifying the information to be judged by adopting a Support Vector Machine (SVM) model algorithm, and re-uploading unreasonable information to be judged to an information generation module as a feedback signal;

S4, improving the signal-to-noise ratio of the feedback signal by adopting a Kalman filtering algorithm EKF, and resending second posture information of the detected target for the second time by utilizing the information generating module and returning to the step S2 until the information to be distinguished is reasonable;

and S5, sending the reasonable judging information to the display module.

According to the safety belt feedback monitoring system and the safety belt feedback monitoring method, the gesture information of the target to be detected can be obtained more effectively and more accurately based on the gesture sensor module, so that more reliable gesture data are provided for the operation discrimination module; the unreasonable dangerous posture information generated by the object to be detected can be accurately identified through the binarization assignment processing in the operation discrimination module in the monitoring system, and feedback information is generated. The signal to noise ratio of feedback information under complex environmental situations can be improved through a filtering algorithm, so that the quality of feedback signals is further improved; through the continuous feedback judgment, unreasonable dangerous posture information is screened out, so that monitoring personnel can take measures in advance for possible abnormal conditions, the monitoring efficiency is greatly improved, and the cost of manpower, material resources and the like consumed in the monitoring process is also reduced.

It will be appreciated that the structure illustrated in the embodiments of the present application does not constitute a specific limitation on the feedback monitoring system and the feedback monitoring method. In embodiments of the application, more or fewer components than shown may be included, or certain components may be combined, or certain components may be split, or different arrangements of components. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

Embodiments of the present application may be implemented as a computer program or program code that is executed on a programmable system having a plurality of adaptations of mechanisms, said computer program code also being writable using a variety of programming languages, such as: in some embodiments, the machine learning based SVM algorithm may be written using the Python language and the EKF algorithm may be written using the Matlab language, and in fact, the mechanisms described in the present application are not limited in scope to any particular programming language.

Those of ordinary skill in the art will appreciate that all or a portion of the steps in a method of implementing an embodiment described above may be implemented by a program to instruct related hardware, and thus, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The foregoing embodiments have been presented in a detail description of the invention, and are presented herein with a particular application to the understanding of the principles and embodiments of the invention, the foregoing embodiments being merely intended to facilitate an understanding of the method of the invention and its core concepts; meanwhile, as those skilled in the art will have variations in the specific embodiments and application scope in accordance with the ideas of the present invention, the present description should not be construed as limiting the present invention in view of the above.

Claims (8)

1. A safety belt feedback monitoring system based on an attitude sensor, wherein the safety belt feedback monitoring system is configured to sense attitude information of a detected target and make rational feedback for output according to the attitude information, the safety belt feedback monitoring system comprising:

The information generation module is used for generating first posture information of the detected target for the first time and transmitting the first posture information to the next module;

The gesture sensor module is used for sensing the first gesture information generated by the information generation module and processing the first gesture information to obtain gesture data;

The operation discrimination module is connected with the attitude sensor module and acts on the attitude sensor module to classify the attitude data and perform binarization processing on the attitude data;

if the gesture data is higher than the preset threshold value, outputting the gesture data through the display module;

If the position data is lower than the preset threshold value, the position data is re-uploaded to the information generation module as a feedback signal, and the information generation module re-sends second position information of the detected target for the second time after receiving the feedback signal and transmits the second position information to the position sensor module until the binarization processing result is higher than the preset threshold value, and the position data is output through the display module.

2. The seat belt feedback monitoring system of claim 1, wherein the attitude sensor module is connected to the information generating module by wireless communication, and the attitude sensor is used as the attitude sensor module to be mounted on the intelligent seat belt.

3. The seat belt feedback monitoring system of claim 1, wherein the attitude sensor module performs directional acquisition of the consolidated attitude information to derive generated numerical information including calculating euler angles of a space in which the detected object is located.

4. The seatbelt feedback monitoring system according to claim 1, wherein the operation discriminating module includes an analyzing unit for estimating and inferring the posture data, and a classifying unit for classifying the posture data according to a result outputted by the analyzing unit, and a discriminating unit for deciding rationalization of the posture data so as to generate display data and feedback data.

5. The seat belt feedback monitoring system of claim 4, wherein the estimating and reasoning of the pose data comprises keypoint detection, pose regression, and deep learning, and the rationalizing decision comprises cross-validation of the pose data and binary reasoning assignment of the pose data.

6. The seat belt feedback monitoring system of claim 4, wherein the classification unit classifies the pose data using a support vector machine model SVM algorithm.

7. The seat belt feedback monitoring system of claim 4 wherein the attitude data is re-uploaded to the information generating module as a feedback signal, the feedback signal being filtered during the re-uploading using a kalman filter algorithm EKF.

8. A method for implementing a seat belt feedback monitoring system according to any of the preceding claims 1-7, characterized in that the method comprises the steps of:

S1, acquiring first posture information of a first detected target generated by an information generation module;

s2, sensing the first gesture information through a gesture sensor module to obtain information to be distinguished;

s3, classifying the information to be judged by adopting a Support Vector Machine (SVM) model algorithm, and re-uploading unreasonable information to be judged to an information generation module as a feedback signal;

S4, improving the signal-to-noise ratio of the feedback signal by adopting a Kalman filtering algorithm EKF, and resending second posture information of the detected target for the second time by utilizing the information generating module and returning to the step S2 until the information to be distinguished is reasonable;

and S5, sending the reasonable judging information to the display module.