Mine transportation safety adaptive optimization system and method based on video monitoring

Abstract

The application discloses a mine transportation safety self-adaptive optimization system and method based on video monitoring and belongs to the technical field of real-time optimization. The application utilizes various sensors to sense the environment in a monitoring area to obtain multi-source environment data in the monitoring area; analyzes and processes the environment sensed by the sensors to extract features of the environment in the monitoring area to obtain first features of the monitoring area; analyzes the environment in the monitoring area to calculate an abnormal environment index in the monitoring area to realize hierarchical early warning of the abnormal environment in the monitoring area; utilizes the first features in the monitoring area obtained by calculation to perform machine learning to construct an abnormal early warning model; analyzes the environment during early warning to extract second features of the monitoring area; calculates linear features of the first features and the second features to realize real-time optimization of early warning classification according to the linear features of the first features and the second features obtained by calculation.

Description

Technical Field

[0001] This invention relates to the field of real-time optimization technology, specifically to an adaptive optimization system and method for mine transportation safety based on video surveillance. Background Technology

[0002] With the advancement of smart mines, mine transportation safety has become a core concern. Traditional video surveillance systems are insufficient in handling massive amounts of data, complex environments, and real-time requirements, making it difficult to meet the efficient and safe operational needs of modern mines. With the development of science and technology, intelligent video surveillance technology is increasingly being applied to mine transportation safety. By using visual sensors, microphones, temperature and humidity sensors, etc., intelligent video surveillance can monitor various data in mine transportation scenarios. With the development of modern information technology, intelligent video surveillance plays a crucial role in mine transportation; by collecting and monitoring various data during mine transportation, it provides comprehensive safety checks and ensures that when abnormal situations occur, timely feedback can be provided to staff for handling. However, due to the complex and ever-changing working environment of mines, many data points change frequently, and the data interact with each other. The criteria for abnormal occurrences during mine transportation are constantly changing, resulting in insufficient timely warnings from intelligent video surveillance technology when the mine transportation environment changes, and an inability to distinguish the influence relationships between data points during mine transportation. Therefore, this invention designs an adaptive optimization system and method for mine transportation safety based on video surveillance to address the above problems. Summary of the Invention

[0003] The purpose of this invention is to provide an adaptive optimization system and method for mine transportation safety based on video surveillance, so as to solve the problems mentioned in the background art.

[0004] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0005] An adaptive optimization method for mine transportation safety based on video surveillance, the method comprising the following steps:

[0006] S100: Utilize multiple sensors to perceive the environment within the monitoring area and obtain multi-source environmental data within the monitoring area;

[0007] Furthermore, the sensors employ a visual sensor, a temperature and humidity sensor, and a microphone sensor, and the multi-source environmental data includes temperature and humidity data, image data, and sound data within the monitored area.

[0008] S200: Process and analyze the environmental data sensed by the sensor and extract features to obtain the first feature of the monitored area;

[0009] Further, the specific steps for feature extraction of the environment in the monitoring area to obtain the first feature of the monitoring area are as follows:

[0010] S201, collecting and classifying the environment data in the monitoring area without abnormality using multiple sensors, and dividing the data collected by different sensors into different categories according to the different sensors to generate n kinds of data respectively , , which are the 1st, 2nd, 3rd, …, nth data, and n is a positive integer; the abnormality in the monitoring area indicates that the environment data in the monitoring area has multiple data eruption changes.

[0011] S202, recording the environment data in the monitoring area with abnormality, classifying the data, dividing the data collected by different sensors into different categories according to the different sensors to generate m kinds of abnormal environment data respectively , , which are the 1st, 2nd, 3rd, …, mth abnormal environment data, and m is a positive integer;

[0012] S203, after classifying and collecting the environment data in the monitoring area without abnormality and with abnormality, comparing the data in the two cases, using the environment data in the monitoring area without abnormality to the environment data in the monitoring area with abnormality , deleting the same values of the environment data in the monitoring area with abnormality and the environment data in the monitoring area without abnormality , and setting the deleted environment data with abnormality as difference data , , which are the 1st, 2nd, 3rd, …, jth difference data, and j is a positive integer;

[0013] S204, according to the above steps, extracting k times of environment data with abnormality in the monitoring area, then comparing the environment data with abnormality in the monitoring area with the environment data without abnormality in the monitoring area for k times to generate k times of difference data, and statistically analyzing the k times of difference data to extract the data category with the most repeated times in the k times of difference data , the subscript y takes values from 1 to j, and is set as the first feature of the monitoring area.

[0014] The first feature generated above reflects the most significant change characteristics of the environment data when the abnormality occurs in the monitoring area, and the first feature can be used to monitor whether the abnormality occurs in the monitoring area.

[0015] S300: Analyze environmental data within the monitoring area, calculate the rate of change of abnormal environmental data within the monitoring area, and realize graded early warning of anomalies within the monitoring area;

[0016] Furthermore, the steps to achieve tiered early warning within the monitored area are as follows:

[0017] S301. Based on the collection of abnormal environmental data within the monitored area, and after comparing the abnormal environmental data with the abnormal environmental data within the monitored area k times using the method in S200, the number of k difference data points is generated as follows: , This represents the number of discrepancies between the number of environmental data points with and without abnormalities within the monitoring area during the 1st, 2nd, 3rd...kth monitoring cycles, where k is a positive integer.

[0018] S302. Arrange the first feature in ascending order. The numerical values ​​are sorted, and the first feature is plotted using the point-plotting method. Number of numerical and differential data A line graph, where p takes values ​​from 1 to k;

[0019] S303. Normalize and fit the line graph to form a smooth function curve. Using a piecewise method, calculate the functional relationship of each segment of the function curve to generate a set of functions. , Let h be the function in segments 1, 2, 3...h of the function curve, where h is a positive integer;

[0020] S304. Perform derivative operations on the piecewise functions separately to obtain the derivative function of each segment of the function curve. , Let h be the derivative of the function curve for segments 1, 2, 3...h, where h is a positive integer; the derivative represents the growth rate of each segment of the function curve.

[0021] S305. Filter the derivative functions obtained above, selecting those with derivative values ​​greater than 0 and sorting them in ascending order. Obtain the corresponding independent variables based on the sorting of the derivative functions. The sorting is , Let the independent variables be the 1st, 2nd, 3rd...ath derivatives that are greater than 0, arranged in ascending order of their derivative values. , where a is a positive integer; calculate the... The data is classified as a warning level;

[0022] The larger the value of the derivative of a curvilinear function, the more likely it is that the independent variable... The faster the variable changes, the faster the increase; it can reflect that when an anomaly occurs in the monitoring area, the changes in the remaining environmental data in the monitoring area are judged by judging the changes in the first characteristic value, the greater the derivative, the faster the speed of the impact caused by the anomaly in the monitoring area, and therefore setting the warning level according to the size of the derivative can effectively classify the speed of the impact of the anomaly in the monitoring area on the environment, and reasonably allocate resources.

[0023] S306, according to the calculation, the classification data is obtained The monitoring area is classified and warned, and the initial value of the independent variable in the broken line graph is selected as the minimum level of the warning _min, and then the warning level is increased in turn according to the order.

[0024] The above-mentioned classification data of the warning in the monitoring area is calculated according to the derivative. After an anomaly occurs in the monitoring area, the speed of the increase in the number of the remaining changed environmental data caused by the change of the first characteristic is also different, and the value of the derivative represents the increase speed of the remaining changed environmental data in the monitoring area. The anomaly is classified and warned according to the different increase speeds. For example: when monitoring whether a fire is likely to occur in mine transportation, in the monitoring area, the temperature is the first characteristic. When the temperature is too high, it is judged that a fire anomaly occurs, and after the fire occurs, the remaining data in the monitoring area will change. The increase speed of the number of changed data proves the size of the fire, so that the classification warning of the anomaly can be realized.

[0025] S400, machine learning is performed on the first characteristic in the monitoring area to construct an anomaly warning model;

[0026] Further, the construction of the anomaly warning model is specifically:

[0027] The classification data Machine learning is performed on the first characteristic value _s in the environmental data in the monitoring area by using a convolutional neural network algorithm to construct a classification warning model; the first characteristic value _s is input into the constructed classification warning model, and it is judged that the environment in the monitoring area is abnormal when the first characteristic value _s> _min, and then _s is compared with the classification data in turn, and when it is judged that _s> , a corresponding level of warning is issued to the monitoring area, and i belongs to 1-a. ​​

[0028] S500, analyze the environment at the early warning time, and extract the second feature of the monitoring area;

[0029] Further, the specific steps of extracting the second feature of the monitoring area are:

[0030] S501, collect data of the environment in the monitoring area that occurs abnormally, extract k times of the environment data that occurs abnormally in the monitoring area according to the method described in S200, classify each extracted environment data, and generate m kinds of abnormal environment data respectively , extract the first feature in the abnormal environment data, and obtain k values of the first feature after k times of extraction , is the value of the first feature in the environment data that occurs abnormally for the first, second, third,..., k times, and k is a positive integer; a curve graph is drawn by using the dotting method, taking the value of the first feature in the abnormal environment data as the independent variable, and taking the values of other kinds of data as the variables , ; m-1 curves are drawn in the graph ; m-1 curves are drawn in the graph

[0031] S502, analyze the drawn curve graph, and calculate the function of the m-1 curves in the curve graph by using the piecewise method, and the function corresponding to each curve is , is the function of the first, second, third,..., m-1 curves, and m-1 is a positive integer

[0032] S503, derivative operation is performed on the m-1 functions calculated above, and the derivative function is obtained , is the derivative function of the first, second, third,..., m-1 curve functions, and m-1 is a positive integer

[0033] S504, according to the above method, collect the environment data that does not occur abnormally in the monitoring area, classify and generate corresponding n kinds of data respectively , draw a curve graph, calculate a curve function, and finally obtain the derivative function of each kind of data in the environment data that does not occur abnormally , is the derivative function of the first, second, third,..., n-1 curve functions, and n-1 is a positive integer

[0034] S505, analyze the two groups of derivative functions calculated above, and extract the functions in which the values of the derivative functions exist only in one positive or negative case, specifically: in the abnormal environment data, only the function curve in which ≥0 or ≤0 is extracted, and in the environment data that does not occur abnormally, only the function curve in which ≥0 or a function curve with ≤0,

[0035] S506, compare the two groups of functions after extraction, and select the data type x corresponding to the variable value in the curve with the same function as the second feature in the monitoring area.

[0036] By screening the curves in the graph, when the derivative function of the curve only has one change, it is proved that this data may have a certain linear relationship with the first feature, and by comparing the curves in the monitoring area with and without abnormality, the same change curve is found in the screened curve. The data type that can linearly affect the first feature in the monitoring area can be located, and the second feature in the monitoring area can be located.

[0037] S600, calculate the linear feature of the first feature and the second feature, and realize real-time optimization of hierarchical early warning according to the calculated linear feature of the first feature and the second feature.

[0038] Further, the linear feature of the first feature and the second feature is specifically:

[0039] According to the above S500, in the process of calculating the second feature, the inverse function of the curve function corresponding to the selected second feature is defined as the linear relationship between the first feature and the second feature, and the linear relationship function between the first feature and the second feature is set as , in which is the value of the first feature, x is the value of the second feature, is the linear relationship between the first feature and the second feature.

[0040] The specific process of realizing real-time optimization of early warning classification is:

[0041] The data of the second feature in the monitoring area is collected by using multiple sensors to obtain the real-time value x_s of the second feature in the monitoring area, and according to the linear relationship function , the value x_s of the second feature is substituted into the function as the independent variable to calculate the theoretical minimum early warning value of the first feature _b, the calculated theoretical minimum early warning value of the first feature is compared with the actual minimum early warning value _min, and the difference _c= _b _min, the minimum early warning value _min is replaced by the theoretical minimum early warning value _b, and the hierarchical data of early warning are all replaced by the difference ​​The addition operation is performed to obtain new grading data of the early warning.

[0042] When early warning is performed in the monitoring area, other data other than the first feature of the judgment standard has an impact on the early warning standard, resulting in the inability to timely early warn abnormal conditions in the monitoring area. For example, when a fire occurs in mine transportation, the temperature is set as the first feature as the early warning standard, but the humidity in the air will still affect the temperature standard of the fire.

[0043] The mine transportation safety self-adaptive optimization system based on video monitoring comprises an environment perception module, a feature extraction module, a grading early warning module, and an intelligent optimization module.

[0044] The environment perception module collects environmental data in the monitoring area using various sensors.

[0045] The feature extraction module is used to extract the first feature and the second feature according to the historical monitoring area without abnormal environmental data and abnormal environmental data.

[0046] The grading early warning module is used to monitor the environmental data in the monitoring area in real time, calculate the grading data of the early warning according to the first feature, and then realize grading early warning when abnormal environment occurs in the monitoring area according to the first feature.

[0047] The intelligent optimization module is used to modify the grading data in the grading early warning module according to the extracted second feature.

[0048] The feature extraction module comprises a first feature extraction unit and a second feature extraction unit.

[0049] The first feature extraction unit is used to collect abnormal environmental data and data without abnormal environment in the monitoring area, compare the two sets of collected data to obtain difference data, and after k comparisons, count the k times of difference data to extract the data type with the most repeated number of times in the k times of difference data. Let be the first feature in the monitoring area.

[0050] The second feature extraction unit is used to collect abnormal environmental data and data without abnormal environment in the monitoring area, draw data curve graphs of the two sets of collected data, calculate and analyze the two curve graphs, obtain the function and the derivative function of each curve, respectively extract the function whose derivative function value only exists in one positive and negative case, and compare the two sets of extracted functions to select the data type x corresponding to the variable value in the curve with the same function as the second feature in the monitoring area.

[0051] The hierarchical early warning module obtains hierarchical data for early warning when an abnormality occurs in the monitoring area by analyzing the speed at which the number of remaining changed environmental data types increases when the first characteristic changes abnormally in the monitoring area.

[0052] The intelligent optimization module comprises a characteristic relationship calculation unit and a real-time optimization unit.

[0053] The characteristic relationship calculation unit is configured to, in the process of calculating the second characteristic, set the inverse function of the curve function corresponding to the selected second characteristic as the linear relationship between the first characteristic and the second characteristic, and set the linear relationship function between the first characteristic and the second characteristic as f (x) = a * x + b, wherein a and b are constants. In the function, a is the value of the first characteristic, x is the value of the second characteristic, and a is the linear relationship between the first characteristic and the second characteristic.

[0054] The real-time optimization unit is configured to collect data of the second characteristic in the monitoring area by using a plurality of sensors, modify the minimum level _min of early warning and the hierarchical data of early warning according to the calculated linear relationship function between the first characteristic and the second characteristic.

[0055] Compared with the prior art, the present application has the following beneficial effects:

[0056] 1. The present application can effectively classify the speed at which the abnormality in the monitoring area affects the environment by setting the level of early warning according to the size of the derivative, and can allocate resources reasonably.

[0057] 2. The present application can ensure that the abnormality and the influence of the abnormality on the environment of the monitoring area are avoided when the monitoring area is monitored in real time by obtaining the second characteristic data that can linearly affect the first characteristic through the relationship between all data in the monitoring area and the first characteristic, and optimizing and modifying the hierarchical data of early warning through the second characteristic data. BRIEF DESCRIPTION OF DRAWINGS

[0058] The accompanying drawings are included to provide a further understanding of the present application, and constitute a part of the specification, illustrate the present application together with the embodiments thereof, and explain the present application, but do not limit the present application. In the drawings:

[0059] Figure 1 is a module distribution diagram of the mine transportation safety adaptive optimization system based on video monitoring of the present application;

[0060] Figure 2 ​​​​is a step schematic diagram of a mine transportation safety adaptive optimization method based on video monitoring of the present application;

[0061] Figure 3 is an abnormal function curve graph of a mine transportation safety adaptive optimization method based on video monitoring of the present application;

[0062] Figure 4 is an abnormal function curve graph of a mine transportation safety adaptive optimization method based on video monitoring of the present application. DETAILED DESCRIPTION

[0063] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0064] Please refer to Figures 1-4 The present application provides technical solutions:

[0065] The mine transportation safety adaptive optimization method based on video monitoring comprises the following steps:

[0066] S100, using multiple sensors to sense the environment in the monitoring area to obtain environment multi-source data in the monitoring area;

[0067] Further, the sensors use visual sensors, temperature and humidity sensors, and microphone sensors, and the environment multi-source data includes temperature and humidity data, image data, and sound data in the monitoring area.

[0068] S200, processing and analyzing the environment data sensed by the sensors and performing feature extraction to obtain the first feature of the monitoring area;

[0069] Further, the specific steps of performing feature extraction on the environment in the monitoring area to obtain the first feature of the monitoring area are:

[0070] S201, collecting the environment data without abnormality in the monitoring area using multiple sensors and classifying, according to different sensors, the data collected by different sensors into different categories to generate corresponding n kinds of data, respectively , The first, second, third,..., and n kinds of data, respectively, and n is a positive integer; the abnormality in the monitoring area indicates that the environment data in the monitoring area has multiple data eruption type changes.

[0071] S202, record the abnormal environment data in the monitoring area, classify the data, divide the data collected by different sensors into different categories, and generate m kinds of abnormal environment data respectively , are the 1st, 2nd, 3rd,..., mth abnormal environment data, and m is a positive integer;

[0072] S203, after classifying and collecting the environment data in the monitoring area when no abnormality occurs and when abnormality occurs, compare the data in the two cases, use the environment data in the monitoring area when no abnormality occurs to filter the environment data in the monitoring area when abnormality occurs , delete the values of the environment data in the monitoring area when abnormality occurs that are the same as the environment data when no abnormality occurs , and define the environment data after deletion when abnormality occurs as difference data , are the 1st, 2nd, 3rd,..., jth difference data, and j is a positive integer;

[0073] S204, according to the above steps, extract k times of environment data in the monitoring area when abnormality occurs, then compare the environment data in the monitoring area when no abnormality occurs with the environment data in the monitoring area when abnormality occurs for k times, generate k times of difference data, and count the k times of difference data to extract the data category that appears most frequently in the k times of difference data , the subscript y takes values from 1 to j, and is defined as the first feature in the monitoring area.

[0074] The first feature generated above reflects the most significant change characteristics of the environment data when abnormality occurs in the monitoring area, and the first feature can be used to monitor whether abnormality occurs in the monitoring area.

[0075] S300, analyze the environment in the monitoring area, calculate the abnormal environment index in the monitoring area, and realize the hierarchical early warning of the abnormality in the monitoring area;

[0076] Further, the hierarchical early warning steps in the monitoring area are as follows:

[0077] S301, according to the collection of the environment data in the monitoring area when abnormality occurs, use the k times of difference data generated after comparing the environment data in the monitoring area when no abnormality occurs with the environment data in the monitoring area when abnormality occurs in S200, and the number of difference data is , is the number of difference data generated by comparing the 1st, 2nd, 3rd,..., kth environment data in the monitoring area when no abnormality occurs with the environment data in the monitoring area when abnormality occurs, and k is a positive integer;

[0078] S302. Arrange the first feature in ascending order. The numerical values ​​are sorted, and the first feature is plotted using the point-plotting method. Number of numerical and differential data A line graph where p takes values ​​from 1 to k;

[0079] S303. Normalize and fit the line graph to form a smooth function curve. Using a piecewise method, calculate the functional relationship of each segment of the function curve to generate a set of functions. , Let h be the function in segments 1, 2, 3...h of the function curve, where h is a positive integer;

[0080] S304. Perform derivative operations on the piecewise functions separately to obtain the derivative function of each segment of the function curve. , Let h be the derivative of the function curve for segments 1, 2, 3...h, where h is a positive integer; the derivative represents the growth rate of each segment of the function curve.

[0081] S305. Filter the derivative functions obtained above, selecting those with derivative values ​​greater than 0 and sorting them in ascending order. Obtain the corresponding independent variables based on the sorting of the derivative functions. The sorting is , Let the independent variables be the 1st, 2nd, 3rd...ath derivatives that are greater than 0, arranged in ascending order of their derivative values. , where a is a positive integer; calculate the... Classified data for early warning purposes;

[0082] The larger the value of the derivative of a curvilinear function, the more likely it is that the independent variable... When the derivative increases, the faster the variable changes; it can reflect the changes in other environmental data in the monitoring area when an anomaly occurs in the monitoring area by judging the change in the first characteristic value. The larger the derivative, the faster the anomaly in the monitoring area causes its impact. Therefore, setting the warning level according to the size of the derivative can effectively classify the speed at which anomalies in the monitoring area affect the environment and make reasonable resource allocation.

[0083] S306. Obtain the grading data based on the calculation. The monitored area is given tiered early warnings, and the independent variables in the line graph are selected. The initial value is the lowest level of the warning. _min, then according to The warning levels increase sequentially.

[0084] The derivative is used to calculate the classification data of the pre-warning in the monitoring area. After the abnormality occurs in the monitoring area, the speed of the increase of the number of the remaining changed environmental data caused by the change of the first feature is also different, and the value of the derivative represents the increase speed of the remaining changed environmental data in the monitoring area. According to the different increase speeds, the abnormality is classified and pre-warned. For example, when the mine transportation is monitored to determine whether a fire is likely to occur, in the monitoring area, the temperature is the first feature. When the temperature is too high, it is judged that a fire abnormality occurs, and after the fire occurs, the remaining data in the monitoring area will change. The increase speed of the number of the changed data proves the size of the fire, so the classification pre-warning of the abnormality can be realized.

[0085] S400, machine learning is performed by using the calculated first feature in the monitoring area to construct an abnormal pre-warning model;

[0086] Further, the construction of the abnormal pre-warning model is specifically:

[0087] The calculated classification data The convolutional neural network algorithm is used for machine learning to construct a classification pre-warning model. The sensor is used to realize real-time sensing of the environment in the monitoring area, and the first feature value in the environmental data in the monitoring area is extracted _s, the extracted first feature value _s is input into the constructed classification pre-warning model, and it is judged that when the extracted first feature value _s> _min, the environment in the monitoring area is abnormal, and then _s is compared with the classification data in turn, when it is judged that _s> the corresponding level of pre-warning is sent to the monitoring area, and i belongs to 1-a.

[0088] S500, the environment during the pre-warning is analyzed, and the second feature of the monitoring area is extracted;

[0089] Further, the specific steps of extracting the second feature of the monitoring area are:

[0090] S501, the data of the abnormal environment in the monitoring area is collected, and the k times of the abnormal environmental data in the monitoring area are extracted according to the method described in S200. Each extracted environmental data is classified to generate m kinds of abnormal environmental data respectively The first feature in the abnormal environmental data is extracted, and k first feature values are obtained after k times of extraction , The value of the first feature in the environment data that occurs abnormally for the first, second, third, …, k time, k is a positive integer; the point drawing method is used to draw a curve graph with the value of the first feature in the abnormal environment data as the independent variable, and the values of other kinds of data as the variables, m-1 curves are drawn in the graph,

[0091] S502, analyze the drawn curve graph, and calculate the function of the m-1 curves in the curve graph using the piecewise method, and the function corresponding to each curve is , The function of the first, second, third, …, m-1 curves, m-1 is a positive integer,

[0092] S503, derivative operation is performed on the m-1 functions calculated above to obtain the derivative function , The derivative function of the first, second, third, …, m-1 curve functions, m-1 is a positive integer,

[0093] S504, according to the above method, collect and classify the environment data in the monitoring area that does not occur abnormally, and generate corresponding n kinds of data, respectively , draw a curve graph, calculate a curve function, and finally obtain the derivative function of each kind of data in the environment data that does not occur abnormally , The derivative function of the first, second, third, …, n-1 curve functions, n-1 is a positive integer,

[0094] S505, analyze the two groups of derivative functions calculated above, and extract the functions whose derivative function values only exist in one positive or negative situation, specifically: in the abnormal environment data, only extract the function curves in the entire function ≥0 or ≤0, in the environment data that does not occur abnormally, only extract the function curves in the entire function ≥0 or ≤0, ,

[0095] S506, compare the two groups of functions after extraction, and select the data type x corresponding to the variable value in the curve with the same function as the second feature in the monitoring area.

[0096] ​​​By screening the curves in the graph, when the derivative function of the curve only has one change, it is proved that the data may have a certain linear relationship with the first feature, and the curves with the same change are found in the screened curves by comparing the curves in the monitoring area with and without abnormality, and the data type that can linearly affect the first feature in the monitoring area is located, and the second feature in the monitoring area is located.

[0097] S600, calculate the linear feature of the first feature and the second feature, and realize real-time optimization of the hierarchical early warning according to the calculated linear feature of the first feature and the second feature.

[0098] Further, the linear feature of the first feature and the second feature is specifically:

[0099] According to the above S500, in the process of calculating the second feature, the inverse function of the curve function corresponding to the second feature is defined as the linear relationship between the first feature and the second feature, and the linear relationship function between the first feature and the second feature is set as , in which is the value of the first feature, x is the value of the second feature, is the linear relationship between the first feature and the second feature.

[0100] The specific process of realizing real-time optimization of early warning classification is:

[0101] The data of the second feature in the monitoring area is collected by using a plurality of sensors to obtain the real-time value x_s of the second feature in the monitoring area, and according to the linear relationship function , the value x_s of the second feature is substituted as the independent variable into the function to calculate the theoretical minimum early warning value of the first feature _b, the calculated theoretical minimum early warning value of the first feature is compared with the actual minimum early warning value _min, the difference _c= _b _min, the minimum early warning value _min is replaced by the theoretical minimum early warning value _b, and the hierarchical data of the early warning are added to the difference _c to obtain new hierarchical data of the early warning.

[0102] When issuing an early warning within a monitored area, other data besides the primary characteristic of the judgment standard may affect the early warning standard, making it impossible to issue timely warnings for abnormal situations within the monitored area. For example, in mining transportation, when issuing an early warning for a fire, temperature is set as the primary characteristic and used as the early warning standard. However, the humidity in the air will still affect the temperature standard for the occurrence of a fire.

[0103] The mine transportation safety adaptive optimization system based on video surveillance includes an environmental perception module, a feature extraction module, a hierarchical early warning module, and an intelligent optimization module.

[0104] The environmental perception module uses multiple sensors to collect environmental data within the monitored area;

[0105] The feature extraction module is used to extract a first feature and a second feature based on historical environmental data of no abnormalities and environmental data of abnormalities within the monitoring area.

[0106] The graded early warning module is used to monitor environmental data in the monitoring area in real time, calculate the graded data of the early warning based on the first feature, and then implement graded early warning when abnormal environment occurs in the monitoring area based on the first feature.

[0107] The intelligent optimization module is used to modify the hierarchical data in the hierarchical early warning module based on the extracted second feature.

[0108] The feature extraction module includes a first feature extraction unit and a second feature extraction unit;

[0109] The first feature extraction unit is used to collect data on abnormal environments and non-abnormal environments within the monitored area, compare the two sets of data to obtain difference data, and after performing k comparisons, statistically analyze the k differences to extract the data type that appears most frequently among the k differences. ,Will Set it as the first feature within the monitored area;

[0110] The second feature extraction unit is used to collect data on abnormal environments and non-abnormal environments within the monitoring area. It plots data curves on the two sets of data, performs calculations and analysis on the two curves to obtain the function and derivative function of each curve, extracts the function whose derivative function has only one positive and one negative value, compares the two sets of extracted functions, and selects the data type x corresponding to the variable value in the curve with the same function as the second feature within the monitoring area.

[0111] The hierarchical early warning module obtains hierarchical data for early warning when an abnormality occurs in the monitored area by analyzing the speed at which the number of types of environmental data that change increases when the first characteristic changes when an abnormality occurs in the monitored area.

[0112] The intelligent optimization module includes a characteristic relationship calculation unit and a real-time optimization unit.

[0113] The characteristic relationship calculation unit is configured to, in calculating the second characteristic, set the inverse function of the curve function corresponding to the selected second characteristic as the linear relationship between the first characteristic and the second characteristic, and set the linear relationship function between the first characteristic and the second characteristic as , where is the value of the first characteristic, x is the value of the second characteristic, is the linear relationship between the first characteristic and the second characteristic.

[0114] The real-time optimization unit is configured to collect data of the second characteristic in the monitored area using multiple sensors, modify the calculated minimum level _min of early warning and the hierarchical data for early warning according to the calculated linear relationship function between the first characteristic and the second characteristic.

[0115] Embodiment 1

[0116] A certain mine transportation working environment is monitored, and the first characteristic data and the early warning hierarchical data in the monitored area have been calculated, where the first characteristic data is , and the early warning hierarchical data is

[0117] The second characteristic in the monitored area is extracted, and the environmental data in the monitored area when an abnormality occurs and when no abnormality occurs are shown in Table 1 and Table 2 as follows:

[0118]

[0119] Table 1

[0120]

[0121] Table 2

[0122] According to the contents in the above tables, a curve graph is drawn, as shown in Figure 3 and Figure 4 , where series 1 represents the curve of and , and series 2 represents the curve of and .

[0123] Through analysis and calculation of the curves in the two graphs, it is obtained that the function in the abnormality curve graph is ​, the function in the abnormal curve graph is 、 , the derivative of the four functions is obtained respectively , ;

[0124] The derivative of the curve function in the two graphs is calculated respectively, and the derivative of the curve function in the abnormal curve graph is There is only one value in the abnormal curve graph, and there is only one value in the abnormal curve graph ; After comparing and screening the two functions, and are both less than 0, so it is determined that the second feature in the monitoring area is the data and represented by the data type.

[0125] It should be noted that in this paper, relationship terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment.

[0126] Finally, it should be noted that: the above only describes the preferred embodiments of the present application, and does not limit the present application, although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions recorded in the foregoing embodiments, or replace some technical features with equivalent ones. Any modification, equivalent replacement, improvement, etc. within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for adaptive optimization of mine haulage safety based on video surveillance, characterized in that: The method comprises the following steps: S100, sensing the environment in the monitoring area by using multiple sensors to obtain multi-source environment data in the monitoring area; S200, processing and analyzing the environment data sensed by the sensors and performing feature extraction to obtain the first feature of the monitoring area; S300, analyzing the environment data in the monitoring area, calculating the speed of change of the abnormal environment data in the monitoring area, and realizing the hierarchical early warning of the abnormal environment in the monitoring area; The hierarchical early warning step in the monitoring area is: S301, according to the collection of the abnormal environment data occurring in the monitoring area, using the k times comparison of the non-abnormal environment data and the abnormal environment data in the monitoring area in S200, generating the number of k times difference data respectively , is the number of difference data generated by the first, second, third...k times comparison of the non-abnormal environment data and the abnormal environment data in the monitoring area, k is a positive integer; S302、According to the first feature The numerical values are sorted, and the first feature The numerical values and the number of difference data The line graph, and p takes values from 1 to k; S303, normalize and fit the broken line graph to form a smooth function curve, calculate the function relationship of each segment of the function curve by using the piecewise method, and generate a function set as , are the functions of the 1st, 2nd, 3rd,..., hth segments of the function curve, and h is a positive integer. S304. Perform derivative operations on the piecewise functions separately to obtain the derivative function of each segment of the function curve. , Let h be the derivative of the function curve for segments 1, 2, 3...h, where h is a positive integer; the derivative represents the growth rate of each segment of the function curve. S305, screening the calculated derivative function, selecting the derivative function whose value is greater than 0, and sorting the derivative function in ascending order, and obtaining the sorting of independent variables corresponding to the derivative function , is the first, second, third,..., a-th independent variable corresponding to the derivative function greater than 0 in ascending order of the value of the derivative function , a is a positive integer; the calculated is defined as the grading data of early warning;​ S306. Obtain the grading data based on the calculation. The monitored area is given tiered early warnings, and the independent variables in the line graph are selected. The initial value is the lowest level of the warning. _min, then according to The warning levels increase sequentially in order; S400, machine learning is performed by using the first feature calculated in the monitoring area to construct an abnormal early warning model; S500, analyzing the environment during early warning and extracting the second feature of the monitoring area; The specific steps of extracting the second feature of the monitoring area are: S501, collect the abnormal environment data occurring in the monitoring area, extract k times of the abnormal environment data occurring in the monitoring area according to the method described in S200, classify each extracted environment data, and generate m kinds of abnormal environment data respectively as , extract the first feature in the abnormal environment data, and obtain k first feature values after k times of extraction , is the value of the first feature in the first, second, third...k times of the abnormal environment data, and k is a positive integer; the point drawing method is used to draw a curve graph with the value of the first feature in the abnormal environment data as the independent variable and the values of other kinds of data as the variables , , ; m-1 curves are drawn in the graph S502, analysis is performed on the drawn graph, and functions are calculated for the m-1 curves in the graph by using the piecewise method, and the function corresponding to each curve is , is the function of the 1st, 2nd, 3rd, …, and (m-1)th curves, and m-1 is a positive integer. S503, derivative operation is performed on the m-1 functions calculated above to obtain derivative functions , are derivative functions of the first, second, third, …, and m-1 curve functions, and m-1 is a positive integer. S504, according to the method described above, the environment data in the monitoring area without exception is collected and classified, and the corresponding n kinds of data are generated , draw a curve, calculate the curve function, and finally get the derivative function of each data in the environment data without exception , is the derivative function of the first, second, third... n-1 curve function, and n-1 is a positive integer; S505, analyze the two groups of derived functions calculated above, and extract the functions whose values of the derived functions exist in only one positive or negative case, specifically: in the abnormal environmental data, only extract the function curves of ≥0 or ≤0 in the entire function, and in the environmental non-abnormal data, only extract the function curves of ≥0 or ≤0 in the entire function, , ≤0, ≤0, ≤0, , ; S506, comparing the two groups of functions after extraction, and selecting the data type x corresponding to the variable value in the curve with the same function as the second feature in the monitoring area; S600, calculating the linear feature of the first feature and the second feature, and realizing real-time optimization of the hierarchical early warning according to the linear feature of the first feature and the second feature calculated; The linear feature of the first feature and the second feature calculated is: According to the above S500, in the process of calculating the second feature, the inverse function of the curve function corresponding to the selected second feature is defined as the linear relationship between the first feature and the second feature, and the linear relationship function between the first feature and the second feature is set as , wherein is the value of the first feature, x is the value of the second feature, is the linear relationship between the first feature and the second feature. The specific process of realizing real-time optimization of the early warning hierarchy is: Data on the second feature within the monitored area is collected using multiple sensors to obtain the real-time value x_s of the second feature within the monitored area. This value is then analyzed based on a linear relationship function. The theoretical minimum warning value of the first feature is calculated by substituting the value of the second feature, x_s, as the independent variable into the function. _b, the calculated theoretical minimum warning value of the first feature and the actual minimum warning value. Compare with _min and calculate the difference. _c= _b _min represents the minimum warning value. _min utilizes the theoretical minimum warning value Replace _b and change the warning level data. Mean and difference The _c values ​​are added together to obtain the new warning classification data.

2. The video surveillance based adaptive optimization method for mine haulage safety according to claim 1, characterized in that: The specific steps of extracting the first feature of the monitoring area in S200 are: S201, collecting and classifying the environment data in the monitoring area where no abnormality occurs by using multiple sensors, and dividing the data collected by different sensors into different categories to generate corresponding n kinds of data respectively , respectively as the 1st, 2nd, 3rd...nth data, and n is a positive integer; S202, record the abnormal environment data occurring in the monitoring area, classify the data, according to the different sensors, divide the data collected by different sensors into different categories, generate m kinds of abnormal environment data respectively , are the 1st, 2nd, 3rd...mth abnormal environment data, and m is a positive integer. S203, after collecting the environment data in the monitoring area when no abnormality occurs and when abnormality occurs, comparing the data in the two cases, using the environment data in the monitoring area when no abnormality occurs to the environment data in the monitoring area when abnormality occurs to filter the environment data in the monitoring area when abnormality occurs with the environment data when no abnormality occurs with the same value, and deleting the environment data after the deletion , is the first, second, third,..., jth difference data, and j is a positive integer. S204、According to the above steps, extract the environmental data that occurs abnormally in the monitoring area k times, then generate k times of difference data after comparing with the environmental data that does not occur abnormally in the monitoring area k times, and extract the data type with the most repeated number of times in k times of difference data , the subscript y takes the value of 1 to j, and The first feature in the monitoring area is set.

3. The video surveillance based adaptive optimization method for mine haulage safety according to claim 1, characterized in that: The specific steps of constructing an abnormal early warning model in S400 are: The calculated hierarchical data Machine learning is performed by using a convolutional neural network algorithm to construct a hierarchical early warning model; the environment in the monitoring area is sensed in real time by using a sensor to extract first characteristic values in the environmental data in the monitoring area _s, the extracted first characteristic values _s are input into the constructed hierarchical early warning model, and it is judged whether the extracted first characteristic values _s> _min, the environment in the monitoring area is abnormal, and then _s is compared with the hierarchical data in sequence, and when it is judged that _s> , a corresponding level of early warning is issued for the monitoring area, i belongs to 1-a.

4. A video monitoring based mine transportation safety adaptive optimization system applying the video monitoring based mine transportation safety adaptive optimization method according to any one of claims 1-3, characterized in that: The mine transportation safety adaptive optimization system comprises an environment sensing module, a feature extraction module, a hierarchical early warning module and an intelligent optimization module; The environment sensing module collects environment data in the monitoring area by using multiple sensors; The feature extraction module is used to extract the first feature and the second feature according to the historical monitoring area without abnormal environment data and with abnormal environment data; The hierarchical early warning module is used to monitor the environment data in the monitoring area in real time, calculate the hierarchical data of early warning according to the first feature, and then realize hierarchical early warning when the abnormal environment occurs in the monitoring area according to the first feature; The intelligent optimization module is used to modify the hierarchical data in the hierarchical early warning module according to the extracted second feature.

5. The video surveillance based adaptive optimization system for mine haulage safety according to claim 4, characterized in that: The feature extraction module comprises a first feature extraction unit and a second feature extraction unit; The first feature extraction unit is used for collecting abnormal environment data and data without abnormal environment occurring in the monitoring area, comparing the two groups of collected data to obtain difference data, and after k comparisons, counting the k times of difference data and extracting the data type with the most repeated times in the k times of difference data , The first feature is set as the monitoring area The second feature extraction unit is used to collect the data of the abnormal environment data and the data without abnormal environment in the monitoring area, draw data curve graphs of the two groups of collected data, calculate and analyze the two kinds of curve graphs, obtain the function and the derivative function of each curve, extract the function whose value of the derivative function only exists in one positive or negative situation, compare the two groups of functions after extraction, and select the data type x corresponding to the variable value in the curve with the same function as the second feature in the monitoring area.

6. The video surveillance based mine haulage safety adaptive optimization system of claim 4, wherein: The hierarchical early warning module analyzes the speed of change of the number of environment data types caused by the change of the first feature when the abnormal environment occurs in the monitoring area, and obtains the hierarchical data of early warning when the abnormal environment occurs in the monitoring area.

7. The video surveillance based mine haulage safety adaptive optimization system of claim 4, wherein: The intelligent optimization module comprises a feature relationship calculation unit and a real-time optimization unit; The feature relationship calculation unit is configured to, in the process of calculating the second feature, set an inverse function of the curve function corresponding to the selected second feature as a linear relationship between the first feature and the second feature, and set a linear relationship function between the first feature and the second feature as , wherein f is a value of the first feature, x is a value of the second feature, and f is the linear relationship between the first feature and the second feature. The real-time optimization unit is configured to collect data of the second feature in the monitoring area by using a plurality of sensors, modify the calculated linear relationship function of the first feature and the second feature as the minimum level of the calculated early warning _min and the classification data of the early warning.

Images