Online Libs laser real-time detection device

Abstract

The invention relates to the technical field of new material related services, in particular to an online Libs laser real-time detection device, which comprises a laser generator, a data acquisition instrument, a power supply, an online monitoring system, a sensor node group and a shell, the online monitoring system comprises a data acquisition module, a data processing module, a first detection tuning module and a second detection tuning module, the computing power is dynamically adjusted according to the change of an actual detection task, and accurate computing power calculation is guaranteed, so that the accuracy of a detection result when a new material is detected is improved, and the detection efficiency is improved. According to the on-line monitoring system, flexible electric quantity distribution is carried out through the second detection tuning module, and electric quantity consumption is reduced, so that the remaining electric quantity is more durable, resource waste is avoided, and the resource utilization rate during new material detection is improved, and therefore the cruising ability of the on-line Libs laser real-time detection device is improved, and the efficiency and accuracy of new material detection are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of new material-related services, and in particular to an online Libs laser real-time detection device. Background Art

[0002] Due to limitations such as the size, weight, and energy supply of drones, the performance of the computing equipment they carry cannot be compared with large-scale ground-based computing centers. However, the data processing process of Libs technology is relatively complex. From the acquisition of plasma spectral signals induced by lasers to feature extraction and elemental analysis of spectral data, a large amount of computing resources is required. In addition, existing Libs laser real-time detection devices usually adopt a fixed computing power allocation mode, which cannot dynamically adjust computing power according to changes in actual detection tasks, nor can they flexibly allocate power, resulting in waste of resources. A device that can optimize computing power and allocate power is indeed needed.

[0003] Chinese patent publication number CN105223187A discloses a device for measuring heavy metal elements in gas based on Libs. However, this solution still suffers from the inability to dynamically adjust computing power according to changes in actual detection tasks, resulting in inaccurate detection results when detecting new materials. It also cannot flexibly allocate power, resulting in waste of resources when detecting new materials. A device that can dynamically adjust computing power and flexibly allocate power is needed to solve these problems. Summary of the Invention

[0004] To this end, the present invention provides an online Libs laser real-time detection device to overcome the problems in the prior art of being unable to dynamically adjust computing power according to changes in actual detection tasks, resulting in inaccurate detection results when detecting new materials, and being unable to flexibly allocate power, resulting in waste of resources when detecting new materials.

[0005] To achieve the above object, the present invention provides an online Libs laser real-time detection device, comprising:

[0006] A laser generator, which is connected to the data acquisition instrument, the online monitoring system, the sensor node group and the housing, and is used to emit a laser beam;

[0007] A data collector, connected to the laser generator and the housing, for collecting target environment data;

[0008] A power supply connected to the online monitoring system for providing power to the online Libs laser real-time detection device;

[0009] An online monitoring system connected to the laser generator and the power supply, and connected to the data acquisition instrument and the sensor node group via internal connecting lines, for controlling the online Libs laser real-time detection device;

[0010] A sensor node group is connected to the laser generator and the housing, wherein the sensor node group includes a voltage sensor, a current sensor, and a temperature sensor for collecting data from the target device;

[0011] The housing is connected to the laser generator, the data acquisition instrument and the sensor node group, and is used to protect the online Libs laser real-time detection device.

[0012] Furthermore, the online monitoring system includes:

[0013] A data acquisition module, used to acquire target environment data and target device data;

[0014] A data processing module is used to process the target environment data according to the data processing method to obtain the laser detection results;

[0015] A first detection and tuning module is used to calculate the computing power requirement based on the target environment data using a computing power requirement calculation method, adjust the data processing method based on the computing power requirement, and optimize the computing power requirement based on the computing power requirement optimization method;

[0016] The second detection and tuning module is used to calculate the total power demand value according to the target device data through the total power demand value calculation method, and is also used to perform secondary computing power adjustment on the computing power demand optimization method according to the total power demand value through the computing power optimization adjustment method.

[0017] Furthermore, the data processing module performs data preprocessing on the target environment data according to a data preprocessing strategy, and the data preprocessing strategy includes:

[0018] Step J01: Based on the spectrum data value set s={s1,s2,s3,...,si}, the spectrum data minimum value S min and the maximum value of the spectral data S max Calculate the normalized spectral data value set Sop = {Sop1, Sop2, Sop3, ..., Sopi} to obtain the normalized spectral data value set Sop = {Sop1, Sop2, Sop3, ..., Sopi}, set

[0019] Step J02: Calculate the processed time-resolved data set t`={t`1,t`2,t`3,...,t`j} according to the segmentation window size h and the time-resolved data set t={t1,t2,t3,...,tj} to obtain the processed time-resolved data set t`={t`1,t`2,t`3,...,t`j}. Set

[0020] Step J03: Obtain projection coordinates (αm, βm, γm) of each spatial position data according to the universal transverse Mercator grid system, where m is the total number of measurement points;

[0021] In step J04, the normalized spectral data value set Sop = {Sop1, Sop2, Sop3, ..., Sopi}, the processed time-resolved data set t` = {t`1, t`2, t`3, ..., t`j} and the projection coordinates (αm, βm, γm) are used as the first target processing data.

[0022] Furthermore, when the data processing module performs data preprocessing on the target environment data according to the data preprocessing strategy, the data processing module constructs a spectral data feature vector model according to a spectral data feature vector model construction method, and the spectral data feature vector model construction method includes:

[0023] Step K01, divide the historical spectral dataset into a 70% spectral training set, a 20% spectral validation set, and a 10% spectral test set;

[0024] Step K02, initializing the spectral parameters of the first convolutional neural network model;

[0025] Step K03: Input the spectral training set into the first convolutional neural network model after parameter initialization for training, input the spectral validation set into the trained first convolutional neural network model, optimize the spectral parameters of the trained first convolutional neural network model, and then input the spectral test set into the first convolutional neural network model after parameter optimization for testing, and output the test accuracy;

[0026] Step K04: outputting the first convolutional neural network model after parameter optimization with an accuracy rate of 90% as a spectral data feature vector model.

[0027] Furthermore, the data processing module performs feature extraction on the first target processing data according to a feature extraction strategy, wherein the feature extraction strategy includes:

[0028] Step D01, inputting a normalized spectral data value set Sop={Sop1, Sop2, Sop3, ..., Sopi} into a spectral data feature vector model to obtain a spectral feature vector set Foh={Foh1, Foh2, Foh3, ..., Fohi} output by the spectral data feature vector model;

[0029] Step D02, inputting the processed time-resolved data set t`={t`1, t`2, t`3, ..., t`j} into the time-resolved feature vector model to obtain the time feature vector set Fom={Fom1, Fom2, Fom3, ..., Fomj} output by the time-resolved feature vector model;

[0030] Step D03, calculate each spatial distance dm according to each projection coordinate (αm, βm, γm) and each preset coordinate (αy, βy, γy), and set

[0031] In step D04 , the spectral feature vector set Foh={Foh1, Foh2, Foh3, ..., Fohi}, the time feature vector set Fom={Fom1, Fom2, Fom3, ..., Fomj} and each spatial distance dm are used as second target processing data.

[0032] Furthermore, the data processing module performs data fusion on the second target processing data according to a data fusion strategy, and the data fusion strategy includes:

[0033] Step E01, matrixing the spectral feature vector set Foh = {Foh1, Foh2, Foh3, ..., Fohi}, the time feature vector set Fom = {Fom1, Fom2, Fom3, ..., Fomj} and each spatial distance dm to obtain a first matrix X;

[0034] Step E02, setting a first data sampling point Xzv, where z is the zth row in the first matrix X, and v is the vth column in the first matrix X;

[0035] Step E03, calculate the mean μv and the standard deviation σv of each column according to the total number of measurement points m and the first data sampling point Xzv, and obtain the mean μv and the standard deviation σv of each column, and set

[0036] Step E04, calculate the second data sampling point Gz0v0 according to the mean value μv of each column and the standard deviation σv of each column, and obtain the second data sampling point Gz0v0, and set

[0037] Step E05, performing matrix processing on the second data sampling point Gz0v0 to obtain a second matrix G;

[0038] Step E06, transposing the second matrix G to obtain a transposed second matrix G';

[0039] Step E07, calculate the covariance matrix C according to the total number of measurement points m, the second matrix G and the transposed second matrix G' to obtain the covariance matrix C, set

[0040] Step E08, performing eigendecomposition on the covariance matrix C to obtain eigenvalues;

[0041] Step E09: Arrange the eigenvalues in a preset order to obtain sorted eigenvalues, perform value processing on the sorted eigenvalues according to the value ratio, and obtain the laser detection result.

[0042] Furthermore, the first detection and tuning module calculates the computing power requirement according to the target environment data using a computing power requirement calculation method, and the computing power requirement calculation method includes:

[0043] Step B01, calculate the spectral data computing power requirement Q1 according to the spectral data volume Iv, obtain the spectral data computing power requirement Q1, and set Q1=4i';

[0044] Step B02: Calculate the time-resolved computing power requirement Q2 based on the time-resolved data point jq and the segmentation window size h to obtain the time-resolved computing power requirement Q2, and set Q2 = 3 × jq - 2 × h + 1;

[0045] Step B03: Calculate the spatial location computing power requirement Q3 based on the total number of measurement points m to obtain the spatial location computing power requirement Q3, and set Q3 = 38 × n3;

[0046] Step B04: Calculate the total computing power requirement Q based on the spectral data computing power requirement Q1, the time-resolved computing power requirement Q2, and the spatial position computing power requirement Q3 to obtain the total computing power requirement Q, and set Q = Q1 + Q2 + Q3;

[0047] When the first detection and tuning module adjusts the computing power of the data processing method according to the computing power demand, the total computing power demand Q is compared with the preset computing power demand Q0, the total computing power demand is judged according to the comparison result, and the computing power of the data processing method is adjusted according to the judgment result, wherein:

[0048] When Q≤Q0, the first detection and tuning module determines that the total computing power demand is low and does not adjust the computing power of the data processing method;

[0049] When Q>Q0, the first detection tuning module determines that the total computing power demand is high, adjusts the computing power of the data processing method, sends the second target processing data to the cloud, performs data fusion through the cloud, and obtains the laser detection results.

[0050] Furthermore, the first detection and tuning module optimizes the computing power requirement according to a computing power requirement optimization method, and the computing power requirement optimization method includes:

[0051] Step L01, calculating the data generation rate Vs according to the data acquisition frequency Cf and the data acquisition amount Ns, obtaining the data generation rate Vs, and setting Vs = Cf × Ns;

[0052] Step L02: Compare the data generation rate Vs with the preset generation rate Vs0, determine the state of the data generation rate Vs based on the comparison result, and update the total computing power requirement Q based on the determination result, where:

[0053] When Vs≤Vs0, the first detection and tuning module determines that the data generation rate Vs is a low rate and does not update the total computing power demand Q;

[0054] When Vs>Vs0, the first detection and tuning module determines that the data generation rate Vs is at a high rate, updates the total computing power demand Q, and sets the computing power update coefficient to ε, setting ε=1.3-0.3e -0.7×(Vs-Vs0) The updated total computing power requirement is Q`, Q`=ε×Q. Compare the updated total computing power requirement Q` with the preset computing power requirement Q0, and re-evaluate the total computing power requirement.

[0055] Step L03: Compare the parallel task quantity Bq with the preset parallel task quantity Bq0, judge the situation of the parallel task quantity Bq based on the comparison result, and make a computing power correction to the preset generation rate Vs0 based on the judgment result, where:

[0056] When Bq≤Bq0, the first detection and tuning module determines that the parallel task amount Bq is a low task amount and does not perform computing power correction on the preset generation rate Vs0;

[0057] When Bq>Bq0, the first detection and tuning module determines that the parallel task volume Bq is a high task volume, performs a computing power correction on the preset generation rate Vs0, sets the rate correction coefficient to θ, and sets The corrected preset generation rate is Vs0`, Vs0`=Vs0×θ, and the preset generation rate Vs0 is replaced by the corrected preset generation rate Vs0`, and the data generation rate Vs is re-compared with the corrected preset generation rate Vs0`.

[0058] Furthermore, the second detection and tuning module calculates the total power requirement value according to the target device data using a total power requirement value calculation method, and the total power requirement value calculation method includes:

[0059] Step P01, calculate the power R1 required for data acquisition based on the working voltage Ug, the working current Ig, and the data acquisition frequency Cf, and obtain the power R1 required for data acquisition, and set R1 = Ug × Ig × Cf;

[0060] Step P02, calculating the power R2 required for data processing based on the data processing power Pc and the data processing time tl, obtaining the power R2 required for data processing, and setting R2 = Pc × tl;

[0061] Step P03: Calculate the power R3 required for computing power analysis based on computing power calculation power Px and computing power calculation time Tx to obtain the power R3 required for computing power analysis, and set R3 = Px × Tx;

[0062] Step P04, calculating the power R4 required for power distribution based on the control circuit power Pk, the number of communications nt, and the communication time tt, to obtain the power R4 required for power distribution, and setting R4 = Pk × nt × tt;

[0063] Step P05, calculate the total power demand R based on the power required for data acquisition R1, the power required for data processing R2, the power required for computing power analysis R3 and the power required for power distribution R4, and obtain the total power demand R, and set R=R1+R2+R3+R4.

[0064] Furthermore, the second detection and tuning module performs a secondary computing power adjustment on the computing power demand optimization method according to the computing power optimization adjustment method, and the computing power optimization adjustment method includes:

[0065] Step U01: compare the total power demand R with the remaining power R0, determine the demand for the total power demand R based on the comparison result, and perform secondary computing power adjustments on the data processing method and computing power demand optimization method based on the determination result, where:

[0066] When R≤R0, the second detection and tuning module determines that the total power demand R is low, and does not perform secondary computing power adjustment on the data processing method and the computing power demand optimization method;

[0067] When R>R0, the second detection and tuning module determines that the demand situation of the total power demand R is high, performs a secondary computing power adjustment on the computing power demand optimization method, sets the adjusted data acquisition frequency to Cf`, Cf`=Cf×(R-R0) / R0, and replaces the data acquisition frequency Cf with the adjusted data acquisition frequency Cf`, recalculates the data generation rate Vs, sets the adjusted preset parallel task volume to Bq0`, Bq0`=Bq0×(1-(R-R0) / R0), replaces the preset parallel task volume Bq0 with the adjusted preset parallel task volume Bq0`, and recomputes the parallel task volume Bq with the adjusted preset parallel task volume Bq0`;

[0068] Step U02: compare the working temperature Wh with the preset working temperature Wh0, determine the state of the working temperature Wh based on the comparison result, and perform a secondary calculation power correction on the remaining power R0 based on the determination result, wherein:

[0069] When Wh≤Wh0, the second detection and tuning module determines that the operating temperature Wh is in a low temperature state and does not perform a secondary computing power correction on the remaining power R0;

[0070] When Wh>Wh0, the second detection and tuning module determines that the operating temperature Wh is in a high temperature state, performs a secondary computing power correction on the power reserve R0, sets the corrected power reserve to R0', R0'=R0×Wh / Wh0, replaces the power reserve R0 with the corrected power reserve R0', and re-compares the total power demand R with the corrected power reserve R0';

[0071] In step U03, the heat dissipation efficiency Vv is compared with the preset heat dissipation efficiency Vv0, and the heat dissipation degree of the heat dissipation efficiency Vv is judged according to the comparison result. The preset working temperature Wh0 is updated with a secondary computing power according to the judgment result, wherein:

[0072] When Vv≤Vv0, the second detection and tuning module determines that the heat dissipation degree of the heat dissipation efficiency Vv is low, and does not perform a secondary computing power update on the preset operating temperature Wh0;

[0073] When Vv>Vv0, the second detection and tuning module determines that the heat dissipation degree of the heat dissipation efficiency Vv is high, and does not perform a secondary computing power update on the preset working temperature Wh0. The updated preset working temperature is set to Wh0`, Wh0`=Wh0×Vv / Vv0, and the preset working temperature Wh0 is replaced with the updated preset working temperature Wh0`, and the working temperature Wh is re-compared with the updated preset working temperature Wh0`.

[0074] Compared with the prior art, the beneficial effect of the present invention is that the online monitoring system acquires, processes, calculates and optimizes computing power, and calculates and distributes power to the target environment data and target device data through the data acquisition module, the data processing module, the first detection tuning module, and the second detection tuning module, so as to standardize the data and maintain detection stability, thereby improving the detection efficiency and detection accuracy of new materials, saving computing power and improving resource utilization, and increasing the endurance of the online Libs laser real-time detection device. Among them, the online monitoring system saves the computing power of the online Libs laser real-time detection device through the first detection tuning module, dynamically adjusts the computing power according to the changes in the actual detection tasks, and ensures accurate computing power calculation, thereby improving the accuracy of the detection results when detecting new materials. The online monitoring system flexibly distributes power through the second detection tuning module to reduce power consumption, so that the power reserve is more durable, avoids resource waste, and improves resource utilization when detecting new materials, thereby increasing the endurance of the online Libs laser real-time detection device and improving the efficiency and accuracy of new material detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] Figure 1 Schematic diagram of the structure of the online Libs laser real-time detection device used in this embodiment;

[0076] Figure 2 Schematic diagram of the structure of the online monitoring system of this embodiment. DETAILED DESCRIPTION

[0077] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below with reference to embodiments. It should be understood that the specific embodiments described herein are merely used to explain the present invention and are not intended to limit the present invention.

[0078] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0079] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside", and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.

[0080] Furthermore, it should be noted that, in the description of the present invention, unless otherwise expressly specified or limited, the terms "mounted," "connected," and "connected" should be understood in a broad sense. For example, they may refer to fixed connections, detachable connections, or integral connections; mechanical connections or electrical connections; direct connections or indirect connections through an intermediate medium; and internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0081] See also Figure 1 As shown, it is a structural diagram of the online Libs laser real-time detection device of this embodiment, and the device includes:

[0082] A laser generator 1, which is connected to a data acquisition device 2, an online monitoring system 4, a sensor node group 5 and a housing 6, and is used to emit a laser beam;

[0083] A data collector 2, connected to the laser generator 1 and the housing 6, for collecting target environment data;

[0084] Power supply 3, which is connected to the online monitoring system 4 and is used to provide power to the online Libs laser real-time detection device;

[0085] An online monitoring system 4, which is connected to the laser generator 1 and the power supply 3, and is connected to the data acquisition instrument 2 and the sensor node group 5 via internal connecting lines, and is used to control the online Libs laser real-time detection device;

[0086] A sensor node group 5 is connected to the laser generator 1 and the housing 6. The sensor node group includes a voltage sensor, a current sensor, and a temperature sensor for collecting data from the target device.

[0087] The housing 6 is connected to the laser generator 1, the data acquisition device 2 and the sensor node group 5, and is used to protect the online Libs laser real-time detection device.

[0088] Specifically, the online Libs laser real-time detection device is applied to the UAV Libs laser detection device. The online Libs laser real-time detection device collects, controls and supplies power to the target environment data and the target device data through the coordinated cooperation of various components, so as to realize the accurate detection of the sample, ensure the stable operation of the device, and extend the service life, thereby improving the overall performance and reliability of the online Libs laser real-time detection device. The online Libs laser real-time detection device emits a laser beam through a laser generator to ensure that the detection process is carried out effectively. The online Libs laser real-time detection device collects the target environment data through a data acquisition instrument to ensure that the acquired target environment data truly reflects the information of the sample. The online Libs laser real-time detection device provides electricity through a power supply to provide stable energy support for the detection process. The online Libs laser real-time detection device monitors the online L through an online monitoring system. The ibs laser real-time detection device is controlled and the computing power is dynamically adjusted according to the changes in the actual detection tasks to ensure accurate computing power calculation, thereby improving the accuracy of the detection results when detecting new materials. The online monitoring system also flexibly distributes power through the second detection tuning module to reduce power consumption, so that the power reserve is more durable, avoid waste of resources, improve resource utilization when detecting new materials, ensure efficient and stable completion of detection tasks, improve the efficiency and accuracy of new material detection, and thus improve the detection accuracy when detecting new materials. The online Libs laser real-time detection device collects target device data through a sensor node group, monitors the working status of the device itself in real time to ensure the safe and reliable operation of the device, and protects the online Libs laser real-time detection device through the outer shell to resist interference and damage to internal components caused by external environmental factors, extend the service life of the device, and thus ensure stable operation of the device.

[0089] Specifically, the data collector 2 collects target environment data, which includes spectral data, time-resolved data, and spatial position data.

[0090] Specifically, the spectral data refers to the data obtained by analyzing the specific light radiation generated after the laser beam emitted by the laser generator is irradiated on the surface of the object. The time-resolved data refers to the spectral data associated with time information, which records the characteristic information of the specific light radiation at different times. The spatial position data refers to the data set that records the specific position of each measurement point on the sample when the online Libs laser real-time detection device detects different positions of the sample. The specific position includes the longitude coordinates, latitude coordinates and altitude of the measurement point. The measurement point refers to the sample position detected by the online Libs laser real-time detection device.

[0091] Specifically, the data acquisition instrument 2 determines the type and content of elements contained in the sample based on the specific light radiation generated after the laser beam emitted by the laser generator is irradiated on the surface of the object, thereby accurately collecting spectral data. The data acquisition instrument 2 records the characteristic information of specific light radiation at different times through time-resolved data, thereby improving the accuracy of the data in the time dimension. The data acquisition instrument 2 obtains the distribution of elements in the sample through spatial position data, thereby discovering impurities, defects or composition differences in different areas of the sample, thereby improving detection accuracy.

[0092] Specifically, the sensor node group 5 collects target device data, and the target device data includes working current, working voltage and working temperature. The sensor node group 5 obtains the working current through the current sensor, the sensor node group 5 obtains the working voltage through the voltage sensor, and the sensor node group 5 obtains the working temperature through the temperature sensor.

[0093] Specifically, the working current refers to the current passing through the online Libs laser real-time detection device during normal operation, the working voltage refers to the power supply voltage required by the online Libs laser real-time detection device during normal operation, the working temperature refers to the real-time temperature of the online Libs laser real-time detection device during normal operation, the current sensor refers to a device for detecting and measuring the current in the circuit, the voltage sensor refers to a device for measuring the voltage in the circuit, and the temperature sensor refers to a device for measuring the temperature of an object.

[0094] Specifically, the sensor node group 5 collects operating current, operating voltage and operating temperature through the sensor node group to monitor the operating status of the device in real time, provide data reference for subsequent power distribution, and improve the reliability and service life of the device.

[0095] See also Figure 2 As shown in FIG, which is a schematic diagram of the structure of the online monitoring system of this embodiment, the online monitoring system includes:

[0096] A data acquisition module, used to acquire target environment data and target device data;

[0097] A data processing module is used to process the target environment data according to the data processing method to obtain the laser detection results. The data processing module is connected to the data acquisition module;

[0098] A first detection and tuning module is configured to calculate computing power requirements based on target environment data using a computing power requirement calculation method, adjust computing power for a data processing method based on the computing power requirements, and optimize computing power requirements based on a computing power requirement optimization method. The first detection and tuning module is connected to the data processing module.

[0099] The second detection and tuning module is used to calculate the total power demand value according to the target device data through the total power demand value calculation method, and is also used to perform secondary computing power adjustment on the computing power demand optimization method according to the total power demand value through the computing power optimization adjustment method. The second detection and tuning module is connected to the first detection and tuning module.

[0100] Specifically, the online monitoring system is applied to the online Libs laser real-time detection device. The online monitoring system acquires, processes, calculates computing power, optimizes computing power, calculates power and distributes target environment data and target device data through the data acquisition module, the data processing module, the first detection tuning module and the second detection tuning module, so as to standardize the target environment data and maintain the detection stability, thereby improving the detection efficiency and detection accuracy, saving computing power and improving resource utilization, and increasing the endurance of the online Libs laser real-time detection device. The online monitoring system acquires data in real time through the data acquisition module. The target environment data and target device data are obtained to improve the detection efficiency. The online monitoring system simplifies the data through the data processing module while retaining important information, thereby improving the detection accuracy and efficiency. The online monitoring system saves the computing power of the online Libs laser real-time detection device through the first detection tuning module, ensures accurate computing power calculation, avoids resource waste, and thus improves the detection efficiency. The online monitoring system reduces power consumption through the second detection tuning module to make the power reserve more durable, avoid power consumption, and improve resource utilization, thereby increasing the endurance of the online Libs laser real-time detection device and improving the efficiency and accuracy of detection.

[0101] Specifically, the data acquisition module acquires target environment data and target device data. The data acquisition module acquires target environment data through a data acquisition instrument, and the data acquisition module acquires target device data through a sensor node group.

[0102] Specifically, the data acquisition module acquires target environment data and target device data in real time, so as to facilitate subsequent data processing of the target environment data and target device data, thereby improving detection efficiency.

[0103] Specifically, the data processing module processes the target environment data according to a data processing method, and the data processing method includes a data preprocessing strategy, a feature extraction strategy, and a data fusion strategy.

[0104] Specifically, the data processing module performs data preprocessing on the target environment data according to a data preprocessing strategy, and the data preprocessing strategy includes:

[0105] Step J01: Based on the spectrum data value set s={s1,s2,s3,...,si}, the spectrum data minimum value S min and the maximum value of the spectral data S max Calculate the normalized spectral data value set Sop = {Sop1, Sop2, Sop3, ..., Sopi} to obtain the normalized spectral data value set Sop = {Sop1, Sop2, Sop3, ..., Sopi}, set

[0106] Step J02: Calculate the processed time-resolved data set t`={t`1,t`2,t`3,...,t`j} according to the segmentation window size h and the time-resolved data set t={t1,t2,t3,...,tj} to obtain the processed time-resolved data set t`={t`1,t`2,t`3,...,t`j}. Set

[0107] Step J03, obtaining projection coordinates (αm, βm, γm) of each spatial position data according to the universal transverse Mercator grid system;

[0108] In step J04, the normalized spectral data value set Sop = {Sop1, Sop2, Sop3, ..., Sopi}, the processed time-resolved data set t` = {t`1, t`2, t`3, ..., t`j} and the projection coordinates (αm, βm, γm) are used as the first target processing data.

[0109] Specifically, the spectral data value set s = {s1, s2, s3, ..., si} refers to a data set composed of the values of each spectral data point, wherein i is the order of the spectral data point, i is a positive integer, the maximum value of i is the total number of spectral data points, s1 is the value of the first spectral data point in the spectral data value set, s2 is the value of the second spectral data point in the spectral data value set, s3 is the value of the third spectral data point in the spectral data value set, si is the value of the i-th spectral data point in the spectral data value set, and the minimum spectral data S min It refers to the value with the minimum signal intensity in the spectral data. This embodiment does not apply to the minimum value S of the spectral data. min The acquisition method is limited, and those skilled in the art can freely choose according to actual needs, as long as the minimum value S of the spectral data is satisfied. min The acquisition requirement can be obtained, such as in this embodiment, by using spectrum analysis software to analyze the minimum value S of the spectrum data.min To obtain the maximum value of the spectral data S max It refers to the value with the maximum signal intensity in the spectral data. This embodiment does not apply to the maximum value S of the spectral data. max The acquisition method is limited. For example, in this embodiment, the maximum value S of the spectral data is obtained by using spectral analysis software. maxThe normalized spectral data value set Sop={Sop1, Sop2, Sop3, ..., Sopi} refers to a data set composed of normalized spectral data values obtained by mapping the spectral data value set s={s1, s2, s3, ..., si} to the interval [0, 1] in step J01, wherein Sop1 is the first normalized spectral data value in the normalized spectral data value set Sop, Sop2 is the second normalized spectral data value in the normalized spectral data value set Sop, Sop3 is the third normalized spectral data value in the normalized spectral data value set Sop, and Sopi is the i-th normalized spectral data value in the normalized spectral data value set Sop. The split window size h refers to selecting a preset number a of data points before and after the data point to be processed, with the data point to be currently processed as the center, and calculating the split window size h according to the data points to be split by the preset number a, and setting h=2a+ 1. The preset number of segments a refers to a preset number value. This embodiment does not limit the preset number of segments a. For example, in this embodiment, the preset number of segments a is set to 2, and the segmentation window size h = 5. The time-resolved data set t = {t1, t2, t3, ..., tj} refers to a data set consisting of the values of each time-resolved data point, where j is the order of the time-resolved data point, j is a positive integer, and the maximum value of j is the total number of time-resolved data points. t1 is the value of the first time-resolved data point in the time-resolved data set, t2 is the value of the second time-resolved data point in the time-resolved data set, t3 is the value of the third time-resolved data point in the time-resolved data set, and tj is the value corresponding to the jth data point in the time-resolved data set. The processed time-resolved data set t` = {t`1, t`2, t`3, ..., t`j} refers to the time-resolved data set t = {t1, t2, t3, ..., tj} processed in step J02.,tj} is a data set composed of the values of each processed time-resolved data point obtained after calculation and processing, wherein t`1 is the value of the first processed time-resolved data point in the processed time-resolved data set, t`2 is the value of the second processed time-resolved data point in the processed time-resolved data set, t`3 is the value of the third processed time-resolved data point in the processed time-resolved data set, and t`j is the value of the jth processed time-resolved data point in the processed time-resolved data set. The universal transverse Mercator grid system refers to a calculation software that takes longitude coordinates, latitude coordinates and altitude as input data and projection coordinates as output data. This embodiment does not limit the implementation method of the universal transverse Mercator grid system, such as through professional geographic information system software. The spatial position data refers to the spatial position data collected by a data acquisition instrument for a preset number of measurement points at the position. The position preset Assume that the number refers to a preset value. This embodiment does not limit the number of preset locations. Persons skilled in the art can freely choose according to actual circumstances, as long as the requirement for limiting the number of preset locations is met. For example, the number of preset locations is set to 3000. The projection coordinates (αm, βm, γm) are plane rectangular coordinates representing the geographic location information of each measurement point. The Earth's surface is divided into 60 projection zones, starting from 180° west longitude and extending every 6° from west to east, numbered 1 to 60. Within each projection zone, the central meridian is the vertical axis, the equator is the horizontal axis, and the intersection of the central meridian and the equator is the origin. On the vertical axis, positive values are in the north and negative values are in the south. On the horizontal axis, positive values are in the east and negative values are in the west. αm refers to the coordinate value indicating the north-south direction, βm refers to the coordinate value indicating the east-west direction, and γm refers to the projection zone number. 0m is the order of the projection coordinates, and m is a positive integer. The maximum value of m is the total number of measurement points.

[0110] Specifically, the data processing module performs data preprocessing on the target environment data so as to standardize each data point in the target environment data, improve data quality and data consistency and comparability, and thus improve detection efficiency.

[0111] Specifically, when the data processing module performs data preprocessing on the target environment data according to the data preprocessing strategy, the data processing module constructs a spectral data feature vector model according to a spectral data feature vector model construction method, and the spectral data feature vector model construction method includes:

[0112] Step K01, divide the historical spectral dataset into a 70% spectral training set, a 20% spectral validation set, and a 10% spectral test set;

[0113] Step K02, initializing the spectral parameters of the first convolutional neural network model;

[0114] Step K03: Input the spectral training set into the first convolutional neural network model after parameter initialization for training, input the spectral validation set into the trained first convolutional neural network model, optimize the spectral parameters of the trained first convolutional neural network model, and then input the spectral test set into the first convolutional neural network model after parameter optimization for testing, and output the test accuracy;

[0115] Step K04, outputting the first convolutional neural network model after parameter optimization with an accuracy rate of 90% as a spectral data feature vector model;

[0116] When the data processing module performs data preprocessing on the target environment data according to the data preprocessing strategy, the data processing module constructs a time-resolved feature vector model according to a time-resolved feature vector model construction method, and the time-resolved feature vector model construction method includes:

[0117] Step C01, divide the historical time-resolved dataset into a 70% time training set, a 20% time validation set, and a 10% time test set;

[0118] Step C02, initializing the time parameters of the second convolutional neural network model;

[0119] Step C03: Input the time training set into the second convolutional neural network model after parameter initialization for training, input the time verification set into the trained second convolutional neural network model, optimize the time parameters of the trained second convolutional neural network model, and then input the time test set into the second convolutional neural network model after parameter optimization for testing, and output the test accuracy;

[0120] In step C04, the second convolutional neural network model after parameter optimization with an accuracy rate of 90% is output as a time-resolved feature vector model.

[0121] Specifically, the historical spectral data set refers to each normalized spectral data value generated historically and the spectral feature vector corresponding to each normalized spectral data value generated historically, each normalized spectral data value is used as input data of the spectral data feature vector model, and the spectral feature vector is used as output data of the spectral data feature vector model, the spectral training set refers to a data set in the historical spectral data set used to train the first convolutional neural network model, the spectral verification set refers to a data set used to adjust the parameters of the first convolutional neural network model during the training process, the spectral test set refers to a data set used to evaluate the final performance of the first convolutional neural network model after the first convolutional neural network model completes the spectral parameter optimization, and the first convolutional neural network model refers to a data set used to train the first convolutional neural network model. The statistical model for describing the linear relationship between each normalized spectral data value and the spectral feature vector is used. The spectral parameter initialization refers to the process of setting initial values for the spectral weights and spectral biases of the first convolutional neural network model when constructing the spectral data feature vector model. The spectral weights refer to the parameters used to measure the connection strength between neurons in the first convolutional neural network model. The spectral bias refers to the parameter output by neurons in the first convolutional neural network model when the input is zero. This embodiment does not limit the spectral parameter initialization method. The relevant technical personnel of this embodiment can freely choose according to actual needs, and only need to meet the need to set initial values for the spectral weights and spectral biases of the first convolutional neural network model, such as random initialization. Spectral parameter optimization refers to the process of optimizing the performance of the first convolutional neural network model on the spectral validation set by adjusting the spectral weights and spectral biases of the first convolutional neural network model when evaluating the trained first convolutional neural network model using the spectral validation set. This embodiment does not limit the manner in which the spectral weights and spectral biases of the first convolutional neural network model are adjusted. Relevant technicians in this field can freely choose according to actual needs. They only need to adjust the spectral weights and spectral biases of the first convolutional neural network model to optimize the performance of the first convolutional neural network model on the spectral validation set, such as by adjusting through computing software. The historical time-resolved data set refers to the processed time-resolved data and historical The generated time-resolved data corresponds to a time feature vector, the processed time-resolved data is used as input data of the time-resolved feature vector model, and the time feature vector is used as output data of the time-resolved feature vector model, the time training set refers to a data set in the historical time-resolved data set for training the second convolutional neural network model, the time verification set refers to a data set used to adjust the parameters of the second convolutional neural network model during the training process, the time test set refers to a data set used to evaluate the final performance of the second convolutional neural network model after the second convolutional neural network model completes time parameter optimization, and the second convolutional neural network model refers to a statistical model used to describe the linear relationship between the processed time-resolved data and the time feature vector,The time parameter initialization refers to the process of setting initial values for the time weights and time biases of the second convolutional neural network model when constructing the time-resolved feature vector model. This embodiment does not limit the time parameter initialization method. Personnel skilled in the art can freely choose according to actual needs, as long as the time weights and time biases of the second convolutional neural network model are set to initial values, such as random initialization. The time parameter optimization refers to the process of adjusting the time weights and time biases of the second convolutional neural network model to achieve optimal performance on the time validation set when evaluating the trained second convolutional neural network model using the time validation set. This embodiment does not limit the method for adjusting the spectral weights and spectral biases of the second convolutional neural network model. Personnel skilled in the art can freely choose according to actual needs, as long as the time weights and time biases of the second convolutional neural network model are adjusted to achieve optimal performance on the spectral validation set, such as adjustment through computing software. The time weight refers to the parameter used in the second convolutional neural network model to measure the connection strength between neurons, and the time bias refers to the parameter output by neurons in the second convolutional neural network model when the input is zero.

[0122] Specifically, the data processing module constructs a spectral data feature vector model and a time-resolved feature vector model to facilitate the subsequent convenient and accurate acquisition of each feature vector, thereby improving detection efficiency.

[0123] Specifically, the data processing module extracts features from the first target processing data according to a feature extraction strategy, wherein the feature extraction strategy includes:

[0124] Step D01, inputting a normalized spectral data value set Sop={Sop1, Sop2, Sop3, ..., Sopi} into a spectral data feature vector model to obtain a spectral feature vector set Foh={Foh1, Foh2, Foh3, ..., Fohi} output by the spectral data feature vector model;

[0125] Step D02, inputting the processed time-resolved data set t`={t`1, t`2, t`3, ..., t`j} into the time-resolved feature vector model to obtain the time feature vector set Fom={Fom1, Fom2, Fom3, ..., Fomj} output by the time-resolved feature vector model;

[0126] Step D03, calculate each spatial distance dm according to each projection coordinate (αm, βm, γm) and each preset coordinate (αy, βy, γy), and set

[0127] In step D04 , the spectral feature vector set Foh={Foh1, Foh2, Foh3, ..., Fohi}, the time feature vector set Fom={Fom1, Fom2, Fom3, ..., Fomj} and each spatial distance dm are used as second target processing data.

[0128] Specifically, the spectral feature vector set Foh = {Foh1, Foh2, Foh3, ..., Fohi} refers to a data set composed of vector values obtained by taking the normalized spectral data value set Sop = {Sop1, Sop2, Sop3, ..., Sopi} as the input data of the spectral data feature vector model, which is used to represent the feature information corresponding to the normalized spectral data value set Sop = {Sop1, Sop2, Sop3, ..., Sopi}, wherein Foh1 is the first vector value in the spectral feature vector set Foh, Foh2 is the second vector value in the spectral feature vector set Foh, Foh3 is the third vector value in the spectral feature vector set Foh, and Fohi is the i-th vector value in the spectral feature vector set Foh, and the time feature vector set Fom = {Fom1, Fom2, Fom3, ..., Fomj} refers to the processed time-resolved data set t` = {t`1, t`2, t`3, ..., t`j} as the time A data set consisting of vector values obtained from the input data of the time-resolved feature vector model is used to represent the feature information corresponding to the processed time-resolved data set t`={t`1, t`2, t`3, ..., t`j}, wherein Fom1 refers to the first vector value in the time feature vector set Fom, Fom2 refers to the second vector value in the time feature vector set Fom, Fom3 refers to the third vector value in the time feature vector set Fom, and Fomj refers to the j-th vector value in the time feature vector set Fom. The preset coordinates (αy, βy, γy) refer to the preset position coordinates of each measurement point. This embodiment does not limit the preset coordinates. Relevant technical personnel in this field can freely choose according to actual needs, as long as the requirement of calculating each spatial distance dm is met. For example, αy=50000, βy=400000, and γy=32N are set. The spatial distance dm refers to the distance between each projection coordinate (αm, βm, γm) and each preset coordinate (αy, βy, γy).

[0129] Specifically, the data processing module extracts features from the first target processing data in order to reduce the data dimension and ease the difficulty of data analysis, thereby obtaining vector features representing the data and improving detection accuracy.

[0130] Specifically, the data processing module performs data fusion on the second target processing data according to a data fusion strategy, and the data fusion strategy includes:

[0131] Step E01, matrixing the spectral feature vector set Foh = {Foh1, Foh2, Foh3, ..., Fohi}, the time feature vector set Fom = {Fom1, Fom2, Fom3, ..., Fomj} and each spatial distance dm to obtain a first matrix X;

[0132] Step E02, setting a first data sampling point Xzv, where z is the zth row in the first matrix X, and v is the vth column in the first matrix X;

[0133] Step E03, calculate the mean μv and the standard deviation σv of each column according to the total number of measurement points m and the first data sampling point Xzv, and obtain the mean μv and the standard deviation σv of each column, and set

[0134] Step E04, calculate the second data sampling point Gz0v0 according to the mean value μv of each column and the standard deviation σv of each column, and obtain the second data sampling point Gz0v0, and set

[0135] Step E05, performing matrix processing on the second data sampling point Gz0v0 to obtain a second matrix G;

[0136] Step E06, transposing the second matrix G to obtain a transposed second matrix G';

[0137] Step E07, calculate the covariance matrix C according to the total number of measurement points m, the second matrix G and the transposed second matrix G' to obtain the covariance matrix C, set

[0138] Step E08, performing eigendecomposition on the covariance matrix C to obtain eigenvalues;

[0139] Step E09: Arrange the eigenvalues in a preset order to obtain sorted eigenvalues, perform value processing on the sorted eigenvalues according to the value ratio, and obtain the laser detection result.

[0140] Specifically, the matrix processing refers to the process of arranging different types of data vectors in rows and columns so that they have specific positions in the matrix. For example, if the total number of normalized spectral data values in the spectral feature vector set Foh = {Foh1, Foh2, Foh3, ..., Fohi} is ik, the total number of processed time-resolved data in the time feature vector set Fom = {Fom1, Fom2, Fom3, ..., Fomj} is jk, and the total number of measurement points is mk, then the first matrix X is mk×(ik+jk+1), the first matrix X refers to the data structure obtained by matrix processing the data points in the second target processed data, the first data sampling point Xzv refers to the data point in the zth row and vth column of the first matrix X, wherein z is the order of the number of rows in the first matrix X, z is a positive integer, and the maximum value of z is the total number of rows of the first matrix X, v is the order of the number of columns in the first matrix X, v is a positive integer, and the maximum value of v is the total number of columns of the first matrix X, and the column mean μv refers to the mean value of the Vth column in the first matrix X. The mean, the standard deviation σv of each column refers to a statistic that measures the degree of dispersion of the data in the vth column of the first matrix X, the second data sampling point Gz0v0 refers to the data point in the z0th row and v0th column of the second matrix G, wherein z0 is the order of the number of rows in the second matrix G, z0 is a positive integer, the maximum value of z0 is the total number of rows of the second matrix G, v0 is the order of the number of columns in the second matrix G, v0 is a positive integer, the maximum value of v0 is the total number of columns of the second matrix G, and the second matrix G refers to the second data sampling point Gz0v0 processed by matrix The data structure obtained after the transposition refers to the operation of performing row-column transformation on the matrix, such as exchanging the rows and columns in the second matrix G, the transposed second matrix G' refers to the new matrix obtained after transposing the second matrix G, the covariance matrix C refers to a matrix used to measure the covariance relationship between variables, such as describing the correlation between spectral eigenvectors, time eigenvectors and spatial distances, the eigendecomposition refers to a method of decomposing a matrix into eigenvalues and eigenvectors, such as calculating the eigenvalues according to the covariance matrix C and the identity matrix IP, setting The unit matrix IP refers to a preset standard matrix. This embodiment does not limit the unit matrix IP. Relevant technicians in this field can freely choose it according to actual needs, as long as the requirement for calculating the eigenvalue is met. For example, the unit matrix IP is set to a 3×3 matrix. The eigenvalue refers to the data obtained after eigendecomposition of the covariance matrix C to measure the importance of data features. The larger the eigenvalue, the greater the data importance. The preset order refers to a preset arrangement order. This embodiment does not limit the specific arrangement method of arranging the eigenvalues according to the preset order. Those skilled in the art can set it according to actual needs. For example, the eigenvalues are arranged from large to small according to the preset order. The sorted eigenvalues refer to the numerical sequence obtained after the eigenvalues are arranged according to the preset order. The value ratio refers to the percentage of values taken in the sorted eigenvalues. This embodiment does not limit the number of values. For example, the number of values is set to 80%. The value processing refers to the process of taking values of the sorted eigenvalues according to the value ratio, such as extracting the first 80% of the sorted eigenvalues, removing the last 20% of the data, and using the first 80% of the sorted eigenvalues as the laser detection results.

[0141] Specifically, the data processing module simplifies the vector data by matrixing the second target processing data, thereby facilitating subsequent data processing. The data processing module obtains the second matrix G by calculating the mean and standard deviation of each column of the first data sampling point Xzv, so as to convert the features of different dimensions in the second target processing data into data with the same scale, thereby improving data accuracy. The data processing module also transposes the second matrix G to analyze the relationship between different dimensions of the data and observe the correlation between different features, thereby improving the accuracy of data information. The data processing module arranges and takes values of the eigenvalues to extract data with high importance, so that the data is simplified while accurately retaining the information in the data, thereby improving data accuracy.

[0142] Specifically, the first detection and tuning module calculates the computing power requirement according to the target environment data using a computing power requirement calculation method, and the computing power requirement calculation method includes:

[0143] Step B01, calculate the spectral data computing power requirement Q1 according to the spectral data volume Iv, obtain the spectral data computing power requirement Q1, and set Q1=4i';

[0144] Step B02: Calculate the time-resolved computing power requirement Q2 based on the time-resolved data point jq and the segmentation window size h to obtain the time-resolved computing power requirement Q2, and set Q2 = 3 × jq - 2 × h + 1;

[0145] Step B03: Calculate the spatial location computing power requirement Q3 based on the total number of measurement points m to obtain the spatial location computing power requirement Q3, and set Q3 = 38 × n3;

[0146] Step B04: Calculate the total computing power requirement Q based on the spectral data computing power requirement Q1, the time-resolved computing power requirement Q2, and the spatial position computing power requirement Q3 to obtain the total computing power requirement Q, and set Q = Q1 + Q2 + Q3;

[0147] When the first detection and tuning module adjusts the computing power of the data processing method according to the computing power demand, the total computing power demand Q is compared with the preset computing power demand Q0, the total computing power demand is judged according to the comparison result, and the computing power of the data processing method is adjusted according to the judgment result, wherein:

[0148] When Q≤Q0, the first detection and tuning module determines that the total computing power demand is low and does not adjust the computing power of the data processing method;

[0149] When Q>Q0, the first detection tuning module determines that the total computing power demand is high, adjusts the computing power of the data processing method, sends the second target processing data to the cloud, performs data fusion through the cloud, and obtains the laser detection results.

[0150] Specifically, the spectral data volume Iv refers to the total number of data points of spectral data in the target environment data, the spectral data computing power requirement Q1 refers to the computing power required for data processing of the spectral data volume Iv, the time-resolved data points jq refers to the total number of data points of time-resolved data in the target environment data, the time-resolved computing power requirement Q2 refers to the computing power required for data processing of the time-resolved data points jq, and the spatial position computing power requirement Q3 refers to the computing power required for data processing of each projection coordinate.

[0151] Specifically, the first detection tuning module calculates the computing power requirements to facilitate subsequent optimization of the computing power requirements and flexibly adjust the computing power, thereby improving detection efficiency.

[0152] Specifically, the first detection and tuning module optimizes the computing power requirement according to a computing power requirement optimization method, and the computing power requirement optimization method includes:

[0153] Step L01, calculating the data generation rate Vs according to the data acquisition frequency Cf and the data acquisition amount Ns, obtaining the data generation rate Vs, and setting Vs = Cf × Ns;

[0154] Step L02: Compare the data generation rate Vs with the preset generation rate Vs0, determine the state of the data generation rate Vs based on the comparison result, and update the total computing power requirement Q based on the determination result, where:

[0155] When Vs≤Vs0, the first detection and tuning module determines that the data generation rate Vs is a low rate and does not update the total computing power demand Q;

[0156] When Vs>Vs0, the first detection and tuning module determines that the data generation rate Vs is at a high rate, updates the total computing power demand Q, and sets the computing power update coefficient to ε, setting ε=1.3-0.3e -0.7×(Vs-Vs0) The updated total computing power requirement is Q`, Q`=ε×Q. Compare the updated total computing power requirement Q` with the preset computing power requirement Q0, and re-evaluate the total computing power requirement.

[0157] Step L03: Compare the parallel task quantity Bq with the preset parallel task quantity Bq0, judge the situation of the parallel task quantity Bq based on the comparison result, and make a computing power correction to the preset generation rate Vs0 based on the judgment result, where:

[0158] When Bq≤Bq0, the first detection and tuning module determines that the parallel task amount Bq is a low task amount and does not perform computing power correction on the preset generation rate Vs0;

[0159] When Bq>Bq0, the first detection and tuning module determines that the parallel task volume Bq is a high task volume, performs a computing power correction on the preset generation rate Vs0, sets the rate correction coefficient to θ, and sets The corrected preset generation rate is Vs0`, Vs0`=Vs0×θ, and the preset generation rate Vs0 is replaced by the corrected preset generation rate Vs0`, and the data generation rate Vs is re-compared with the corrected preset generation rate Vs0`.

[0160] Specifically, the preset computing power demand Q0 refers to a preset value for judging the demand situation of the total computing power demand Q. This embodiment does not limit the preset computing power demand Q0. Relevant technical personnel in this field can freely choose according to actual needs. It is only necessary to meet the demand for limiting the preset computing power demand Q0. For example, this embodiment sets the preset computing power demand Q0 = 800KFLOPS. The cloud refers to a remote server cluster based on the Internet for data fusion and storage of the second target processing data in the data processing module. This embodiment does not limit the choice of the cloud. Relevant technical personnel in this field can freely choose according to actual needs. It is only necessary to meet the demand for data fusion of the second target processing data. For example, the Internet, the total computing power demand Q0 is 800KFLOPS. The power demand situation refers to the total computing power demand level determined by the preset computing power demand Q0. The total computing power demand situation includes a high demand for the total computing power demand situation and a low demand for the total computing power demand situation. The data acquisition frequency Cf refers to the number of times the online Libs laser real-time detection device collects data through the data acquisition instrument within the preset time. This embodiment does not limit the method for obtaining the data acquisition frequency Cf. Relevant technical personnel in this field can freely choose according to actual needs. It is only necessary to meet the demand for obtaining the data acquisition frequency Cf. For example, this embodiment obtains it through the instruction manual of the online Libs laser real-time detection device. The preset time refers to a pre-set time length. This embodiment does not limit the preset time. This embodiment Relevant technical personnel in the field can make free choices according to actual needs, and only need to meet the requirements of limiting the preset time. For example, if the preset time is set to 1 hour, the data collection volume Ns refers to the total amount of data collected when the online Libs laser real-time detection device performs a single data collection. This embodiment does not limit the method of obtaining the data collection volume Ns. Relevant technical personnel in this field can make free choices according to actual needs, and only need to meet the requirements of obtaining the data collection volume Ns. For example, if data monitoring software is installed, the data generation rate Vs refers to the rate at which the online Libs laser real-time detection device collects data within the preset time, which is used to indicate the speed of generating laser detection results. The preset generation rate Vs0 refers to the data The state of the data generation rate Vs is judged by a preset value. This embodiment does not limit the preset generation rate Vs0. Relevant technicians in this field can freely choose according to actual needs. They only need to meet the demand for limiting the preset generation rate Vs0. For example, setting Vs0 = 3MB / s, the state of the data generation rate Vs refers to the speed of the data generation rate judged by the preset generation rate Vs0. The state of the data generation rate Vs includes the state of the generation rate Vs being a low rate and the state of the data generation rate Vs being a high rate. The parallel task volume Bq refers to the number of detection tasks running simultaneously by the online Libs laser real-time detection device. This embodiment does not limit the method for obtaining the parallel task volume Bq.Those skilled in the art can freely choose according to actual needs, and only need to meet the need to obtain the parallel task volume Bq. For example, in this embodiment, it is obtained through the instruction manual of the online Libs laser real-time detection device. The preset parallel task volume Bq0 refers to the preset value for judging the situation of the parallel task volume Bq. This embodiment does not limit the preset parallel task volume Bq0. Those skilled in the art can freely choose according to actual needs, and only need to meet the need to limit the preset parallel task volume Bq0. For example, in this embodiment, Bq0=5 is set. The situation of the parallel task volume Bq refers to the degree of heaviness of the parallel tasks judged by the preset parallel task volume Bq0. The situation of the parallel task volume Bq includes the situation of the parallel task volume Bq being a low task volume and the situation of the parallel task volume Bq being a high task volume.

[0161] Specifically, the first detection tuning module judges the demand situation of the total computing power demand Q, so that when the total computing power demand Q is high, the data fusion process is calculated by the cloud, thereby saving the computing power of the online Libs laser real-time detection device and ensuring the detection effect. The first detection tuning module judges the state of the data generation rate Vs, so that the total computing power demand Q increases with the increase of the data generation rate Vs, avoiding inaccurate computing power demand Q when the data generation rate Vs is too high, thereby timely adjusting the data processing method to maintain the detection stability of the online Libs laser real-time detection device. The first detection tuning module judges the situation of the parallel task volume Bq, so that the preset generation rate Vs0 decreases with the increase of the parallel task volume Bq, avoiding inaccurate state judgment of the data generation rate Vs when the parallel task volume Bq is too much, thereby ensuring accurate computing power calculation of the online Libs laser real-time detection device, avoiding waste of resources, and improving the accuracy of detection results and detection efficiency.

[0162] Specifically, the second detection and tuning module calculates the total power requirement value according to the target device data using a total power requirement value calculation method, and the total power requirement value calculation method includes:

[0163] Step P01, calculate the power R1 required for data acquisition based on the working voltage Ug, the working current Ig, and the data acquisition frequency Cf, and obtain the power R1 required for data acquisition, and set R1 = Ug × Ig × Cf;

[0164] Step P02, calculating the power R2 required for data processing based on the data processing power Pc and the data processing time tl, obtaining the power R2 required for data processing, and setting R2 = Pc × tl;

[0165] Step P03: Calculate the power R3 required for computing power analysis based on computing power calculation power Px and computing power calculation time Tx to obtain the power R3 required for computing power analysis, and set R3 = Px × Tx;

[0166] Step P04, calculating the power R4 required for power distribution based on the control circuit power Pk, the number of communications nt, and the communication time tt, to obtain the power R4 required for power distribution, and setting R4 = Pk × nt × tt;

[0167] Step P05, calculate the total power demand R based on the power required for data acquisition R1, the power required for data processing R2, the power required for computing power analysis R3 and the power required for power distribution R4, and obtain the total power demand R, and set R=R1+R2+R3+R4.

[0168] Specifically, the power R1 required for data acquisition refers to the power value required for data acquisition, the data processing power Pc refers to the amount of data processed by the data processing module within the preset processing time when performing data processing, and is used to measure the processing capacity and efficiency of the data processing module. This embodiment does not limit the preset processing time. For example, the preset processing time is set to 20 minutes. The data processing time t1 refers to the time required for measuring a single measurement point according to the data processing power Pc to obtain the target environment data. The power R2 required for data processing refers to the power required for data processing by the data processing module. The computing power calculation power Px refers to the amount of data calculated within the preset calculation time when the first detection and tuning module calculates the computing power demand. This embodiment does not limit the preset calculation time. For example, the preset calculation time is set to 10 minutes. The computing power calculation time Tx refers to the time when the first detection and tuning module calculates the computing power demand for the target environment data. The power R3 required for computing power analysis refers to the power required for the first detection and tuning module. The control circuit power Pk refers to the electric power consumed by the control circuit of the online Libs laser real-time detection device during operation, the communication number nt refers to the number of times the online Libs laser real-time detection device transmits the laser detection results to the cloud, the communication time tt refers to the time when the online Libs laser real-time detection device transmits the laser detection results to the cloud, the power required for power distribution R4 refers to the power consumed for power distribution and data upload of the online Libs laser real-time detection device, this embodiment does not limit the acquisition method of data processing power Pc, data processing time tl, computing power calculation power Px, computing power calculation time Tx, control circuit power Pk, communication number nt and communication time tt, relevant technical personnel in this field can freely choose according to actual needs, and only need to meet the requirements for obtaining data processing power Pc and data processing time tl, such as obtaining through the instruction manual of the online Libs laser real-time detection device in this embodiment, the total power demand R refers to the total power required for the online monitoring system to complete control.

[0169] Specifically, the second detection and tuning module calculates the total power demand R to facilitate subsequent adjustment of the power, thereby improving resource utilization.

[0170] Specifically, the second detection and tuning module performs a secondary computing power adjustment on the computing power demand optimization method according to the computing power optimization adjustment method, and the computing power optimization adjustment method includes:

[0171] Step U01: compare the total power demand R with the remaining power R0, determine the demand for the total power demand R based on the comparison result, and perform secondary computing power adjustments on the data processing method and computing power demand optimization method based on the determination result, where:

[0172] When R≤R0, the second detection and tuning module determines that the total power demand R is low, and does not perform secondary computing power adjustment on the data processing method and the computing power demand optimization method;

[0173] When R>R0, the second detection and tuning module determines that the demand situation of the total power demand R is high, performs a secondary computing power adjustment on the computing power demand optimization method, sets the adjusted data acquisition frequency to Cf`, Cf`=Cf×(R-R0) / R0, and replaces the data acquisition frequency Cf with the adjusted data acquisition frequency Cf`, recalculates the data generation rate Vs, sets the adjusted preset parallel task volume to Bq0`, Bq0`=Bq0×(1-(R-R0) / R0), replaces the preset parallel task volume Bq0 with the adjusted preset parallel task volume Bq0`, and recomputes the parallel task volume Bq with the adjusted preset parallel task volume Bq0`;

[0174] Step U02: compare the working temperature Wh with the preset working temperature Wh0, determine the state of the working temperature Wh based on the comparison result, and perform a secondary calculation power correction on the remaining power R0 based on the determination result, wherein:

[0175] When Wh≤Wh0, the second detection and tuning module determines that the operating temperature Wh is in a low temperature state and does not perform a secondary computing power correction on the remaining power R0;

[0176] When Wh>Wh0, the second detection and tuning module determines that the operating temperature Wh is in a high temperature state, performs a secondary computing power correction on the power reserve R0, sets the corrected power reserve to R0', R0'=R0×Wh / Wh0, replaces the power reserve R0 with the corrected power reserve R0', and re-compares the total power demand R with the corrected power reserve R0';

[0177] In step U03, the heat dissipation efficiency Vv is compared with the preset heat dissipation efficiency Vv0, and the heat dissipation degree of the heat dissipation efficiency Vv is judged according to the comparison result. The preset working temperature Wh0 is updated with a secondary computing power according to the judgment result, wherein:

[0178] When Vv≤Vv0, the second detection and tuning module determines that the heat dissipation degree of the heat dissipation efficiency Vv is low, and does not perform a secondary computing power update on the preset operating temperature Wh0;

[0179] When Vv>Vv0, the second detection and tuning module determines that the heat dissipation degree of the heat dissipation efficiency Vv is high, and does not perform a secondary computing power update on the preset working temperature Wh0. The updated preset working temperature is set to Wh0`, Wh0`=Wh0×Vv / Vv0, and the preset working temperature Wh0 is replaced with the updated preset working temperature Wh0`, and the working temperature Wh is re-compared with the updated preset working temperature Wh0`.

[0180] Specifically, the remaining power R0 refers to the amount of power contained in the power supply at the current moment, and the preset operating temperature Wh0 refers to a preset value for judging the state of the operating temperature Wh. This embodiment does not limit the preset operating temperature Wh0, and relevant technical personnel in this field can freely choose according to actual needs, and only need to meet the requirements of limiting the preset operating temperature Wh0. For example, this embodiment sets Wh0 = 25 ° C. The heat dissipation efficiency Vv refers to the ability of the online Libs laser real-time detection device to dissipate the internally generated heat to the surrounding environment. This embodiment does not limit the method of obtaining the heat dissipation efficiency Vv. For example, this embodiment is obtained through the instruction manual of the online Libs laser real-time detection device. The preset heat dissipation efficiency Vv0 refers to a preset value for judging the heat dissipation degree of the heat dissipation efficiency Vv. This embodiment does not limit the preset heat dissipation efficiency Vv0, and relevant technical personnel in this field Technicians can make free choices based on actual needs, and only need to meet the requirements of limiting the preset heat dissipation efficiency Vv0. For example, in this embodiment, Vv0 is set to 30%. The demand situation of the total power demand R refers to the degree of demand for total power judged by the power margin R0, and the demand situation of the total power demand R includes the demand situation of the total power demand R as low demand and the demand situation of the total power demand R as high demand. The state of the working temperature Wh refers to the high and low conditions of the working temperature judged by the preset working temperature Wh0, and the state of the working temperature Wh includes the state of the working temperature Wh as a low temperature state and the state of the working temperature Wh as a high temperature state. The heat dissipation degree of the heat dissipation efficiency Vv refers to the heat dissipation capacity judged by the preset heat dissipation efficiency Vv0, and the heat dissipation degree of the heat dissipation efficiency Vv includes the heat dissipation degree of the heat dissipation efficiency Vv as low and the heat dissipation degree of the heat dissipation efficiency Vv as high.

[0181] Specifically, the second detection and tuning module judges the demand situation of the total power demand R, so that when the total power demand R is high, the data acquisition frequency Cf and the preset parallel task amount Bq0 in the computing power demand optimization method are adjusted, so that the data acquisition frequency Cf decreases as the total power demand R increases, and the preset parallel task amount Bq0 decreases as the total power demand R increases, thereby reducing power consumption, saving power margin, and avoiding resource waste. The second detection and tuning module judges the state of the operating temperature Wh, so that the power margin R0 increases as the operating temperature increases. When the working temperature Wh is high, the remaining power R0 will be more durable, avoiding inaccurate judgment of the remaining power R0 when the working temperature Wh is high, thereby improving the comprehensiveness and accuracy of the detection. The second detection tuning module judges the heat dissipation degree of the heat dissipation efficiency Vv to make the working temperature Wh0 increase with the increase of the heat dissipation efficiency Vv, avoiding inaccurate judgment of the state of the working temperature Wh0 when the heat dissipation degree is high, resulting in power consumption, thereby improving the detection efficiency, saving power, and increasing the endurance of the online Libs laser real-time detection device.

[0182] Thus far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art may make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will fall within the scope of protection of the present invention.

Claims

1. A device for online Libs laser real-time detection, characterized in that: The device comprises: A laser generator, which is connected to the data acquisition instrument, the online monitoring system, the sensor node group and the housing, and is used to emit a laser beam; A data collector, connected to the laser generator and the housing, for collecting target environment data; A power supply connected to the online monitoring system for providing power to the online Libs laser real-time detection device; An online monitoring system connected to the laser generator and the power supply, and connected to the data acquisition instrument and the sensor node group via internal connecting lines, for controlling the online Libs laser real-time detection device; A sensor node group is connected to the laser generator and the housing, wherein the sensor node group includes a voltage sensor, a current sensor, and a temperature sensor for collecting data from the target device; The housing is connected to the laser generator, the data acquisition instrument and the sensor node group, and is used to protect the online Libs laser real-time detection device.

2. The device for online Libs laser real-time detection according to claim 1, characterized in that: The online monitoring system comprises: A data acquisition module, used to acquire target environment data and target device data; A data processing module is used to process the target environment data according to the data processing method to obtain the laser detection results; A first detection and tuning module is used to calculate the computing power requirement based on the target environment data using a computing power requirement calculation method, adjust the data processing method based on the computing power requirement, and optimize the computing power requirement based on the computing power requirement optimization method; The second detection and tuning module is used to calculate the total power demand value according to the target device data through the total power demand value calculation method, and is also used to perform secondary computing power adjustment on the computing power demand optimization method according to the total power demand value through the computing power optimization adjustment method.

3. The device for online Libs laser real-time detection according to claim 2, characterized in that: The data processing module performs data preprocessing on the target environment data according to a data preprocessing strategy, wherein the data preprocessing strategy includes: Step J01: Based on the spectrum data value set s={s1,s2,s3,...,si}, the spectrum data minimum value S min and the maximum value of the spectral data S max Calculating the normalized spectral data value set Sop={Sop1, Sop2, Sop3, ..., Sopi} to obtain the normalized spectral data value set Sop={Sop1, Sop2, Sop3, ..., Sopi}; Step J02, calculating the processed time-resolved data set t`={t`1, t`2, t`3, ..., t`j} according to the segmentation window size h and the time-resolved data set t={t1, t2, t3, ..., tj} to obtain the processed time-resolved data set t`={t`1, t`2, t`3, ..., t`j}; Step J03: Obtain projection coordinates (αm, βm, γm) of each spatial position data according to the universal transverse Mercator grid system, where m is the total number of measurement points; In step J04, the normalized spectral data value set Sop = {Sop1, Sop2, Sop3, ..., Sopi}, the processed time-resolved data set t` = {t`1, t`2, t`3, ..., t`j} and the projection coordinates (αm, βm, γm) are used as the first target processing data.

4. The device for online Libs laser real-time detection according to claim 3, characterized in that: When the data processing module performs data preprocessing on the target environment data according to the data preprocessing strategy, the data processing module constructs a spectral data feature vector model according to a spectral data feature vector model construction method, and the spectral data feature vector model construction method includes: Step K01, divide the historical spectral dataset into a 70% spectral training set, a 20% spectral validation set, and a 10% spectral test set; Step K02, initializing the spectral parameters of the first convolutional neural network model; Step K03: Input the spectral training set into the first convolutional neural network model after parameter initialization for training, input the spectral validation set into the trained first convolutional neural network model, optimize the spectral parameters of the trained first convolutional neural network model, and then input the spectral test set into the first convolutional neural network model after parameter optimization for testing, and output the test accuracy; Step K04: outputting the first convolutional neural network model after parameter optimization with an accuracy rate of 90% as a spectral data feature vector model.

5. The device for online Libs laser real-time detection according to claim 4, characterized in that: The data processing module extracts features from the first target processing data according to a feature extraction strategy, wherein the feature extraction strategy includes: Step D01, inputting a normalized spectral data value set Sop={Sop1, Sop2, Sop3, ..., Sopi} into a spectral data feature vector model to obtain a spectral feature vector set Foh={Foh1, Foh2, Foh3, ..., Fohi} output by the spectral data feature vector model; Step D02, inputting the processed time-resolved data set t`={t`1, t`2, t`3, ..., t`j} into the time-resolved feature vector model to obtain the time feature vector set Fom={Fom1, Fom2, Fom3, ..., Fomj} output by the time-resolved feature vector model; Step D03, calculating each spatial distance dm according to each projection coordinate (αm, βm, γm) and each preset coordinate (αy, βy, γy); In step D04 , the spectral feature vector set Foh={Foh1, Foh2, Foh3, ..., Fohi}, the time feature vector set Fom={Fom1, Fom2, Fom3, ..., Fomj} and each spatial distance dm are used as second target processing data.

6. The on-line Libs laser real-time detection device according to claim 5, characterized in that: The data processing module performs data fusion on the second target processing data according to a data fusion strategy, wherein the data fusion strategy includes: Step E01, matrixing the spectral feature vector set Foh = {Foh1, Foh2, Foh3, ..., Fohi}, the time feature vector set Fom = {Fom1, Fom2, Fom3, ..., Fomj} and each spatial distance dm to obtain a first matrix X; Step E02, setting a first data sampling point Xzv, where z is the zth row in the first matrix X, and v is the vth column in the first matrix X; Step E03, calculating the mean μv of each column and the standard deviation σv of each column according to the total number of measurement points m and the first data sampling point Xzv, to obtain the mean μv of each column and the standard deviation σv of each column; Step E04, calculating the second data sampling point Gz0v0 according to the mean value μv of each column and the standard deviation σv of each column to obtain the second data sampling point Gz0v0; Step E05, performing matrix processing on the second data sampling point Gz0v0 to obtain a second matrix G; Step E06, transposing the second matrix G to obtain a transposed second matrix G'; Step E07, calculating the covariance matrix C according to the total number of measurement points m, the second matrix G, and the transposed second matrix G' to obtain the covariance matrix C; Step E08, performing eigendecomposition on the covariance matrix C to obtain eigenvalues; Step E09: Arrange the eigenvalues in a preset order to obtain sorted eigenvalues, perform value processing on the sorted eigenvalues according to the value ratio, and obtain the laser detection result.

7. The device for online Libs laser real-time detection according to claim 2, characterized in that: The first detection and tuning module calculates the computing power requirement according to the target environment data using a computing power requirement calculation method, wherein the computing power requirement calculation method includes: Step B01, calculate the spectral data computing power requirement Q1 according to the spectral data volume Iv, obtain the spectral data computing power requirement Q1, and set Q1=4i'; Step B02: Calculate the time-resolved computing power requirement Q2 based on the time-resolved data point jq and the segmentation window size h to obtain the time-resolved computing power requirement Q2, and set Q2 = 3 × jq - 2 × h + 1; Step B03: Calculate the spatial location computing power requirement Q3 based on the total number of measurement points m to obtain the spatial location computing power requirement Q3, and set Q3 = 38 × n3; Step B04: Calculate the total computing power requirement Q based on the spectral data computing power requirement Q1, the time-resolved computing power requirement Q2, and the spatial position computing power requirement Q3 to obtain the total computing power requirement Q, and set Q = Q1 + Q2 + Q3; When the first detection and tuning module adjusts the computing power of the data processing method according to the computing power demand, the total computing power demand Q is compared with the preset computing power demand Q0, the total computing power demand is judged according to the comparison result, and the computing power of the data processing method is adjusted according to the judgment result, wherein: When Q≤Q0, the first detection and tuning module determines that the total computing power demand is low and does not adjust the computing power of the data processing method; When Q>Q0, the first detection tuning module determines that the total computing power demand is high, adjusts the computing power of the data processing method, sends the second target processing data to the cloud, performs data fusion through the cloud, and obtains the laser detection results.

8. The device for online Libs laser real-time detection according to claim 7, characterized in that: The first detection and tuning module optimizes the computing power requirement according to a computing power requirement optimization method, wherein the computing power requirement optimization method includes: Step L01, calculating the data generation rate Vs according to the data acquisition frequency Cf and the data acquisition amount Ns, obtaining the data generation rate Vs, and setting Vs = Cf × Ns; Step L02: Compare the data generation rate Vs with the preset generation rate Vs0, determine the state of the data generation rate Vs based on the comparison result, and update the total computing power requirement Q based on the determination result, where: When Vs≤Vs0, the first detection and tuning module determines that the data generation rate Vs is a low rate and does not update the total computing power demand Q; When Vs>Vs0, the first detection and tuning module determines that the data generation rate Vs is at a high rate, updates the total computing power demand Q, and sets the computing power update coefficient to ε, setting ε=1.3-0.3e -0.7×(Vs-Vs0) The updated total computing power requirement is Q`, Q`=ε×Q. Compare the updated total computing power requirement Q` with the preset computing power requirement Q0, and re-evaluate the total computing power requirement. Step L03: Compare the parallel task quantity Bq with the preset parallel task quantity Bq0, judge the situation of the parallel task quantity Bq based on the comparison result, and make a computing power correction to the preset generation rate Vs0 based on the judgment result, where: When Bq≤Bq0, the first detection and tuning module determines that the parallel task amount Bq is a low task amount and does not perform computing power correction on the preset generation rate Vs0; When Bq>Bq0, the first detection and tuning module determines that the parallel task volume Bq is a high task volume, performs computing power correction on the preset generation rate Vs0, and sets the rate correction coefficient to θ; the corrected preset generation rate is Vs0`, Vs0`=Vs0×θ, and the preset generation rate Vs0 is replaced with the corrected preset generation rate Vs0`, and the data generation rate Vs is re-compared with the corrected preset generation rate Vs0`.

9. The device for online Libs laser real-time detection according to claim 2, characterized in that: The second detection and tuning module calculates the total power requirement value according to the target device data using a total power requirement value calculation method, wherein the total power requirement value calculation method includes: Step P01, calculate the power R1 required for data acquisition based on the working voltage Ug, the working current Ig, and the data acquisition frequency Cf, and obtain the power R1 required for data acquisition, and set R1 = Ug × Ig × Cf; Step P02, calculating the power R2 required for data processing based on the data processing power Pc and the data processing time tl, obtaining the power R2 required for data processing, and setting R2 = Pc × tl; Step P03: Calculate the power R3 required for computing power analysis based on computing power calculation power Px and computing power calculation time Tx to obtain the power R3 required for computing power analysis, and set R3 = Px × Tx; Step P04, calculating the power R4 required for power distribution based on the control circuit power Pk, the number of communications nt, and the communication time tt, to obtain the power R4 required for power distribution, and setting R4 = Pk × nt × tt; Step P05, calculate the total power demand R based on the power required for data acquisition R1, the power required for data processing R2, the power required for computing power analysis R3 and the power required for power distribution R4, and obtain the total power demand R, and set R=R1+R2+R3+R4.

10. The on-line Libs laser real-time detection device according to claim 9, characterized in that: The second detection and tuning module performs a secondary computing power adjustment on the computing power demand optimization method according to the computing power optimization adjustment method, and the computing power optimization adjustment method includes: Step U01: compare the total power demand R with the remaining power R0, determine the demand for the total power demand R based on the comparison result, and perform secondary computing power adjustments on the data processing method and computing power demand optimization method based on the determination result, where: When R≤R0, the second detection and tuning module determines that the total power demand R is low, and does not perform secondary computing power adjustment on the data processing method and the computing power demand optimization method; When R>R0, the second detection and tuning module determines that the demand situation of the total power demand R is high, performs a secondary computing power adjustment on the computing power demand optimization method, sets the adjusted data acquisition frequency to Cf`, Cf`=Cf×(R-R0) / R0, and replaces the data acquisition frequency Cf with the adjusted data acquisition frequency Cf`, recalculates the data generation rate Vs, sets the adjusted preset parallel task volume to Bq0`, Bq0`=Bq0×(1-(R-R0) / R0), replaces the preset parallel task volume Bq0 with the adjusted preset parallel task volume Bq0`, and recomputes the parallel task volume Bq with the adjusted preset parallel task volume Bq0`; Step U02: compare the working temperature Wh with the preset working temperature Wh0, determine the state of the working temperature Wh based on the comparison result, and perform a secondary calculation power correction on the remaining power R0 based on the determination result, wherein: When Wh≤Wh0, the second detection and tuning module determines that the operating temperature Wh is in a low temperature state and does not perform a secondary computing power correction on the remaining power R0; When Wh>Wh0, the second detection and tuning module determines that the operating temperature Wh is in a high temperature state, performs a secondary computing power correction on the power reserve R0, sets the corrected power reserve to R0', R0'=R0×Wh / Wh0, replaces the power reserve R0 with the corrected power reserve R0', and re-compares the total power demand R with the corrected power reserve R0'; In step U03, the heat dissipation efficiency Vv is compared with the preset heat dissipation efficiency Vv0, and the heat dissipation degree of the heat dissipation efficiency Vv is judged according to the comparison result. The preset working temperature Wh0 is updated with a secondary computing power according to the judgment result, wherein: When Vv≤Vv0, the second detection and tuning module determines that the heat dissipation degree of the heat dissipation efficiency Vv is low, and does not perform a secondary computing power update on the preset operating temperature Wh0; When Vv>Vv0, the second detection and tuning module determines that the heat dissipation degree of the heat dissipation efficiency Vv is high, and does not perform a secondary computing power update on the preset working temperature Wh0. The updated preset working temperature is set to Wh0`, Wh0`=Wh0×Vv / Vv0, and the preset working temperature Wh0 is replaced with the updated preset working temperature Wh0`, and the working temperature Wh is re-compared with the updated preset working temperature Wh0`.

Images