A method, system, and storage medium for cargo hold item status detection based on multi-point measurement.

Abstract

This application relates to the technical field of cargo storage status monitoring, and discloses a method, system, and storage medium for cargo hold item status detection based on multi-point measurement. The method achieves cargo hold item status detection through multi-point measurement, setting up a first and second distance sensor and a human body sensor on the same height plane. The two distance sensors scan adjacent areas to generate distance data, and the algorithm calculates the change data; the human body sensor data in the overlapping area is used to determine the personnel activity status and generate human movement markers. The system, based on the unpositioned human movement markers, fuses and processes the change data to calculate a comprehensive change value; if the value exceeds a threshold, an item location anomaly alert is triggered. Through multi-sensor data fusion and dynamic calibration mechanisms, the method effectively distinguishes between personnel activity and item anomalies, significantly improving the accuracy and reliability of cargo hold item status monitoring.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of cargo storage state monitoring, in particular to a cargo hold article state detection method and system based on multi-point measurement and a storage medium. BACKGROUND

[0002] With the global emphasis on energy saving and emission reduction and sustainable development, the cargo ship industry is experiencing a critical stage of transformation from traditional fuel power to new energy power. The application of new energy power such as electric power drive and hydrogen energy drive not only helps to reduce carbon emissions and reduce dependence on fossil energy, but also complies with the development requirements of the International Maritime Organization for green shipping.

[0003] Compared with traditional power, the power response of new energy cargo ships is more rapid, and the dynamic changes of the ship are more sensitive during acceleration, deceleration and turning operations. Although this feature improves the control performance of the cargo ship, it also makes the attitude change of the cargo ship more complex and frequent during navigation. Whether facing complex sea conditions or executing fine control instructions, the dynamic performance of the ship body generated during the control process of the new energy cargo ship puts higher requirements on the stability of the cargo in the cargo hold.

[0004] When the cargo ship shakes due to control adjustment or sea state influence, the goods in the cargo hold are prone to displacement. Especially in long-distance transportation, small displacements can accumulate over a long period of time, which may cause the cargo to be scattered, the center of gravity to shift, and the like. This not only increases the risk of damage to the cargo, but also may affect the navigation safety and stability of the cargo ship, and even cause serious consequences such as cargo collapse and ship imbalance. Therefore, it is urgent to effectively detect the cargo hold article state for real-time monitoring and early warning. SUMMARY

[0005] In order to monitor and warn the cargo hold article state in real time, the present application provides a cargo hold article state detection method and system based on multi-point measurement and a storage medium.

[0006] In a first aspect, the present application provides a cargo hold article state detection method based on multi-point measurement, which adopts the following technical solution:

[0007] A cargo hold article state detection method based on multi-point measurement, based on a human body sensor and a first distance sensor and a second distance sensor arranged on the same height plane, the method comprising the following steps:

[0008] The first distance sensor scans and detects a first area to generate first distance data, and the second distance sensor scans and detects a second area to generate second distance data;

[0009] The first distance data is calculated into first change data according to a distance change algorithm based on the associated scanning angle, and the first change data and second change data are calculated based on the scanning angle and the first distance data;

[0010] There is an intersection region between the first region and the second region, and the human body sensor is located in the intersection region; human body sensing data of the human body sensor is acquired, and a human body action amplitude is calculated according to the human body sensing data;

[0011] If the human body action amplitude is greater than a preset reference amplitude, a human body movement flag is set; otherwise, the human body movement flag is reset;

[0012] Based on the human body movement flag that is not set, a comprehensive change value is calculated according to the first change data and the second change data;

[0013] If the comprehensive change value is greater than a preset reference comprehensive value, a cargo compartment article position abnormality prompt is performed.

[0014] By adopting the above technical solution, the distance data corresponding to the cargo compartment articles can be obtained based on the first distance sensor and the second distance sensor, the relative displacement between the cargo compartment articles and the distance sensor can be obtained according to the distance change algorithm based on the distance data, the human body sensor in the intersection region can identify the human body situation in the cargo compartment, the influence of the large human body action amplitude on the calculated distance data is prevented, and the accuracy of the calculation result is reduced. When the human body action amplitude is small, measurement is still performed, so that an accurate abnormality prompt can be obtained, and the cargo compartment article state is monitored and warned in real time.

[0015] Optionally, the step of setting the human body movement flag further includes the following sub-steps:

[0016] Based on the human body movement warning, a sum value of the first change data and the second change data is calculated as comprehensive change data;

[0017] The comprehensive change data is converted into a change graph through a graphical algorithm;

[0018] A graph matching value between the change graph and a preset graph template is calculated, and if the graph matching value is greater than a preset reference matching value, the human body movement flag is reset.

[0019] By adopting the above technical solution, when the human body action amplitude is small, the human body action shape is identified and matched, so that the human body movement flag can be reset when the human body is nearly in a stationary state, and the influence of the existence of the human body on the detection time is reduced.

[0020] Optionally, the step of setting the human body movement flag further includes the following sub-steps:

[0021] After resetting the human body movement marker, the changed pattern and the pattern template are fused to update the pattern template with a fusion result;

[0022] The reference comprehensive value is corrected according to the pattern matching value inverse coefficient, that is, the greater the pattern matching value, the smaller the reference comprehensive value, and the smaller the pattern matching value, the greater the reference comprehensive value.

[0023] By using the above technical solution, updating the pattern template can adapt the pattern template to the weak motion amplitude of the human body, reduce the deviation of subsequent calculation, and combine the correction of the reference comprehensive value to further improve the accuracy of the calculation result.

[0024] Optionally, there is an intersection region between the first region and the second region, and a calibration anchor point is arranged in the intersection region.

[0025] First calibration data is extracted from the first distance data, and second calibration data is extracted from the second distance data, and the first calibration data and the second calibration data are both associated with the calibration anchor point.

[0026] A first difference value between the first calibration data and preset first reference data is calculated, and a second difference value between the second calibration data and preset second reference data is calculated.

[0027] A deviation coefficient is calculated according to the first difference value and the second difference value.

[0028] If the deviation coefficient is greater than or equal to a preset reference coefficient, an anchor abnormality alarm prompt is performed.

[0029] By using the above technical solution, based on the calibration anchor point, the state of the distance sensor can be compared, the calibration data can feedback the sensing error of the distance sensor, the deviation coefficient can be calculated through the difference value, which can provide a calculation basis for the anchor abnormality alarm prompt, so as to detect the accuracy of the distance sensor itself, and to ensure the stability of the calculation result.

[0030] Optionally, the step of comparing the deviation coefficient further includes the following sub-steps:

[0031] If the anchor abnormality alarm prompt is not performed, the first distance data is corrected according to the first difference value inverse coefficient, and the second distance data is corrected according to the second difference value inverse coefficient.

[0032] The size of the intersection region is corrected according to the deviation coefficient inverse coefficient.

[0033] The calibration anchor point is reselected in the intersection region, and the volume of the calibration anchor point is associated with the deviation coefficient inverse coefficient.

[0034] By adopting the technical scheme, no anchor abnormality alarm is given, and the sensing error is low. At this time, the position data of the cargo in the cargo hold can be corrected according to the difference, so as to reduce the influence of the double errors caused by the sensing error and the cargo hold shaking on the accuracy of the detection result.

[0035] Optionally, the method further comprises the following steps:

[0036] The ratio of the first difference value and the second difference value is calculated as a deviation ratio value at equal periods;

[0037] If the deviation ratio value is outside a preset reference deviation range, anchor abnormality early warning is given.

[0038] By adopting the technical scheme, the deviation ratio value can be used as reference data between the two distance sensors. If the sensing errors of the two distance sensors are too different, anchor abnormality early warning is given, which means that the sensor may be damaged or the anchor point may have a position abnormality.

[0039] Optionally, the method further comprises the following steps:

[0040] A plurality of deviation ratio values are recorded in a latest set time period, and a ratio dispersion value of the plurality of deviation ratio values is calculated;

[0041] If the ratio dispersion value is greater than a preset reference dispersion value, anchor shaking alarm is given.

[0042] By adopting the technical scheme, the dispersion value of the deviation ratio value can reflect the change of the sensing error between the two distance sensors. If the dispersion value is large, anchor abnormality early warning is given, which means that the data of the distance sensor is abnormal for many times, the distance sensor is damaged, or the anchor point has a position abnormality.

[0043] Optionally, the method further comprises the following steps:

[0044] A plurality of deviation ratio values are recorded in a latest set time period to form a deviation curve;

[0045] A deviation frequency spectrum is obtained by performing fast Fourier calculation on the deviation curve;

[0046] An energy distribution frequency band of the deviation frequency spectrum is calculated;

[0047] If the energy distribution frequency band is in a preset first frequency band, anchor shaking alarm is given;

[0048] If the energy distribution frequency band is in a preset second frequency band, sensor shaking alarm is given;

[0049] If the energy distribution frequency band is located in a preset third frequency band, a sensor data anomaly alarm prompt is performed.

[0050] The frequency in the first frequency band < the frequency in the second frequency band < the frequency in the third frequency band.

[0051] By using the above technical solution, the frequency response of the sensor data is analyzed by performing frequency spectrum analysis on the deviation curve formed by the deviation ratio, so that the state of the distance sensor or the calibration anchor point can be matched in the frequency domain.

[0052] In a second aspect, the application provides a cargo compartment article state detection system based on multi-point measurement, which adopts the following technical solution:

[0053] A cargo compartment article state detection system based on multi-point measurement includes a processor, and the processor executes the steps of the cargo compartment article state detection method based on multi-point measurement according to any one of the above.

[0054] In a third aspect, the application provides a storage medium, which adopts the following technical solution:

[0055] A storage medium stores a program, and the program is executed by a processor to implement the steps of the cargo compartment article state detection method based on multi-point measurement according to any one of the above.

[0056] In summary, the application has at least one of the following beneficial technical effects: based on the first distance sensor and the second distance sensor, distance data corresponding to the cargo compartment article can be obtained, the distance sensor is anchored by the calibration anchor point as a reference, which helps to ensure the accuracy of the distance data. The calibration anchor point itself and the distance sensor itself can also be fault warned by the calibration anchor point; according to the distance data, the relative displacement between the cargo compartment article and the distance sensor can be obtained according to the distance change algorithm, the human body sensor in the intersection area can identify the human body condition in the cargo compartment, and it can adapt to the nearly stationary human body to prevent large human body action amplitude from affecting the calculated distance data and reducing the accuracy of the calculation result. When the human body action amplitude is small, measurement is still performed, so that accurate abnormal prompts can be obtained, and the cargo compartment article state can be monitored and warned in real time. BRIEF DESCRIPTION OF DRAWINGS

[0057] Figure 1 is a step diagram of the cargo compartment article state detection method based on multi-point measurement.

[0058] Figure 2 is a position diagram of the first distance sensor, the second distance sensor, and the human body sensor.

[0059] Figure 3is a step diagram of a detection method based on a calibration anchor point arranged in a cross region.

[0060] Reference signs: 1, human body sensor; 2, first distance sensor; 3, second distance sensor; 4, calibration anchor point. DETAILED DESCRIPTION

[0061] Embodiments of the present application are described in detail below, examples of which are shown in the accompanying drawings.

[0062] In the description of the present specification, the description of the terms "certain embodiments", "one embodiment", "some embodiments", "illustrative embodiments", "example", "specific example" or "some examples" means that the specific features, structures, materials or characteristics described in connection with the described embodiments or examples are included in at least one embodiment or example of the present application. In the present specification, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0063] The embodiments of the present application disclose a cargo hold article state detection method based on multi-point measurement, referring to Figure 1 and Figure 2 , based on a human body sensor and a first distance sensor and a second distance sensor arranged on the same height plane, the method comprises the following steps:

[0064] The first distance sensor adopts a laser radar to scan and detect the first region at a frequency of 10Hz, to generate first distance data containing distance information of each detection point; the second distance sensor synchronously scans the second region to generate second distance data. In order to facilitate calculation, the first region and the second region are symmetrical relative to the symmetry line of the cargo hold, and in other embodiments, they can also be in an asymmetrical relationship.

[0065] Based on the scanning angles of the first distance sensor and the second distance sensor, the Kalman filtering algorithm is applied to the first distance data to calculate the first change data, and the scanning angles and the first distance data are combined to calculate the second change data to reflect the distance change of the articles in the corresponding region.

[0066] There is a cross region between the first region and the second region, and the human body sensor is located in the cross region. The human body sensor is an infrared sensing device and can perform image analysis. The human body sensing data of the human body sensor is obtained through infrared sensing technology, and the data features are extracted by using a wavelet transform algorithm according to the human body sensing data to calculate the human body action amplitude.

[0067] When the human body action amplitude is greater than the preset reference amplitude, such as the moving distance per second exceeding 0.5 meters, the human body movement flag is set; otherwise, the human body movement flag is reset.

[0068] In the case that the human body movement flag is not set, the first change data and the second change data are weighted and fused to calculate a comprehensive change value. If the comprehensive change value is greater than a preset reference comprehensive value, it is determined that the cargo compartment article position is abnormal, and the cargo compartment article position abnormality is prompted through the sound and light alarm device.

[0069] The method constructs a comprehensive monitoring network with the aid of the first and second distance sensors, collects distance data between the cargo compartment articles and the sensors in real time, and uses an advanced distance change algorithm to accurately analyze the relative displacement of the articles. By capturing the human activity characteristics, the system accurately distinguishes between personnel operation and abnormal article movement. When a large human action is detected, the system automatically enters an anti-interference mode, temporarily optimizing the data collection strategy to avoid false positives caused by personnel movement; when a small human action is detected, the system still maintains high sensitivity monitoring to ensure the continuity and accuracy of data collection. Through this dynamic adaptive processing mechanism, the system can quickly and accurately identify the abnormal state of the cargo compartment articles, issue an early warning signal in a timely manner, and realize all-weather and high-precision monitoring of the cargo compartment article state.

[0070] The step of setting the human body movement flag further includes the following sub-steps:

[0071] When the human body movement flag is set to trigger the human body movement warning, the first change data calculated by the first distance sensor and the second change data calculated by the second distance sensor are summed to obtain comprehensive change data. For example, the first change data contains distance change values of 100 sampling points, and the second change data also contains distance change values of 100 sampling points. Adding the values of the corresponding sampling points, 100 values of the comprehensive change data are obtained.

[0072] The linear interpolation graphical algorithm is used to convert the comprehensive change data into a change graph. Specifically, the sampling point serial number is taken as the horizontal coordinate, and the value of the comprehensive change data is taken as the vertical coordinate. The discrete comprehensive change data points are connected into a continuous curve through linear interpolation, thereby forming a change graph.

[0073] A plurality of preset graphic templates are provided, each of which corresponds to a different human body action appearance feature. A graphic matching algorithm based on feature point matching is used to calculate a graphic matching value between the changed graphic and the preset graphic template. The specific calculation process is as follows: the contour feature points of the changed graphic and the graphic template are extracted, the Euclidean distance and the angle difference between the feature points are calculated, and a comprehensive matching score is obtained as the graphic matching value. If the graphic matching value is greater than a preset reference matching value, such as 0.8, it is determined that the current human body action appearance matches the preset graphic template of the static state or the small action state, and at this time the human body movement flag is reset.

[0074] The graphic template can be pre-collected and trained according to common human body static or small action postures, such as standing still, slowly lifting hands, etc. Through the above implementation, when the human body action amplitude is small, the recognition and matching mechanism of the human body action appearance can reset the human body movement flag in time when the human body is nearly static, effectively reducing the detection delay caused by the human body, and improving the real-time performance and efficiency of the cargo compartment article state detection.

[0075] The step of setting the human body movement flag further includes the following sub-steps:

[0076] After resetting the human body movement flag, the changed graphic and the graphic template are fused. The fusion method can specifically be to take the intermediate value of the changed graphic and the graphic template; and the graphic template is updated with the fusion result.

[0077] The reference comprehensive value is corrected according to the inverse coefficient of the graphic matching value. The larger the graphic matching value, the smaller the reference comprehensive value; the smaller the graphic matching value, the larger the reference comprehensive value. The correction coefficient γ=k / (graphic matching value+ε); wherein k is a scaling factor (default 0.8), and ε is a smoothing parameter (default 0.1) to prevent the denominator from being zero. The reference comprehensive value is updated according to the following formula: new reference comprehensive value=original reference comprehensive value×γ; and an adjustment boundary is set: 0.6≤γ≤1.5 to prevent over-correction.

[0078] The fusion method of taking the intermediate value enables the graphic template to quickly adapt to the weak action features of the human body. The inverse coefficient correction mechanism of the reference comprehensive value effectively compensates for environmental interference, reduces the false positive rate in the case of slight cargo shaking, and thus further improves the accuracy of the calculation result.

[0079] Reference Figure 2 and Figure 3, the first region and the second region have a cross region, and a calibration anchor point is arranged in the cross region. The calibration anchor point adopts a high-reflectivity metal sphere as the calibration anchor point, and a matte coating is plated on the surface to reduce diffuse reflection interference. Three calibration anchor points are uniformly distributed in the cross region, only one calibration anchor point is drawn in the drawing as an example, forming an equilateral triangle with a side length of 2m, and the center-to-center spacing error of each anchor point is ≤2mm. The anchor point is fixed to the cargo compartment wall by a magnetic support, and the installation height is the same as the sensor.

[0080] First calibration data is extracted from the first distance data, and second calibration data is extracted from the second distance data, both of which are associated with the calibration anchor point. A Hough transform algorithm is used to identify the circular features in the distance data, and the center coordinates are extracted as calibration data. The first and second distance data are time-stamped synchronized (error ≤1ms), and a bilinear interpolation method is used to compensate for the sampling time difference. The extracted calibration data is applied to a Kalman filter, with process noise covariance Q=0.01 and measurement noise covariance R=0.1.

[0081] The first difference between the first calibration data and the preset first reference data is calculated, and the second difference between the second calibration data and the preset second reference data is calculated. The first difference = |first calibration data-first reference data|; the second difference = |second calibration data-second reference data|.

[0082] The deviation coefficient is calculated according to the first difference and the second difference; the deviation coefficient = (first difference²+second difference²) / (first reference data×second reference data).

[0083] If the deviation coefficient is greater than or equal to the preset reference coefficient, an anchor abnormality alarm is prompted. The reference coefficient is dynamically adjusted according to the sensor accuracy level, and the default reference coefficient is 0.03.

[0084] The anchor abnormality alarm adopts a three-level early warning:

[0085] Yellow warning: 0.03≤deviation coefficient<0.05, slight deviation of the sensor;

[0086] Orange warning: 0.05≤deviation coefficient<0.08, manual review is required;

[0087] Red warning: deviation coefficient≥0.08, the system automatically suspends detection.

[0088] A complete sensor state monitoring system is constructed by setting a calibration anchor point in the intersection area of the two distance sensors. The system extracts calibration data related to the anchor point from the first and second distance data, calculates the difference between the data and the preset reference data, and then obtains the deviation coefficient. The deviation coefficient can accurately quantify the measurement error of the sensor. Once the preset reference coefficient is exceeded, the system will trigger an anchor abnormality alarm to timely feedback the accuracy problem of the sensor itself. This comparison mechanism based on the calibration anchor point effectively realizes real-time monitoring and error compensation of the distance sensor state, provides a stable and reliable data basis for cargo hold item state detection, and greatly improves the accuracy of the detection result.

[0089] The step of comparing the deviation coefficients further includes the following sub-steps:

[0090] If no anchor abnormality alarm is prompted, the first distance data is corrected according to the first difference inverse coefficient, and the second distance data is corrected according to the second difference inverse coefficient. The larger the first difference is, the smaller the first distance data is; the smaller the first difference is, the larger the first distance data is. The larger the second difference is, the smaller the second distance data is; the smaller the second difference is, the larger the second distance data is. Wherein, the first inverse coefficient = 1 / (1+|first difference|); the second inverse coefficient = 1 / (1+|second difference|). The first distance data after correction = the first distance data x the first inverse coefficient; the second distance data after correction = the second distance data x the second inverse coefficient.

[0091] The size of the intersection area is corrected according to the deviation coefficient inverse coefficient; the deviation inverse coefficient = 1 / (1+deviation coefficient); the initial area of the intersection area is S0, and the adjusted area S = S0 x the deviation inverse coefficient. The adjustment range is limited between 0.5S0 and 1.5S0.

[0092] The calibration anchor point in the intersection area is reselected by using the simulated annealing algorithm, and the volume of the calibration anchor point is related to the deviation coefficient inverse coefficient. The initial volume of the calibration anchor point is V0, and the volume is adjusted according to the deviation coefficient inverse coefficient, and the formula is: new anchor point volume = V0 x (1 / (1+deviation coefficient)); the volume adjustment range is limited between 0.3V0 and 1.8V0.

[0093] When the deviation coefficient is lower than the preset threshold value and the anchor abnormality alarm is not triggered, it indicates that the sensor is in a relatively stable low-error working state. At this time, the distance data is finely corrected by the inverse coefficients of the first difference and the second difference, which can effectively compensate for the inherent measurement deviation of the sensor. At the same time, the size of the intersection area is dynamically adjusted combined with the deviation coefficient inverse coefficient, and the volume and position of the anchor point are optimized for adaptive error compensation. The superimposed influence of the sensor error and the cargo hold shaking can be significantly weakened to ensure that accurate cargo hold item position data is obtained, thereby improving the reliability of the detection result.

[0094] The method further comprises the following steps:

[0095] The ratio of the first difference value and the second difference value is calculated periodically as a deviation ratio. The fixed calculation period is set as T=30 seconds, and the latest first difference value and second difference value data are extracted at the end of each period. The calculation is performed by using the formula deviation ratio=first difference value / second difference value, if the second difference value is 0, the previous period valid data is taken for calculation; if both data are 0, it is marked as invalid data and the current calculation is skipped. The sliding average filtering is applied to the calculated deviation ratio, and the window size is set as 5 to eliminate the influence of instantaneous fluctuations.

[0096] If the deviation ratio is located outside the preset reference deviation range, an anchor abnormality early warning is performed. The preset reference deviation range is [0.8, 1.2], which can be dynamically adjusted according to the sensor factory precision parameters. The anchor abnormality early warning also adopts a three-level early warning mechanism:

[0097] Yellow warning: the deviation ratio is located in [0.7, 0.8) U (1.2, 1.3], which indicates that there may be a slight sensor error difference;

[0098] Orange warning: the deviation ratio is located in [0.6, 0.7) U (1.3, 1.4), which suggests that manual inspection of the sensor or anchor is recommended;

[0099] Red warning: the deviation ratio is less than 0.6 or greater than 1.4, which immediately triggers a shutdown for maintenance prompt.

[0100] The deviation ratio is a key indicator for measuring the consistency of the cooperative work of the double distance sensors. By calculating the proportional relationship of the first and second difference values in real time, a joint monitoring model of sensor performance and anchor state is established. When the ratio exceeds the preset reference range, it indicates that the measurement data of the two sensors deviates significantly: it may be caused by the detection misalignment of a sensor due to hardware failure or environmental interference, or the displacement or loosening of the calibration anchor due to external force. Accordingly, the anchor abnormality early warning mechanism is triggered to quickly locate the potential risk point, realize accurate identification of sensor failure and anchor abnormality, and reduce the misjudgment of the cargo status caused by equipment hidden dangers.

[0101] The method further comprises the following steps:

[0102] A plurality of deviation ratios are recorded in the latest set time period, and a ratio dispersion value of the plurality of deviation ratios is calculated; a sliding time window mechanism is adopted, the set time period is 30 minutes, the sampling frequency is 1 time / minute, and 30 continuous deviation ratio data points are stored in the window. The ratio dispersion is the dispersion degree of the plurality of deviation ratios.

[0103] If the ratio discrete value is greater than the preset reference discrete value, an anchor shaking alarm is prompted. Based on the statistical characteristics of the sensor historical data, the reference discrete value threshold is determined by kernel density estimation, and the default threshold is 0.25.

[0104] Early warning classification:

[0105] Yellow warning: 0.25≤ratio discrete value<0.35, the system enters monitoring mode, and the data sampling frequency is increased to 2 times / minute;

[0106] Orange warning: 0.35≤ratio discrete value<0.45, trigger anchor position checking process, start standby calibration system;

[0107] Red warning: ratio discrete value≥0.45, immediately suspend detection service, and lock current sensor state data.

[0108] The same level of warning signal needs to last for 5 minutes to be officially triggered, avoiding false alarms caused by transient fluctuations.

[0109] The bias ratio discrete value is a key indicator for measuring the stability of the dual distance sensor, which quantifies the fluctuation degree of the bias ratio within a certain time window, and accurately describes the dynamic change characteristics of the sensor error. When the discrete value exceeds the preset threshold, it indicates that the measurement data of the two sensors fluctuates significantly, which may exist two fault modes: one, the sensor causes abnormal fluctuation of detection data due to hardware aging, circuit failure or environmental interference; the other, the calibration anchor is loosened or displaced due to external force, causing the sensor reference to change.

[0110] The method further comprises the following steps:

[0111] Record a plurality of bias ratios in the latest set time period to form a bias curve; a sliding time window mechanism is used, the set time period is the latest 60 minutes, the sampling frequency is 2 times / minute, 120 continuous bias ratio data points are stored in the window, and the bias curve is generated.

[0112] The bias curve is subjected to fast Fourier transform to calculate the bias frequency spectrum;

[0113] The energy distribution frequency band of the bias frequency spectrum is calculated, i.e. the distribution of most energy.

[0114] If the energy distribution frequency band is located in the preset first frequency band, an anchor shaking alarm is prompted;

[0115] If the energy distribution frequency band is located in the preset second frequency band, a sensor shaking alarm is prompted;

[0116] If the energy distribution frequency band is located in the preset third frequency band, a sensor data anomaly alarm is prompted;

[0117] Wherein, the frequency in the first frequency band < the frequency in the second frequency band < the frequency in the third frequency band.

[0118] The first frequency band (0.001Hz~0.01Hz): corresponding to low-frequency vibration, mainly caused by anchor point mechanical shaking;

[0119] The second frequency band (0.01Hz~0.1Hz): corresponding to medium-frequency vibration, related to the loosening of the sensor installation support;

[0120] The third frequency band (0.1Hz~1Hz): corresponding to high-frequency noise, mostly caused by abnormal internal circuit of the sensor.

[0121] When the energy ratio of the first frequency band is >60%, trigger the anchor shaking alarm;

[0122] When the energy ratio of the second frequency band is >50%, trigger the sensor shaking alarm;

[0123] When the energy ratio of the third frequency band is >40%, trigger the sensor data abnormal alarm.

[0124] Through the introduction of spectrum energy distribution feature analysis, fault feature frequency library and severity evaluation model, the fine identification and early warning of sensor and anchor point faults are realized.

[0125] The embodiment of the application further discloses a cargo compartment article state detection system based on multi-point measurement, comprising a processor, wherein the processor executes the steps of the cargo compartment article state detection method based on multi-point measurement as described in any one of the above.

[0126] The embodiment of the application further discloses a storage medium, wherein the storage medium stores a program, and the program is executed by a processor to realize the steps of the cargo compartment article state detection method based on multi-point measurement as described in any one of the above.

[0127] Although the embodiments of the application have been shown and described above, it can be understood that the above-mentioned embodiments are exemplary and cannot be understood as limiting the application, and those skilled in the art can make changes, modifications, replacements and variations to the above-mentioned embodiments within the scope of the application.

Claims

1. A cargo hold item status detection method based on multipoint measurement, characterized by, The method comprises the following steps based on a human body sensor and a first distance sensor and a second distance sensor arranged on the same height plane: The first distance sensor scans and detects a first area to generate first distance data, and the second distance sensor scans and detects a second area to generate second distance data; Based on the associated scanning angles, the first distance data is calculated according to a distance change algorithm to obtain first change data, and the first change data and second change data are calculated based on the scanning angles and the first distance data; There is an intersection area between the first area and the second area, and the human body sensor is located in the intersection area; human body sensing data of the human body sensor is obtained, and a human body action amplitude is calculated according to the human body sensing data; If the human body action amplitude is greater than a preset reference amplitude, a human body movement flag is set; otherwise, the human body movement flag is reset; Based on the human body movement flag that is not set, a comprehensive change value is calculated according to the first change data and the second change data; If the comprehensive change value is greater than a preset reference comprehensive value, a cargo compartment article position abnormality prompt is performed; In the step of setting the human body movement flag, the following sub-steps are further included: Based on the human body movement warning, a sum value of the first change data and the second change data is calculated as comprehensive change data; The comprehensive change data is converted into a change pattern through a graphical algorithm; A pattern matching value between the change pattern and a preset pattern template is calculated, if the pattern matching value is greater than a preset reference matching value, the human body movement flag is reset; each pattern template corresponds to different human body action contour characteristics; In the step of setting the human body movement flag, the following sub-steps are further included: After resetting the human body movement flag, the change pattern and the pattern template are fused, and the pattern template is updated with a fusion result; The reference comprehensive value is corrected according to a pattern matching value inverse coefficient, the greater the pattern matching value, the smaller the reference comprehensive value; the smaller the pattern matching value, the greater the reference comprehensive value, the correction coefficient γ=k / (pattern matching value+ε); wherein k is a scaling factor, and ε is a smoothing parameter to prevent the denominator from being zero.

2. The multipoint measurement-based cargo hold item state detection method according to claim 1, characterized by, There is an intersection area between the first area and the second area, and a calibration anchor point is arranged in the intersection area; First calibration data is extracted from the first distance data, and second calibration data is extracted from the second distance data, the first calibration data and the second calibration data are both associated with the calibration anchor point; A first difference value between the first calibration data and a preset first reference data is calculated, and a second difference value between the second calibration data and a preset second reference data is calculated; A deviation coefficient is calculated according to the first difference value and the second difference value; If the deviation coefficient is greater than or equal to a preset reference coefficient, an anchor abnormality alarm prompt is performed.

3. The multi-point measurement based cargo hold item status detection method according to claim 2, characterized in that, In the step of calculating the deviation coefficient, the following sub-steps are further included: If the anchor abnormality alarm is not performed, the first distance data is corrected according to a first difference inverse coefficient, and the second distance data is corrected according to a second difference inverse coefficient; wherein the first difference inverse coefficient = 1 / (1+|first difference|); the second difference inverse coefficient = 1 / (1+|second difference|); The size of the intersection region is corrected according to a deviation coefficient inverse coefficient; the deviation coefficient inverse coefficient = 1 / (1+deviation coefficient); The calibration anchor point is reselected in the intersection region, and the volume of the calibration anchor point is associated with the deviation coefficient inverse coefficient.

4. The multipoint measurement-based cargo hold item state detection method according to claim 2, characterized by, The method further comprises the following steps: The ratio of the first difference and the second difference is calculated as a deviation ratio at equal intervals; If the deviation ratio is outside a preset reference deviation range, an anchor abnormality warning is prompted.

5. The multipoint measurement-based cargo hold item state detection method according to claim 4, characterized by, The method further comprises the following steps: A plurality of deviation ratios are recorded within a latest set time period, and a ratio dispersion value of the plurality of deviation ratios is calculated; If the ratio dispersion value is greater than a preset reference dispersion value, an anchor swing alarm is prompted.

6. The multipoint measurement-based cargo hold item state detection method according to claim 4, characterized by, The method further comprises the following steps: A plurality of deviation ratios are recorded within a latest set time period to form a deviation curve; A fast Fourier calculation is performed on the deviation curve to obtain a deviation frequency spectrum; An energy distribution frequency band of the deviation frequency spectrum is calculated; If the energy distribution frequency band is in a preset first frequency band, an anchor swing alarm is prompted; If the energy distribution frequency band is in a preset second frequency band, a sensor swing alarm is prompted; If the energy distribution frequency band is in a preset third frequency band, a sensor data abnormality alarm is prompted; Wherein, the frequency in the first frequency band < the frequency in the second frequency band < the frequency in the third frequency band.

7. A multipoint measurement based cargo hold item status detection system, characterized in that, The processor executes the steps of the cargo hold item state detection method based on multi-point measurement as claimed in any one of claims 1-6.

8. A storage medium, characterized by The medium stores a program, and the program is executed by the processor to realize the steps of the cargo hold item state detection method based on multi-point measurement as claimed in any one of claims 1-6.

Images