Wireless intelligent material moisture detection method and device

Abstract

The application discloses a wireless intelligent material moisture detection method and device, relates to the technical field of material moisture detection, and comprises the following steps: using an ultrasonic wave to pre-detect tobacco in a tobacco box to obtain tobacco density distribution, generating a detection path according to the tobacco density distribution; acquiring label data on the tobacco box, obtaining batch data of the tobacco according to the label data, generating detection calibration parameters, and performing detection calibration by a detection device according to the detection calibration parameters; inserting the detection device into the tobacco box, releasing a mobile end head by a detection needle head of the detection device, and obtaining moisture distribution data of the tobacco in the tobacco box; acquiring temperature and humidity data and gas data in the tobacco box according to the detection device, generating a compressed data chain in combination with the moisture distribution data, and sending the compressed data chain to a cloud end; and the cloud end performs multi-dimensional fusion prediction to generate a comprehensive quality report of the tobacco. The application has the effect of improving the efficiency and accuracy of moisture detection in the tobacco.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of material moisture detection, and in particular to a wireless intelligent material moisture detection method and device. BACKGROUND

[0002] Tobacco is a kind of herbaceous plant, which can be processed into cigarettes for sale, and is a kind of raw material with high economic value and great importance. In the process of tobacco purchase and processing, the moisture content in tobacco needs to be controlled, on the one hand to ensure that the tobacco does not rot and damage during storage, and on the other hand to improve the subsequent processing quality.

[0003] In the prior art, when detecting the moisture in tobacco, the detection device is generally inserted into a tobacco box or a tobacco bag, and then the moisture content in the tobacco is detected by infrared absorption, time domain frequency domain reflection and other methods. The above methods are all sampling detection of tobacco, however, in the detection process, the density of tobacco is not uniform, resulting in different local detection results, so that the final result is inaccurate. At the same time, in the detection process, only the moisture content in the tobacco is referred to, and other factors affecting the detection process and results are ignored. In addition, the dielectric properties of different batches of tobacco are different, and the manual calibration of the equipment before each detection is time-consuming and laborious and has low efficiency. SUMMARY

[0004] The purpose of the present application is to provide a wireless intelligent material moisture detection method and device to solve the problems raised in the background art.

[0005] In a first aspect, the present application provides a wireless intelligent material moisture detection method, which comprises:

[0006] using ultrasonic waves to pre-detect the tobacco in the tobacco box to obtain ultrasonic data of the tobacco, obtaining the density distribution of the tobacco in the tobacco box according to the ultrasonic data, and generating a detection path according to the density distribution of the tobacco;

[0007] obtaining label data on the tobacco box, obtaining batch data of the tobacco according to the label data, generating detection calibration parameters according to the batch data, and sending the detection calibration parameters to a detection device, wherein the detection device performs detection calibration according to the detection calibration parameters;

[0008] inserting the detection device into the tobacco box, releasing a moving end of a detection needle of the detection device, and moving the moving end to detect according to the detection path to obtain moisture distribution data of the tobacco in the tobacco box;

[0009] According to the detection device, temperature and humidity data and gas data in the tobacco box are obtained. The moisture distribution data, temperature and humidity data and gas data are combined and preprocessed to generate a compressed data chain, which is then sent to the cloud.

[0010] The cloud platform performs multi-dimensional fusion prediction on the tobacco in the tobacco box based on the compressed data chain, generates a predicted risk of spoilage of the tobacco, and generates a comprehensive quality report of the tobacco based on the predicted risk of spoilage.

[0011] Preferably, the steps of using ultrasound to pre-detect the tobacco in the tobacco box to obtain ultrasonic data of the tobacco, obtaining the tobacco density distribution in the tobacco box based on the ultrasonic data, and generating a detection path based on the tobacco density distribution are as follows:

[0012] An ultrasonic transmitting / receiving array is arranged around the tobacco box to transmit ultrasonic waves to the tobacco in the tobacco box for pre-detection, thereby obtaining ultrasonic data of the tobacco.

[0013] The ultrasonic data is denoised and filtered, and the denoised and filtered ultrasonic data is time-domain compensated to obtain the target ultrasonic data.

[0014] The sound wave velocity, attenuation, and scattering intensity are extracted from the target ultrasonic data, and the tobacco density distribution in the tobacco box is obtained by inversion based on the sound wave velocity, attenuation, and scattering intensity.

[0015] Based on the tobacco density distribution, the tobacco box is spatially divided to obtain multiple tobacco blocks, and the density value of each tobacco block is labeled.

[0016] Two tobacco blocks are pre-selected as initial blocks, and the tobacco block connected to the initial blocks and with the closest density value is selected as the target block;

[0017] Obtain the distance value between two target blocks, determine whether the distance value exceeds a preset distance threshold, and if the distance value exceeds the distance threshold, reselect the target block;

[0018] Based on the target block, the selection is repeated to generate a detection path.

[0019] Preferably, the step of generating detection calibration parameters based on the batch data and sending the detection calibration parameters to the detection equipment, wherein the detection equipment performs detection calibration based on the detection calibration parameters, specifically includes:

[0020] Acquire historical test data and historical batch data, and generate a historical batch correspondence table by mapping the historical test data and the historical batch data one by one.

[0021] Based on the historical batch correspondence table, extract the historical batch characteristics of tobacco from the historical detection data corresponding to each batch;

[0022] Based on the characteristics of multiple historical batches, specific detection parameters are obtained for each of the historical batch characteristics;

[0023] Based on each of the aforementioned detection parameters, a calibration parameter set is generated, and detection calibration parameters are selected from the calibration parameter set according to the batch data.

[0024] The detection and calibration parameters are sent to the detection equipment, which then performs self-detection and calibration based on the parameters.

[0025] Preferably, the step of inserting the detection device into the tobacco box, releasing the moving end of the detection needle of the detection device, and moving the moving end according to the detection path to obtain the moisture distribution data of the tobacco in the tobacco box is as follows:

[0026] Insert the detection device into the tobacco box and extract the spatial positions of the two detection needles of the detection device in the tobacco box;

[0027] Based on the spatial position of the needle and the multiple tobacco blocks, a starting tobacco block is obtained, and the detection path is corrected according to the starting tobacco block to generate a target detection path.

[0028] Based on the target detection path, each of the two detection needles releases its moving end, and the two moving ends move in the tobacco box according to the target detection path, and collect the capacitance change data in the tobacco box in real time, and upload it to the detection device;

[0029] The capacitance change data is correlated with the tobacco blocks to obtain local moisture data. Multiple local moisture data are integrated to obtain the moisture distribution data of the tobacco in the tobacco box.

[0030] Preferably, before the step of each of the two detection needles releasing its moving end based on the target detection path, the method further includes:

[0031] A temperature probe is disposed between the two detection needles, the temperature probe being longer than the two detection needles;

[0032] When the detection device is inserted into the tobacco box, the temperature probe collects the temperature data in the tobacco box and compares the temperature data with the preset standard temperature data to obtain the temperature difference;

[0033] Based on the temperature difference, the temperature probe performs heating or cooling operations to bring the two detection needles into a standard temperature environment.

[0034] The two detection needles are subjected to calibrated testing in the standard temperature environment to obtain initial detection data, and detection compensation parameters that vary with temperature are generated based on the initial detection data.

[0035] Preferably, the step of the two mobile terminals moving within the tobacco box according to the target detection path and collecting real-time data on changes in capacitance within the tobacco box specifically includes:

[0036] The two mobile ends move in the tobacco box according to the target detection path, and collect temperature and humidity data, pressure data and initial capacitance change data in real time during the movement;

[0037] Extract the temperature and humidity timestamps, pressure timestamps, and capacity timestamps from the temperature and humidity data, the pressure data, and the initial capacity change data, respectively.

[0038] Align the temperature and humidity timestamps, the pressure timestamps, and the capacitance timestamps, and then concatenate the timetamp-aligned temperature and humidity data, pressure data, and initial capacitance change data to generate series capacitance change data.

[0039] After adding the detection compensation parameters to the series capacitance change data, internal data fusion is performed to generate capacitance change data.

[0040] Preferably, the step of preprocessing the moisture distribution data, temperature and humidity data, and gas data to generate a compressed data chain, and then sending the compressed data chain to the cloud, specifically includes:

[0041] Extract key data on moisture distribution, temperature and humidity, and gas from the moisture distribution data, temperature and humidity data, and gas data to generate a key dataset.

[0042] A key data bar chart is generated based on the key dataset, and an alarm red line is set on the key data bar chart to detect whether there are any data bars that reach or exceed the alarm red line.

[0043] If a data column is detected to exceed the alarm threshold, the data type of the data column is extracted, and a corresponding alarm is generated based on the data type and sent to the detection device to trigger an alarm.

[0044] The key dataset is compressed to generate a data header, and the moisture distribution data, temperature and humidity data, and gas data are integrated to generate a compressed feature vector;

[0045] The data header and the compressed feature vector are assembled to generate a compressed data chain, which is then sent to the cloud.

[0046] Preferably, the step of the cloud performing multi-dimensional fusion prediction on the tobacco in the tobacco box based on the compressed data chain to generate a predicted risk of spoilage of the tobacco, and generating a comprehensive quality report of the tobacco based on the predicted risk of spoilage, specifically includes:

[0047] The cloud platform performs data expansion and restoration on the compressed data chain to obtain a complete detection dataset.

[0048] Based on the gas data in the detection dataset, the air composition data and air chemistry data in the tobacco box are obtained;

[0049] Based on the air composition data and air chemistry data in the detection dataset, a deterioration risk value for tobacco in the tobacco box is generated;

[0050] Based on the temperature and humidity data in the detection dataset, the deterioration promoting value of the current temperature and humidity in the tobacco box on the tobacco deterioration is obtained;

[0051] Based on the moisture distribution data in the detection dataset, predict the deterioration distribution map of the tobacco in the current tobacco box;

[0052] By combining the deterioration distribution map, the deterioration risk value, and the deterioration promotion value, a predicted deterioration risk of tobacco is generated, and a comprehensive quality report of tobacco is generated based on the predicted deterioration risk.

[0053] Secondly, this application provides a wireless intelligent material moisture detection device, the device comprising:

[0054] Pre-detection module: used to pre-detect the tobacco in the tobacco box using ultrasound to obtain ultrasonic data of the tobacco, obtain the density distribution of the tobacco in the tobacco box based on the ultrasonic data, and generate a detection path based on the tobacco density distribution;

[0055] OTA calibration module: used to acquire label data on tobacco boxes, obtain batch data of tobacco based on the label data, generate detection calibration parameters based on the batch data, and send the detection calibration parameters to the detection equipment, which performs detection calibration based on the detection calibration parameters;

[0056] Moisture detection module: used to insert the detection device into the tobacco box, the detection needle of the detection device releases a moving end, the moving end moves to detect according to the detection path, and obtains the moisture distribution data of the tobacco in the tobacco box;

[0057] Data preprocessing module: used to acquire temperature and humidity data and gas data in the tobacco box according to the detection device, combine the moisture distribution data, temperature and humidity data and gas data to perform preprocessing to generate a compressed data chain, and send the compressed data chain to the cloud;

[0058] Risk prediction module: used by the cloud to perform multi-dimensional fusion prediction on the tobacco in the tobacco box based on the compressed data chain, generate the predicted risk of tobacco deterioration, and generate a comprehensive quality report of tobacco based on the predicted risk of deterioration.

[0059] In summary, this application includes at least one of the following beneficial technical effects:

[0060] Ultrasonic testing is used to pre-inspect the tobacco box, obtaining the tobacco density distribution within. A detection path is then generated based on this distribution. Batch data is obtained from the label data on the tobacco box. Historical testing data and historical batch data are acquired to determine the historical batch characteristics of each batch during testing. Specific testing parameters are generated based on these characteristics, followed by calibration parameters. The testing equipment is then remotely calibrated using these calibration parameters. The testing equipment, consisting of two detection probes and a temperature probe, is then inserted into the tobacco box. Each probe has a movable detection tip. After insertion, the temperature probe identifies and adjusts the current temperature, while the movable tips remain stationary. Initial testing is performed under standard temperature conditions, generating detection compensation parameters. The detection path is then adjusted based on the spatial position of the movable tips to obtain the target detection path. Moisture distribution data is obtained by correcting the detection results using the detection compensation parameters. Finally, temperature, humidity, and gas data from the tobacco box are collected, pre-processed, compressed, and sent to the cloud. The cloud platform predicts the deterioration of tobacco based on received data and generates a comprehensive quality report. This improves the efficiency and accuracy of moisture detection in tobacco. Attached Figure Description

[0061] Fig. 1 This is a flowchart illustrating the steps of a wireless intelligent material moisture detection method provided in an embodiment of this application;

[0062] Fig. 2 This is a block diagram of a wireless intelligent material moisture detection device provided in an embodiment of this application.

[0063] Explanation of reference numerals in the attached diagram: 1. Pre-detection module; 2. OTA calibration module; 3. Moisture detection module; 4. Data preprocessing module; 5. Risk prediction module. Detailed Implementation

[0064] The following combinationFigs. 1-2 This application will be described in further detail, but the embodiments of the present invention are not limited thereto.

[0065] The application embodiment discloses a wireless intelligent material moisture detection method and device.

[0066] In this embodiment, a wireless intelligent material moisture detection method is provided, the method comprising:

[0067] S100: Use ultrasound to pre-detect the tobacco in the tobacco box to obtain ultrasonic data of the tobacco, obtain the density distribution of the tobacco in the tobacco box based on the ultrasonic data, and generate a detection path based on the tobacco density distribution;

[0068] S200: Obtain label data from tobacco boxes, obtain batch data of tobacco based on label data, generate detection and calibration parameters based on batch data, and send the detection and calibration parameters to the detection equipment. The detection equipment performs detection and calibration based on the detection and calibration parameters.

[0069] S300: Insert the detection device into the tobacco box, release the moving end of the detection needle of the detection device, and move the moving end according to the detection path to obtain the moisture distribution data of the tobacco in the tobacco box;

[0070] S400: Based on the detection equipment, acquire temperature and humidity data and gas data in the tobacco box, combine moisture distribution data, temperature and humidity data and gas data for preprocessing to generate a compressed data chain, and send the compressed data chain to the cloud.

[0071] S500: The cloud performs multi-dimensional fusion prediction on the tobacco in the tobacco box based on the compressed data chain, generates the predicted risk of tobacco deterioration, and generates a comprehensive quality report of the tobacco based on the predicted risk of deterioration.

[0072] It should be noted that the above modules are only the basic steps of this embodiment. In the specific implementation process, some steps may be added, reduced or modified appropriately without affecting the overall implementation effect.

[0073] The steps involve using ultrasound to pre-detect the tobacco in a tobacco box to obtain ultrasonic data, determining the tobacco density distribution in the tobacco box based on the ultrasonic data, and generating a detection path based on the tobacco density distribution. Specifically:

[0074] An ultrasonic transmitting / receiving array is arranged around the tobacco box to transmit ultrasonic waves to the tobacco in the tobacco box for pre-detection, thereby obtaining ultrasonic data of the tobacco.

[0075] The ultrasonic data is denoised and filtered, and the denoised and filtered ultrasonic data is then time-domain compensated to obtain the target ultrasonic data.

[0076] The sound wave velocity, attenuation, and scattering intensity are extracted from the target ultrasonic data. The tobacco density distribution in the tobacco box is obtained by inversion based on the sound wave velocity, attenuation, and scattering intensity.

[0077] Based on the tobacco density distribution, the tobacco box is spatially divided into multiple tobacco blocks, and the density value of each tobacco block is labeled.

[0078] Two tobacco blocks are pre-selected as initial blocks, and the tobacco block connected to the initial blocks with the closest density value is selected as the target block.

[0079] Get the distance between two target blocks, and determine whether the distance exceeds a preset distance threshold. If the distance exceeds the distance threshold, reselect the target block.

[0080] Based on the target block, the selection is repeated to generate a detection path.

[0081] In application, taking a tobacco box from batch A of a tobacco warehouse in 2023 as an example, eight sets of ultrasonic transmitting / receiving arrays were arranged around the box. A 1MHz ultrasonic wave was used for scanning, acquiring raw ultrasonic data containing 1200 echo signals. After wavelet denoising and Butterworth filtering, the raw data was processed using a time-domain compensation algorithm to eliminate the 5μs signal delay caused by the wooden material of the box, obtaining the target ultrasonic data. The extracted compensated sound velocity was 1520m / s (normal tobacco should be 1480-1500m / s), the attenuation coefficient reached 3.2dB / cm (normal value 2.5dB / cm), and the scattering intensity distribution showed a high (0.45) on the east side and a low (0.32) on the west side. The tobacco density distribution map was reconstructed based on the acoustic inversion model. The 1.2m×0.8m×0.6m enclosure was divided into 480 tobacco blocks of 10cm³ each, with labeled density values ​​ranging from 380-420kg / m³ (normal density 400±15kg / m³). Two initial blocks, D12 and D35 (density 402kg / m³), were pre-selected. Neighboring blocks with a density difference of less than 5kg / m³ were connected using a neighborhood search algorithm. When the distance between D12 and E12 exceeded 15cm, a threshold alarm was triggered, and block E15 (density 398kg / m³) was selected to construct the path. The final generated S-shaped detection path contained 68 connection nodes, covering 85% of the enclosure area.

[0082] The steps involve generating testing and calibration parameters based on batch data, sending these parameters to the testing equipment, and then the testing equipment performing testing and calibration based on these parameters.

[0083] Acquire historical testing data and historical batch data, and generate a historical batch mapping table by matching the historical testing data and historical batch data one by one.

[0084] Based on the historical batch correspondence table, extract the historical batch characteristics of tobacco from the historical testing data corresponding to each batch;

[0085] Based on the characteristics of multiple historical batches, specific detection parameters are obtained for each historical batch.

[0086] A calibration parameter set is generated for each detection parameter, and the detection calibration parameter is selected from the calibration parameter set according to the batch data.

[0087] The testing and calibration parameters are sent to the testing equipment, which then performs self-testing and calibration based on these parameters.

[0088] In application, taking a tobacco box from batch A of 2023 in a tobacco warehouse as an example, the label of this batch A shows a production date of May 12, 2023, and a standard moisture content of 12.5%. The system retrieves historical data from the past three months to construct a batch correspondence table and finds that the tobacco in batch 202303 exhibits an abnormal dielectric constant (ε=28.5, normal ε=32±1.5) due to different raw material origins. Targeted calibration parameters are generated: the capacitance detection reference value is adjusted from 120pF to 115pF, and the signal amplification gain is increased by 8%. After the calibration parameters are wirelessly transmitted to the detection equipment, the equipment automatically corrects the bias voltage in the detection circuit, reducing the zero-point drift from ±0.3pF to ±0.1pF. The system also loads the batch-specific moisture-capacitance curve equation: C=0.85W²+12.3W+98.6 (W is the moisture percentage), replacing the standard equation C=0.78W²+11.5W+102.4.

[0089] The steps for inserting the detection device into the tobacco box, releasing the moving end of the detection needle, and moving the moving end according to the detection path to obtain the moisture distribution data of the tobacco in the tobacco box are as follows:

[0090] Insert the detection device into the tobacco box and extract the spatial position of the two detection needles of the detection device in the tobacco box;

[0091] Based on the spatial position of the needle and multiple tobacco blocks, a starting tobacco block is obtained. The detection path is then corrected according to the starting tobacco block to generate the target detection path.

[0092] Based on the target detection path, each of the two detection needles releases its moving end. The two moving ends move in the tobacco box according to the target detection path and collect the capacitance change data in the tobacco box in real time, which is then uploaded to the detection device.

[0093] By mapping the capacitance change data to tobacco blocks, local moisture data is obtained. By integrating multiple local moisture data, the moisture distribution data of tobacco in the tobacco box is obtained.

[0094] In application, taking a tobacco box from batch A of a tobacco warehouse in 2023 as an example, after the detection equipment was inserted into the box at a 45-degree angle, the two probes were positioned at three-dimensional coordinates (0.2, 0.3, 0.15) m and (0.5, 0.3, 0.15) m respectively. The path planning module optimized the detection path into a spiral progression based on the initial position, increasing the number of nodes from 68 to 75, achieving a coverage rate of 92%. The moving end moved along the path at a speed of 5 mm / s, collecting data every 0.5 seconds. A sudden change in capacitance was detected in tobacco block No. 3 (coordinates 0.3, 0.4, 0.2), jumping from the baseline value of 115 pF to 138 pF (corresponding to a moisture content of 16.7%), generating local abnormal data. A total of 1500 sets of capacitance data were acquired along the entire path. After spatial interpolation, the generated moisture distribution map showed a high-humidity area with a diameter of 25 cm and a moisture content of 18.5% in the northeast corner of the box.

[0095] Based on the target detection path, before the step of each of the two detection probes releasing its moving end, the following steps are also included:

[0096] A temperature probe is placed between the two detection needles, and the length of the temperature probe is longer than the two detection needles.

[0097] After the detection device is inserted into the tobacco box, the temperature probe collects the temperature data in the tobacco box and compares the temperature data with the preset standard temperature data to obtain the temperature difference;

[0098] Based on the temperature difference, the temperature probe performs heating or cooling operations to keep the two detection needles in a standard temperature environment.

[0099] Two detection needles are calibrated in a standard temperature environment to obtain initial detection data, and detection compensation parameters that vary with temperature are generated based on the initial detection data.

[0100] In application, taking a tobacco box from batch A of a tobacco warehouse in 2023 as an example, a temperature probe was set up in the middle of the detection process. The 15cm long temperature probe measured the temperature inside the box at 22℃ (standard 25℃), triggering the heating device to work for 3 minutes to raise the probe's ambient temperature to 25±0.5℃. Calibration detection showed that the reference capacitance value changed from 115pF at room temperature to 117pF, establishing a temperature compensation coefficient K=0.15pF / ℃. When a temperature fluctuation to 24.3℃ was detected in real-time during subsequent moving detection, compensation was automatically applied to the collected 125pF data: 125 + (25-24.3)*0.15 = 125.105pF. This compensation mechanism reduced the moisture detection error from ±0.8% to ±0.3%.

[0101] The steps for two mobile terminals to move within the tobacco box according to the target detection path and collect real-time data on changes in capacitance within the tobacco box are as follows:

[0102] Two mobile heads move within the tobacco box according to the target detection path, collecting temperature and humidity data, pressure data, and initial capacitance change data in real time during the movement.

[0103] Extract the temperature and humidity timestamps, pressure timestamps, and capacity timestamps from the temperature and humidity data, pressure data, and initial capacity change data, respectively.

[0104] Align the temperature and humidity timestamps, pressure timestamps, and capacitance timestamps, and then concatenate the timetamp-aligned temperature and humidity data, pressure data, and initial capacitance change data to generate series capacitance change data.

[0105] After adding the detection compensation parameters to the series capacitance change data, internal data fusion is performed to generate capacitance change data.

[0106] In application, taking tobacco boxes from batch A of a tobacco warehouse in 2023 as an example, the mobile terminal synchronously collected temperature 24.8℃, pressure 32.6kPa, and original capacitance 128pF at path node 47 (coordinates 0.6, 0.5, 0.3). The temperature and humidity data (timestamp 2023-07-12T14:23:45.200) at 0.2-second intervals were linearly interpolated and aligned with the capacitance data (timestamp 2023-07-12T14:23:45.205) using a timestamp alignment module. The compensated capacitance data of 129.2pF and the pressure data were used to construct a feature vector [129.2, 32.6, 24.8]. After PCA dimensionality reduction, a fused capacitance value of 130.5pF (corresponding to 15.2% moisture content) was generated. The variance of the data at this node decreased from 2.3 for individual detection to 0.7 after fusion.

[0107] The steps involved in preprocessing moisture distribution data, temperature and humidity data, and gas data to generate a compressed data chain, and then sending the compressed data chain to the cloud, are as follows:

[0108] Extract key data on moisture distribution, temperature and humidity, and gas from gas data to generate a key dataset.

[0109] A key data bar chart is generated based on the key dataset, and an alarm red line is set on the key data bar chart to detect whether there are data bars that reach or exceed the alarm red line.

[0110] If a data column is detected to exceed the alarm threshold, the data type of the data column is extracted, and a corresponding alarm is generated based on the data type and sent to the detection device to trigger an alarm.

[0111] The key dataset is compressed to generate a data header, and the moisture distribution data, temperature and humidity data, and gas data are integrated to generate a compressed feature vector.

[0112] The data header and compressed feature vector are assembled to generate a compressed data chain, which is then sent to the cloud.

[0113] In application, taking tobacco boxes from batch A of a tobacco warehouse in 2023 as an example, key values ​​were extracted from the moisture data: highest 18.5%, lowest 12.1%, variance 2.3. Key temperature and humidity data: average temperature 24.5℃ (standard deviation 0.8), relative humidity 72%. The CO2 concentration in the gas data reached 2300ppm (threshold 2000ppm). The constructed bar chart showed that the CO2 bar exceeded the red alarm line by 15%, triggering a warning signal. Data compression used Huffman coding, compressing the original 2.1MB data into a 356KB data chain, including a 128-byte header (recording the key value range) and a feature vector matrix. The compression ratio reached 6:1, and the wireless transmission time was reduced from 58 seconds to 9 seconds.

[0114] The cloud-based process involves multi-dimensional fusion and prediction of tobacco in tobacco boxes based on compressed data chains to generate predicted spoilage risks. A comprehensive quality report is then generated based on these predicted spoilage risks. The specific steps are as follows:

[0115] The compressed data chain is expanded and restored in the cloud to obtain a complete detection dataset.

[0116] Based on the gas data in the detection dataset, we obtain the air composition data and air chemistry data in the tobacco box;

[0117] The risk value of tobacco deterioration in the tobacco box is generated based on the air composition data and air chemistry data in the detection dataset.

[0118] Based on the temperature and humidity data in the detection dataset, the deterioration promoting value of the current temperature and humidity in the tobacco box on the tobacco deterioration is obtained;

[0119] Based on the moisture distribution data in the detection dataset, predict the deterioration distribution map of the tobacco in the current tobacco box;

[0120] By combining the deterioration distribution map, deterioration risk value, and deterioration promotion value, a predicted deterioration risk of tobacco is generated, and a comprehensive quality report of tobacco is generated based on the predicted deterioration risk.

[0121] In practice, taking tobacco boxes from batch A of a tobacco warehouse in 2023 as an example, after data reconstruction in the cloud, CO2 levels were found to be excessive and acetaldehyde content was 0.15 ppm (safe value <0.1 ppm), resulting in a spoilage risk index of 78 points (threshold 60). Temperature and humidity data yielded a spoilage promotion factor of 1.25 (formula: e^(0.05*(T-20)+0.03*(H-65))). A moisture distribution prediction model showed that high humidity areas (>16%) had an 85% probability of expansion within 48 hours. A comprehensive assessment generated a quality report: current safety level B, recommending dehumidification within 24 hours, with an estimated cost of 320 RMB, reducing the spoilage risk to 32 points. The report was stored on blockchain and synchronized to the warehouse management system.

[0122] This invention provides a wireless intelligent material moisture detection device, using any one of the wireless intelligent material moisture detection methods described above. The device includes the following:

[0123] Pre-detection module 1: Used to pre-detect the tobacco in the tobacco box using ultrasound to obtain ultrasonic data of the tobacco, obtain the density distribution of the tobacco in the tobacco box based on the ultrasonic data, and generate a detection path based on the tobacco density distribution;

[0124] OTA calibration module 2: Used to acquire label data on tobacco boxes, obtain batch data of tobacco based on label data, generate detection calibration parameters based on batch data, and send the detection calibration parameters to the detection equipment. The detection equipment performs detection calibration based on the detection calibration parameters.

[0125] Moisture detection module 3: Used to insert the detection device into the tobacco box. The detection needle of the detection device releases the moving end, and the moving end moves to detect according to the detection path to obtain the moisture distribution data of the tobacco in the tobacco box.

[0126] Data preprocessing module 4: It is used to obtain temperature and humidity data and gas data in the tobacco box according to the detection equipment, combine the moisture distribution data, temperature and humidity data and gas data to preprocess and generate a compressed data chain, and send the compressed data chain to the cloud.

[0127] Risk Prediction Module 5: Used in the cloud to perform multi-dimensional fusion prediction of tobacco in tobacco boxes based on compressed data links, generate predicted deterioration risk of tobacco, and generate a comprehensive quality report of tobacco based on the predicted deterioration risk.

[0128] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A wireless intelligent material moisture detection method, characterized in that, Includes the following steps: Ultrasonic waves are used to pre-detect the tobacco in the tobacco box to obtain ultrasonic data of the tobacco. The density distribution of the tobacco in the tobacco box is obtained based on the ultrasonic data, and a detection path is generated based on the tobacco density distribution. Obtain label data from tobacco boxes, obtain batch data of tobacco based on the label data, generate detection calibration parameters based on the batch data, and send the detection calibration parameters to the detection equipment. The detection equipment performs detection calibration based on the detection calibration parameters. The detection device is inserted into the tobacco box, the detection needle of the detection device releases the moving end, and the moving end moves to detect according to the detection path to obtain the moisture distribution data of the tobacco in the tobacco box; According to the detection device, temperature and humidity data and gas data in the tobacco box are obtained. The moisture distribution data, temperature and humidity data and gas data are combined and preprocessed to generate a compressed data chain, which is then sent to the cloud. The cloud platform performs multi-dimensional fusion prediction on the tobacco in the tobacco box based on the compressed data chain, generates a predicted risk of spoilage of the tobacco, and generates a comprehensive quality report of the tobacco based on the predicted risk of spoilage. The steps of using ultrasound to pre-detect tobacco in a tobacco box to obtain ultrasonic data, obtaining the tobacco density distribution in the tobacco box based on the ultrasonic data, and generating a detection path based on the tobacco density distribution are as follows: An ultrasonic transmitting / receiving array is arranged around the tobacco box to transmit ultrasonic waves to the tobacco in the tobacco box for pre-detection, thereby obtaining ultrasonic data of the tobacco. The ultrasonic data is denoised and filtered, and the denoised and filtered ultrasonic data is time-domain compensated to obtain the target ultrasonic data. The sound wave velocity, attenuation, and scattering intensity are extracted from the target ultrasonic data, and the tobacco density distribution in the tobacco box is obtained by inversion based on the sound wave velocity, attenuation, and scattering intensity. Based on the tobacco density distribution, the tobacco box is spatially divided to obtain multiple tobacco blocks, and the density value of each tobacco block is labeled. Two tobacco blocks are pre-selected as initial blocks, and the tobacco block connected to the initial blocks and with the closest density value is selected as the target block; Obtain the distance value between two target blocks, determine whether the distance value exceeds a preset distance threshold, and if the distance value exceeds the distance threshold, reselect the target block; Based on the target block, the selection is repeated to generate a detection path.

2. The wireless intelligent material moisture detection method according to claim 1, characterized in that, The steps of generating detection calibration parameters based on the batch data, sending the detection calibration parameters to the detection equipment, and the detection equipment performing detection calibration based on the detection calibration parameters are as follows: Acquire historical test data and historical batch data, and generate a historical batch correspondence table by mapping the historical test data and the historical batch data one by one. Based on the historical batch correspondence table, extract the historical batch characteristics of tobacco from the historical detection data corresponding to each batch; Based on the characteristics of multiple historical batches, specific detection parameters are obtained for each of the historical batch characteristics; Based on each of the aforementioned detection parameters, a calibration parameter set is generated, and detection calibration parameters are selected from the calibration parameter set according to the batch data. The detection and calibration parameters are sent to the detection equipment, which then performs self-detection and calibration based on the parameters.

3. The wireless intelligent material moisture detection method according to claim 2, characterized in that, The steps of inserting the detection device into the tobacco box, releasing the moving end of the detection needle of the detection device, and moving the moving end according to the detection path to obtain the moisture distribution data of the tobacco in the tobacco box are as follows: Insert the detection device into the tobacco box and extract the spatial positions of the two detection needles of the detection device in the tobacco box; Based on the spatial position of the needle and the multiple tobacco blocks, a starting tobacco block is obtained, and the detection path is corrected according to the starting tobacco block to generate a target detection path. Based on the target detection path, each of the two detection needles releases its moving end, and the two moving ends move in the tobacco box according to the target detection path, and collect the capacitance change data in the tobacco box in real time, and upload it to the detection device; The capacitance change data is correlated with the tobacco blocks to obtain local moisture data. Multiple local moisture data are integrated to obtain the moisture distribution data of the tobacco in the tobacco box.

4. The wireless intelligent material moisture detection method according to claim 3, characterized in that, Based on the target detection path, before the step of each of the two detection needles releasing its moving end, the method further includes: A temperature probe is disposed between the two detection needles, the temperature probe being longer than the two detection needles; When the detection device is inserted into the tobacco box, the temperature probe collects the temperature data in the tobacco box and compares the temperature data with the preset standard temperature data to obtain the temperature difference; Based on the temperature difference, the temperature probe performs heating or cooling operations to bring the two detection needles into a standard temperature environment. The two detection needles are subjected to calibrated testing in the standard temperature environment to obtain initial detection data, and detection compensation parameters that vary with temperature are generated based on the initial detection data.

5. The wireless intelligent material moisture detection method according to claim 4, characterized in that, The steps of having the two mobile terminals move within the tobacco box according to the target detection path and collect real-time data on changes in capacitance within the tobacco box are as follows: The two mobile ends move in the tobacco box according to the target detection path, and collect temperature and humidity data, pressure data and initial capacitance change data in real time during the movement; Extract the temperature and humidity timestamps, pressure timestamps, and capacity timestamps from the temperature and humidity data, the pressure data, and the initial capacity change data, respectively. Align the temperature and humidity timestamps, the pressure timestamps, and the capacitance timestamps, and then concatenate the timetamp-aligned temperature and humidity data, pressure data, and initial capacitance change data to generate series capacitance change data. After adding the detection compensation parameters to the series capacitance change data, internal data fusion is performed to generate capacitance change data.

6. The wireless intelligent material moisture detection method according to claim 5, characterized in that, The steps of preprocessing the moisture distribution data, temperature and humidity data, and gas data to generate a compressed data chain, and then sending the compressed data chain to the cloud, are as follows: Extract key data on moisture distribution, temperature and humidity, and gas from the moisture distribution data, temperature and humidity data, and gas data to generate a key dataset. A key data bar chart is generated based on the key dataset, and an alarm red line is set on the key data bar chart to detect whether there are any data bars that reach or exceed the alarm red line. If a data column is detected to exceed the alarm threshold, the data type of the data column is extracted, and a corresponding alarm is generated based on the data type and sent to the detection device to trigger an alarm. The key dataset is compressed to generate a data header, and the moisture distribution data, temperature and humidity data, and gas data are integrated to generate a compressed feature vector; The data header and the compressed feature vector are assembled to generate a compressed data chain, which is then sent to the cloud.

7. The wireless intelligent material moisture detection method according to claim 6, characterized in that, The steps of the cloud platform performing multi-dimensional fusion prediction on the tobacco in the tobacco box based on the compressed data chain, generating a predicted risk of spoilage, and generating a comprehensive quality report of the tobacco based on the predicted risk of spoilage are as follows: The cloud platform performs data expansion and restoration on the compressed data chain to obtain a complete detection dataset. Based on the gas data in the detection dataset, the air composition data and air chemistry data in the tobacco box are obtained; Based on the air composition data and air chemistry data in the detection dataset, a deterioration risk value for tobacco in the tobacco box is generated; Based on the temperature and humidity data in the detection dataset, the deterioration promoting value of the current temperature and humidity in the tobacco box on the tobacco deterioration is obtained; Based on the moisture distribution data in the detection dataset, predict the deterioration distribution map of the tobacco in the current tobacco box; By combining the deterioration distribution map, the deterioration risk value, and the deterioration promotion value, a predicted deterioration risk of tobacco is generated, and a comprehensive quality report of tobacco is generated based on the predicted deterioration risk.

8. A wireless intelligent material moisture detection device, wherein the device uses the wireless intelligent material moisture detection method as described in any one of claims 1-7, characterized in that, include: Pre-detection module: used to pre-detect the tobacco in the tobacco box using ultrasound to obtain ultrasonic data of the tobacco, obtain the density distribution of the tobacco in the tobacco box based on the ultrasonic data, and generate a detection path based on the tobacco density distribution; OTA calibration module: used to acquire label data on tobacco boxes, obtain batch data of tobacco based on the label data, generate detection calibration parameters based on the batch data, and send the detection calibration parameters to the detection equipment, which performs detection calibration based on the detection calibration parameters; Moisture detection module: used to insert the detection device into the tobacco box, the detection needle of the detection device releases a moving end, the moving end moves to detect according to the detection path, and obtains the moisture distribution data of the tobacco in the tobacco box; Data preprocessing module: used to acquire temperature and humidity data and gas data in the tobacco box according to the detection device, combine the moisture distribution data, temperature and humidity data and gas data to perform preprocessing to generate a compressed data chain, and send the compressed data chain to the cloud; Risk prediction module: used by the cloud to perform multi-dimensional fusion prediction on the tobacco in the tobacco box based on the compressed data chain, generate the predicted risk of tobacco deterioration, and generate a comprehensive quality report of tobacco based on the predicted risk of deterioration.

Images