File storage health degree monitoring method based on feature extraction

Abstract

The invention discloses an archive storage health degree monitoring method based on feature extraction, and relates to the technical field of archive management.The archive storage health degree monitoring method can accurately recognize trace feature gas released by paper acidification and embrittlement by constructing an archive damage feature odor library, so that non-contact early warning of the archive health degree is achieved, and the safety of archive management is improved. And the limitation of pure dependence on environment temperature and humidity traditionally is broken through. Meanwhile, in combination with gradient source searching logic of distributed sensing or mobile inspection, the system can automatically invert and lock three-dimensional space coordinates of the abnormal gas and directly position the abnormal gas to a specific damaged file box. Moreover, a physical isolation and adjacent file association monitoring mechanism is innovatively introduced, so that a cross pollution source of the acid gas can be rapidly cut off through a transfer device, the monitoring level of surrounding high-risk suspected objects can be automatically improved, and the robustness and the intelligent level of a monitoring result are remarkably improved.

Description

Technical Field

[0001] This invention belongs to the field of archives management technology, and specifically relates to a method for monitoring the health of archives storage based on feature extraction. Background Technology

[0002] Paper archives are primarily composed of cellulose and hemicellulose. During long-term storage, they inevitably undergo slow chemical degradation reactions due to environmental factors such as temperature, humidity, light, and their own acidity, releasing volatile organic compounds, such as acetic acid produced during paper hydrolysis and furfural generated during cellulose dehydration. Existing archival monitoring technologies mainly focus on macroscopic control of temperature and humidity in the storage space. This singular environmental monitoring method can only reflect external preservation conditions and cannot directly characterize the actual aging or deterioration state of the archival paper itself, making it difficult to detect early paper acidification. Furthermore, health checks on archival materials often rely on manual sampling or periodic handling, which is not only inefficient and has limited coverage, but frequent manual handling may also cause secondary physical damage to already fragile ancient books. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides a method for monitoring the health of archive storage based on feature extraction, thereby solving the aforementioned technical problems.

[0004] A method for monitoring the health of archive storage based on feature extraction includes the following: A database of odors representing damaged archives is pre-installed in the central control unit. This database contains gas fingerprint feature vectors corresponding to different types of archive damage. The gas sensing component collects environmental gas data in real time from the storage space, and extracts the current gas feature value after noise reduction processing of the gas data. The current gas characteristic value is compared with the archive damage characteristic odor database to calculate the matching degree; when the matching degree exceeds the preset threshold, it is determined that there is an abnormal archive in the storage space and the type of abnormal gas is determined. In response to the anomaly identification results, a source localization strategy is triggered. Based on the deployment method of the gas sensing component, the concentration spatial distribution gradient of the gas data is used to calculate the three-dimensional spatial coordinates of the abnormal gas release source in the storage space and lock the target file box. The dispatching and transporting device moves to the three-dimensional spatial coordinates to grab the target file box and transports it to the preset physical isolation area; The system obtains information about adjacent files at the original storage location of the target file box, marks files located at adjacent coordinates as high-risk suspected objects, and upgrades the monitoring level of the high-risk suspected objects.

[0005] Preferably, the gas fingerprint feature vector includes at least one or a combination of the following two gas indicators: Acetic acid concentration is used to characterize the degree of hydrolysis and acidification of hemicellulose in archival paper; Furfural concentration is used to characterize the degree of embrittlement caused by cellulose dehydration in archival paper.

[0006] Preferably, the deployment method of the gas sensing component and the corresponding data acquisition logic include at least one of the following; Distributed deployment: Several gas sensor nodes are fixedly distributed in different locations on the archive storage shelf, and a three-dimensional concentration field is constructed by the difference in readings between the nodes; Mobile deployment: Gas sensors are installed on transport devices or independent inspection mechanisms, and the concentration change trajectory is constructed by continuously sampling data during the movement of the transport device along a preset path.

[0007] Preferably, when the gas sensing component is deployed in a distributed manner, the source location strategy specifically includes the following steps: A three-dimensional discrete concentration field model of the storage space is constructed, and the real-time concentration readings of each sensor node are mapped to the grid points of the model. Concentration values ​​in non-grid point regions are estimated using spatial interpolation algorithms to generate a continuous concentration distribution surface; Calculate the gradient vector of the concentration distribution surface and search for local maxima along the gradient ascent direction; The geometric center coordinates corresponding to the local maximum point are determined as the three-dimensional spatial coordinates of the abnormal gas release source.

[0008] Preferably, when the gas sensing component is deployed in a mobile manner, the source location strategy specifically includes the following steps: The transfer device carries a gas sensor to perform a global inspection of all file storage shelves; When the gas characteristic value at the target location is detected to exceed the trigger threshold, the global inspection is interrupted and the precise positioning mode is entered. The transport device uses the target location as the origin and collects the gas concentration in different directions in the three-dimensional space where the target location is located. Calculate the directional derivative of the rate of change of gas concentration with the displacement vector of the transport device, and move the transport device in the direction of the fastest increase in concentration; Repeat the above calculation and movement steps until the detected concentration change rate changes from positive to negative, and lock the physical coordinates corresponding to the concentration peak point as the location of the target file box.

[0009] Preferably, the physical isolation zone is a closed chamber with an independent exhaust system and a negative pressure environment. After the target file box is moved to the physical isolation zone, the following steps are also included: The exhaust system of the physical isolation zone is automatically activated, and the exhaust volume is adjusted to keep the air pressure in the zone lower than that in the external storage space to prevent abnormal gas backflow. Within a physically isolated zone, continuous monitoring of the release of individual gases from the target archive boxes was conducted, and the rate of concentration increase per unit time was recorded to assess the degradation activity of the archives.

[0010] Preferably, the noise reduction processing of the gas data further includes temperature and humidity compensation processing of the gas data, specifically including the following steps: Call the preset temperature and humidity-sensitivity drift curve, which describes the baseline drift of the gas sensing component under different temperature and humidity environments; Based on the currently acquired temperature and humidity data, determine the corresponding correction coefficient from the drift curve; The original gas concentration data is weighted and corrected using the aforementioned correction coefficient to eliminate signal errors caused by fluctuations in ambient temperature and humidity.

[0011] Preferably, when upgrading the monitoring level for the high-risk suspected individuals, the following execution strategies are specifically included: If the gas sensing components are deployed in a distributed manner, the sampling frequency of sensor nodes in the area adjacent to the original location of the target file box will be increased, and the sampling interval will be shortened to the preset emergency sampling cycle. If the gas sensing component is deployed on a mobile basis, then in the subsequent inspection task planning of the mobile device, a fixed-point sampling instruction is inserted at the coordinates adjacent to the original position of the target file box to increase the dwell time for sampling in that area.

[0012] Preferably, the following steps are also included: An alarm is generated through a human-computer interaction interface. The alarm content includes the target file box number, the type of abnormal gas detected, the peak concentration, and a list of adjacent high-risk files that are affected. When monitoring data within the physical isolation zone shows that the gas release rate of the target file box exceeds the irreversible damage threshold, a paper deacidification or repair recommendation form is generated.

[0013] Preferably, the characteristic odor library also includes interfering gas characteristics of non-paper materials; If the current gas characteristic value has the highest matching degree with the interfering gas characteristic of non-paper materials, it is determined to be an abnormality that is not due to damage to archives. It does not trigger the scheduling target file box action and physical isolation steps, and generates environmental interference early warning information.

[0014] The beneficial effects of this invention are as follows: By constructing a database of characteristic odors from damaged archives, it can accurately identify trace characteristic gases released by paper acidification and embrittlement, thereby achieving non-contact early warning of archive health and overcoming the limitations of traditional reliance solely on environmental temperature and humidity. Simultaneously, by combining distributed sensing or mobile inspection with gradient source-finding logic, the system can automatically invert and lock the three-dimensional spatial coordinates of abnormal gases, directly locating the specific damaged archive box. Furthermore, it innovatively introduces a physical isolation and adjacent archive correlation monitoring mechanism, which can not only quickly cut off cross-contamination sources of acidic gases through a transport device, but also automatically upgrade the monitoring level of surrounding high-risk suspected objects, significantly improving the robustness and intelligence of the monitoring results. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 A flowchart illustrating the steps of a feature extraction-based method for monitoring the health of archive storage provided by this invention; Figure 2 A flowchart illustrating the steps of a distributed deployment method for monitoring the health of archive storage based on feature extraction, as provided by this invention. Figure 3 The flowchart illustrates the steps of mobile deployment of a feature extraction-based archive storage health monitoring method provided by this invention. Detailed Implementation

[0017] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," "counterclockwise," "axial," "radial," and "circumferential" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0018] The following disclosure provides many different embodiments or examples for implementing different structures of the present invention. To simplify the disclosure of the present invention, the components and arrangements of specific examples are described below. Of course, these are merely examples and are not intended to limit the present invention.

[0019] The embodiments of the invention will now be described in detail with reference to the accompanying drawings.

[0020] like Figure 1 As shown, a method for monitoring the health of archive storage based on feature extraction includes the following: A database of odors representing damaged archives is pre-installed in the central control unit. This database contains gas fingerprint feature vectors corresponding to different types of archive damage. The gas sensing component collects environmental gas data in real time from the storage space, and extracts the current gas feature value after noise reduction processing of the gas data. The current gas characteristic value is compared with the archive damage characteristic odor database to calculate the matching degree; when the matching degree exceeds the preset threshold, it is determined that there is an abnormal archive in the storage space and the type of abnormal gas is determined. In response to the anomaly identification results, a source localization strategy is triggered. Based on the deployment method of the gas sensing component, the concentration spatial distribution gradient of the gas data is used to calculate the three-dimensional spatial coordinates of the abnormal gas release source in the storage space and lock the target file box. The dispatching and transporting device moves to the three-dimensional spatial coordinates to grab the target file box and transports it to the preset physical isolation area; The system obtains information about adjacent files at the original storage location of the target file box, marks files located at adjacent coordinates as high-risk suspected objects, and upgrades the monitoring level of the high-risk suspected objects.

[0021] In practical implementation, gas volatilization experiments are simulated in a controlled laboratory to examine archives of different materials at different aging stages, such as acidification, embrittlement, and mold growth in materials like Xuan paper and wood pulp paper. This allows for the construction of a high-dimensional database of characteristic odors related to archive damage and the establishment of standard fingerprint vectors for marker products like acetic acid and furfural. Single-type sensors often exhibit cross-sensitivity, making it difficult to distinguish specific pollutant types. The gas fingerprint feature vector technology uses a sensor array instead of a single sensor. The array consists of several metal-oxide-semiconductor sensors with different doping materials or different sensitivity characteristics. When the array is exposed to an environment where acetic acid gas is produced by the hydrolysis of archive paper, each sensor will produce a change in resistance or voltage based on its own chemical properties. Arranging these discrete response values ​​in a fixed order constitutes a multi-dimensional vector.

[0022] In one implementation, a distributed array monitoring system is used: miniature gas sensor nodes are deployed on the back panel of each layer of the filing cabinet, and each node transmits data back in real time via a ZigBee or LoRa network; the central control unit uses a Kalman filter algorithm to filter out random environmental noise, extracts feature vectors, and compares them with data in the database; for example, once acetic acid is identified and the matching degree is >90%, the system uses a Kriging space interpolation algorithm to process the concentration data of each node, constructs a three-dimensional concentration field of the storage room, and calculates the geometric center of the concentration gradient convergence as the coordinates (x, y, z) of the anomaly source. In another implementation, a mobile robot is used for inspection: when a fixed sensor detects an anomaly in an area but cannot pinpoint the location, an AGV equipped with a high-sensitivity electronic nose is dispatched to enter the area; the AGV executes a gradient source-finding algorithm, calculates the derivative of the concentration change rate ΔC and the displacement vector ΔP in real time, and makes micro-steps in the direction of the steepest increase in concentration until the concentration peak point is confirmed, thus locking the specific target file box location. After locking the target, the robotic arm grabs the file box according to the coordinates and transports it to a physical isolation chamber with independent ventilation. Subsequently, the system queries the warehouse topology in the database, identifies the IDs of the file boxes adjacent to the original location (top, bottom, left, and right), marks their status as "high-risk," and instructs local sensors to increase the sampling frequency of this adjacent area from once per hour to once every ten minutes, while adjusting the alarm threshold for agile monitoring. This scheme follows the laws of mass diffusion and the principle of risk prevention. Acetic acid and furfural released during paper aging create a concentration gradient field centered on the source, which can be traced back to the source through array sensing and algorithmic inversion. Since the acid released from paper accelerates the aging of adjacent papers, physical isolation aims to cut off the autocatalytic effect of acidic gases, while adjacent marking is based on the pollutant penetration mechanism to prevent potential hidden damage.

[0023] More specifically, the gas fingerprint feature vector includes at least one or a combination of the following two gas indicators: Acetic acid concentration is used to characterize the degree of hydrolysis and acidification of hemicellulose in archival paper; Furfural concentration is used to characterize the degree of embrittlement caused by cellulose dehydration in archival paper.

[0024] In this scheme, acetic acid and furfural are selected as core gas indicators to construct a fingerprint feature vector. The principle is based on the release mechanism of molecular markers during the paper aging process: acetic acid is produced by the acidic hydrolysis of hemicellulose, and its concentration is negatively correlated with the pH value of the paper; furfural is a characteristic product formed by the dehydration of cellulose under acidic conditions, and its release rate is directly related to the degree of cellulose chain breakage.

[0025] like Figure 2 , 3 As shown, more specifically, the deployment method of the gas sensing component and the corresponding data acquisition logic include at least one of the following; Distributed deployment: Several gas sensor nodes are fixedly distributed in different locations on the archive storage shelf, and a three-dimensional concentration field is constructed by the difference in readings between the nodes; Mobile deployment: Gas sensors are installed on transport devices or independent inspection mechanisms, and the concentration change trajectory is constructed by continuously sampling data during the movement of the transport device along a preset path.

[0026] In one implementation of distributed deployment, four low-power metal-oxide-semiconductor sensor nodes are installed on each layer of a standard mobile shelving system, with a grid distribution of 1.2 meters between nodes. A self-organizing network is constructed using the ZigBee 3.0 protocol to collect acetic acid / furfural concentration data in real time. The central control system constructs a three-dimensional concentration field model using a cubic spline interpolation algorithm. When a local concentration gradient is detected to exceed 0.3 ppm / m, the coordinates of the gradient convergence point are calculated using the Gauss-Newton iterative method to achieve continuous monitoring in a static environment.

[0027] In one implementation of mobile deployment, an electrochemical sensor and a PID detector are integrated on top of a file transport AGV. A serpentine inspection path is preset along the centerline of the channel, with a moving speed of 0.25 m / s. Gas concentration and laser SLAM positioning data are collected synchronously every 0.4 seconds. When a sudden change in concentration is detected, such as ΔC / Δt > 0.1 ppm / s, the system automatically switches to gradient ascent mode, controlling the AGV to finely adjust in 0.1 m steps in the direction of concentration increase, and the peak position is determined by calculating the directional derivative.

[0028] More specifically, when the gas sensing component is deployed in a distributed manner, the source location strategy specifically includes the following steps: A three-dimensional discrete concentration field model of the storage space is constructed, and the real-time concentration readings of each sensor node are mapped to the grid points of the model. Concentration values ​​in non-grid point regions are estimated using spatial interpolation algorithms to generate a continuous concentration distribution surface; Calculate the gradient vector of the concentration distribution surface and search for local maxima along the gradient ascent direction; The geometric center coordinates corresponding to the local maximum point are determined as the three-dimensional spatial coordinates of the abnormal gas release source.

[0029] In one implementation, an inverse distance weighting method is used to construct a static concentration field. The system divides the storage space with a grid resolution of 1.5 meters. For any non-sampling point, its concentration estimate is obtained by a weighted average of the readings of the sensors in the neighborhood. The weights are inversely proportional to the square of the Euclidean distance, thus generating a smooth concentration scalar field. Subsequently, the gradient vector of each grid point in the field is calculated using the central difference method. The gradient ascent algorithm searches along the positive gradient direction with an iteration step size of 0.1 meters until the gradient magnitude approaches zero, locking the coordinates of the local maximum. Example 2 uses Kriging interpolation for large-span warehouses, introducing a semi-variogram to quantify spatial autocorrelation. While generating the concentration surface, the estimated variance is calculated, effectively suppressing edge effect errors caused by uneven sensor layout. Compared to existing technologies that can only indicate the approximate range through area alarms, this solution uses algorithms to fill the blind spots between sensors, solving the pain point of difficult precise positioning under sparse deployment conditions, and controlling the anomaly source locking error within ±15.

[0030] In this scheme, the inverse distance weighting method is preferred for warehouses with relatively static environments because of its high computational efficiency; in warehouses with complex airflow, the Kriging method is preferred to reduce errors by utilizing spatial statistical properties.

[0031] More specifically, when the gas sensing component is deployed in a mobile manner, the source location strategy specifically includes the following steps: The transfer device carries a gas sensor to perform a global inspection of all file storage shelves; When the gas characteristic value at the target location is detected to exceed the trigger threshold, the global inspection is interrupted and the precise positioning mode is entered. The transport device uses the target location as the origin and collects the gas concentration in different directions in the three-dimensional space where the target location is located. Calculate the directional derivative of the rate of change of gas concentration with the displacement vector of the transport device, and move the transport device in the direction of the fastest increase in concentration; Repeat the above calculation and movement steps until the detected concentration change rate changes from positive to negative, and lock the physical coordinates corresponding to the concentration peak point as the location of the target file box.

[0032] In practical implementation, the transfer device is equipped with a high-sensitivity electrochemical sensor array. First, it performs a full-coverage inspection of the warehouse along a preset path. When the acetic acid concentration is detected to exceed the threshold, the system immediately pauses the global task and starts a precise positioning process based on the current coordinates: the device moves 0.2 meters along the three orthogonal axes of X, Y, and Z to conduct exploratory sampling and construct a local concentration gradient field. By calculating the dot product of the concentration change rate and the displacement vector in each direction, the direction of movement with the most significant increase in concentration is determined. Then, it iteratively advances along this direction with smaller step sizes, monitoring the concentration change trend in real time. When two consecutive samples show that the concentration increment changes from a positive value to a negative value, it is determined that the peak point has been passed, and the coordinates of the previous position are used as the precise position of the target file box.

[0033] More specifically, the physical isolation zone is a closed chamber with an independent exhaust system and a negative pressure environment. After the target file box is moved to the physical isolation zone, the following steps are also included: The exhaust system of the physical isolation zone is automatically activated, and the exhaust volume is adjusted to keep the air pressure in the zone lower than that in the external storage space to prevent abnormal gas backflow. Within a physically isolated zone, continuous monitoring of the release of individual gases from the target archive boxes was conducted, and the rate of concentration increase per unit time was recorded to assess the degradation activity of the archives.

[0034] In practice, the physical isolation zone adopts a double-layer stainless steel sealed chamber structure. The inner layer is equipped with a high-efficiency activated carbon filter unit, and the outer layer is equipped with a variable frequency exhaust fan unit. When the target file box is moved in, the control system automatically starts the exhaust system. The differential pressure sensor monitors the pressure difference between the inside and outside of the chamber in real time and dynamically adjusts the fan speed to maintain a stable negative pressure environment of -50Pa inside the chamber, ensuring that any gas flow direction always permeates unidirectionally from the external clean area to the inside of the isolation zone. At the same time, the high-precision gas sensor built into the isolation zone collects acetic acid and furfural concentration data every 30 seconds. The slope of the concentration change per unit time is calculated by a linear regression algorithm to generate a degradation activity index. When the degradation activity index value exceeds the threshold, the emergency treatment process is triggered.

[0035] More specifically, the noise reduction processing of the gas data also includes temperature and humidity compensation processing of the gas data, specifically including the following steps: Call the preset temperature and humidity-sensitivity drift curve, which describes the baseline drift of the gas sensing component under different temperature and humidity environments; Based on the currently acquired temperature and humidity data, determine the corresponding correction coefficient from the drift curve; The original gas concentration data is weighted and corrected using the aforementioned correction coefficient to eliminate signal errors caused by fluctuations in ambient temperature and humidity.

[0036] In practical implementation, a temperature and humidity-sensitivity drift curve is first constructed through a controlled environment chamber experiment. This involves adjusting the temperature (10℃-40℃) and relative humidity (20%-90%) in a constant zero-gas or standard-concentration gas environment, continuously recording the resistance or voltage response values ​​of the sensor array. A binary high-order polynomial function f(T,H) is generated using the least squares method, or a three-dimensional lookup table is constructed as the compensation benchmark. In one implementation, embedded real-time correction is employed. Edge computing nodes read local temperature and humidity sensor values, substitute them into the experimentally derived polynomial function to calculate the current sensitivity correction coefficient K, and perform calculations on the original sampled values ​​to achieve signal calibration. In another implementation, cloud-based data cleaning is used. For historical stored data, a bilinear interpolation algorithm is used to match the corresponding baseline offset on the drift curve surface for post-processing noise reduction. Compared to existing technologies that rely solely on hardware circuits for coarse temperature compensation, this solution achieves fine calibration across the entire measurement range, reducing measurement errors caused by warehouse air conditioning operation or seasonal changes, and eliminating false positives due to environmental fluctuations.

[0037] More specifically, when upgrading the monitoring level for the high-risk suspected individuals, the following implementation strategies are included: If the gas sensing components are deployed in a distributed manner, the sampling frequency of sensor nodes in the area adjacent to the original location of the target file box will be increased, and the sampling interval will be shortened to the preset emergency sampling cycle. If the gas sensing component is deployed on a mobile basis, then in the subsequent inspection task planning of the mobile device, a fixed-point sampling instruction is inserted at the coordinates adjacent to the original position of the target file box to increase the dwell time for sampling in that area.

[0038] In specific implementation, in one distributed deployment implementation, when the system determines that a file box is a high-risk object, it automatically increases the sampling frequency of sensor nodes within a horizontal radius of 1.5 meters and vertically within two layers of that location from the usual 10 minutes / time to 30 seconds / time. At the same time, it activates the collaborative verification mode between adjacent nodes, eliminating sporadic errors of single-point sensors through cross-comparison of multi-node data. In one mobile deployment implementation, the scheduling algorithm reconstructs the inspection path, inserts 7 fixed-point sampling coordinates within a 0.8-meter radius around the original high-risk location, extends the dwell time at each point from the standard 8 seconds to 45 seconds, and performs enhanced re-inspection of the area every 2 hours in the following 24 hours.

[0039] More specifically, it also includes the following steps: An alarm is generated through a human-computer interaction interface. The alarm content includes the target file box number, the type of abnormal gas detected, the peak concentration, and a list of adjacent high-risk files that are affected. When monitoring data within the physical isolation zone shows that the gas release rate of the target file box exceeds the irreversible damage threshold, a paper deacidification or repair recommendation form is generated.

[0040] In practical implementation, the human-computer interaction system adopts a layered alarm architecture. When abnormal gas is detected, the central control unit automatically aggregates sensor network data, generates structured alarm information, and presents it on the monitoring screen in the form of a heat map overlay. The alarm content accurately includes the physical number of the target file box, the main abnormal gas component, the peak concentration data, and a list of adjacent high-risk files calculated based on the diffusion radius algorithm. The list is sorted by risk level and marked with the expected damage time window. At the same time, the system continuously analyzes the release rate change curve of the target file box in the physical isolation zone. For example, when the acetic acid release rate exceeds the irreversible damage threshold of 0.15 ppm / min for 30 consecutive minutes, it automatically calls the preset repair knowledge base, combines the historical data of paper pH and fiber strength test results, and generates a customized repair suggestion form containing parameters such as deacidification solution ratio, processing time, and drying temperature. The suggestion form is pushed to the operation terminal of the file restoration studio through an encrypted network.

[0041] More specifically, the characteristic odor library also includes interfering gas characteristics of non-paper materials; If the current gas characteristic value has the highest matching degree with the interfering gas characteristic of non-paper materials, it is determined to be an abnormality that is not due to damage to archives. It does not trigger the scheduling target file box action and physical isolation steps, and generates environmental interference early warning information.

[0042] In practice, the characteristic odor database pre-sets mass spectrometry characteristic spectra and concentration change patterns of 32 non-paper material interfering gases, including terpenes released by wood preservatives, aldehydes and ketones volatilized from metal cabinets, and halogenated hydrocarbons from cleaning agent residues. When the monitoring system collects abnormal gases, it uses a dynamic time warping algorithm to calculate the similarity between the current gas characteristics and all templates in the database. If the matching degree with the interfering gas template exceeds 92% and is more than 15 percentage points higher than the archive damage characteristic template, it is determined to be an environmental interference event. The system then terminates the subsequent archive scheduling and isolation process and generates environmental interference early warning information containing the type of interference source, intensity level, and suggested investigation area, which is pushed to the operation and maintenance terminal through the management platform.

[0043] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A method for monitoring the health of archive storage based on feature extraction, characterized in that, Includes the following: A database of odors representing damaged archives is pre-installed in the central control unit. This database contains gas fingerprint feature vectors corresponding to different types of archive damage. The gas sensing component collects environmental gas data in real time from the storage space, and extracts the current gas feature value after noise reduction processing of the gas data. The current gas characteristic value is compared with the archive damage characteristic odor database to calculate the matching degree; When the matching degree exceeds the preset threshold, it is determined that there is an abnormal file in the storage space and the type of abnormal gas is determined. In response to the anomaly identification results, a source localization strategy is triggered. Based on the deployment method of the gas sensing component, the concentration spatial distribution gradient of the gas data is used to calculate the three-dimensional spatial coordinates of the abnormal gas release source in the storage space and lock the target file box. The dispatching and transporting device moves to the three-dimensional spatial coordinates to grab the target file box and transports it to the preset physical isolation area; The system obtains information about adjacent files at the original storage location of the target file box, marks files located at adjacent coordinates as high-risk suspected objects, and upgrades the monitoring level of the high-risk suspected objects.

2. The method for monitoring the health of archive storage based on feature extraction according to claim 1, characterized in that, The gas fingerprint feature vector contains at least one or a combination of the following two gas indicators: Acetic acid concentration is used to characterize the degree of hydrolysis and acidification of hemicellulose in archival paper; Furfural concentration is used to characterize the degree of embrittlement caused by cellulose dehydration in archival paper.

3. The method for monitoring the health of archive storage based on feature extraction according to claim 1, characterized in that, The deployment method of the gas sensing component and the corresponding data acquisition logic include at least one of the following; Distributed deployment: Several gas sensor nodes are fixedly distributed in different locations on the archive storage shelf, and a three-dimensional concentration field is constructed by the difference in readings between the nodes; Mobile deployment: Gas sensors are installed on transport devices or independent inspection mechanisms, and the concentration change trajectory is constructed by continuously sampling data during the movement of the transport device along a preset path.

4. The method for monitoring the health of archive storage based on feature extraction according to claim 3, characterized in that, When the gas sensing component is deployed in a distributed manner, the source location strategy specifically includes the following steps: A three-dimensional discrete concentration field model of the storage space is constructed, and the real-time concentration readings of each sensor node are mapped to the grid points of the model. Concentration values ​​in non-grid point regions are estimated using spatial interpolation algorithms to generate a continuous concentration distribution surface; Calculate the gradient vector of the concentration distribution surface and search for local maxima along the gradient ascent direction; The geometric center coordinates corresponding to the local maximum point are determined as the three-dimensional spatial coordinates of the abnormal gas release source.

5. The method for monitoring the health of archive storage based on feature extraction according to claim 3, characterized in that, When the gas sensing component is deployed in a mobile manner, the source location strategy specifically includes the following steps: The transfer device carries a gas sensor to perform a global inspection of all file storage shelves; When the gas characteristic value at the target location is detected to exceed the trigger threshold, the global inspection is interrupted and the precise positioning mode is entered. The transport device uses the target location as the origin and collects the gas concentration in different directions in the three-dimensional space where the target location is located. Calculate the directional derivative of the rate of change of gas concentration with the displacement vector of the transport device, and move the transport device in the direction of the fastest increase in concentration; Repeat the above calculation and movement steps until the detected concentration change rate changes from positive to negative, and lock the physical coordinates corresponding to the concentration peak point as the location of the target file box.

6. The method for monitoring the health of archive storage based on feature extraction according to claim 1, characterized in that, The physical isolation zone is a closed chamber with an independent exhaust system and a negative pressure environment. After the target file box is moved to the physical isolation zone, the following steps are also included: The exhaust system of the physical isolation zone is automatically activated, and the exhaust volume is adjusted to keep the air pressure in the zone lower than that in the external storage space to prevent abnormal gas backflow. Within a physically isolated zone, continuous monitoring of the release of individual gases from the target archive boxes was conducted, and the rate of concentration increase per unit time was recorded to assess the degradation activity of the archives.

7. The method for monitoring the health of archive storage based on feature extraction according to claim 1, characterized in that, The noise reduction processing of the gas data also includes temperature and humidity compensation processing, specifically including the following steps: Call the preset temperature and humidity-sensitivity drift curve, which describes the baseline drift of the gas sensing component under different temperature and humidity environments; Based on the currently acquired temperature and humidity data, determine the corresponding correction coefficient from the drift curve; The original gas concentration data is weighted and corrected using the aforementioned correction coefficient to eliminate signal errors caused by fluctuations in ambient temperature and humidity.

8. The method for monitoring the health of archive storage based on feature extraction according to claim 3, characterized in that, When upgrading the monitoring level for the high-risk suspected individuals, the following specific implementation strategies are included: If the gas sensing components are deployed in a distributed manner, the sampling frequency of sensor nodes in the area adjacent to the original location of the target file box will be increased, and the sampling interval will be shortened to the preset emergency sampling cycle. If the gas sensing component is deployed on a mobile basis, then in the subsequent inspection task planning of the mobile device, a fixed-point sampling instruction is inserted at the coordinates adjacent to the original position of the target file box to increase the dwell time for sampling in that area.

9. The method for monitoring the health of archive storage based on feature extraction according to claim 1, characterized in that, It also includes the following steps: An alarm is generated through a human-computer interaction interface. The alarm content includes the target file box number, the type of abnormal gas detected, the peak concentration, and a list of adjacent high-risk files that are affected. When monitoring data within the physical isolation zone shows that the gas release rate of the target file box exceeds the irreversible damage threshold, a paper deacidification or repair recommendation form is generated.

10. The method for monitoring the health of archive storage based on feature extraction according to claim 1, characterized in that, The characteristic odor library also includes interfering gas characteristics from non-paper materials; If the current gas characteristic value has the highest matching degree with the interfering gas characteristic of non-paper materials, it is determined to be an abnormality that is not due to damage to archives. It does not trigger the scheduling target file box action and physical isolation steps, and generates environmental interference early warning information.

Images