An underground water sample sampling auxiliary and storage management intelligent integrated system and method

Abstract

The present application relates to the technical field of groundwater monitoring, and particularly relates to a groundwater sample sampling auxiliary and storage management intelligent integrated system and method. The method comprises the following steps: in response to a sampling task starting, controlling an integrated storage platform to adaptively expand, constructing a field operation microenvironment based on environmental parameters; identifying a material electronic identification to construct a real-time material mapping model, and checking with a standard task list; obtaining a sampling personnel operation instruction, matching sampling procedure data, and fusing with an operation scene to generate enhanced guidance information; collecting key node data for real-time evaluation, and if the evaluation is passed, automatically generating a record and outputting a label; after the task is completed, automatically storing, and based on the next point coordinate driving the mobile platform to autonomously navigate. The present application integrates environmental protection, material management, intelligent interaction and autonomous movement functions, realizes full-process intelligent closed-loop management and control of sampling operation, and effectively improves the standardization, efficiency and data traceability of field sampling.

Description

Technical Field

[0001] This invention relates to the field of groundwater detection technology, and in particular to an intelligent integrated system and method for assisting groundwater sample collection and management. Background Technology

[0002] Groundwater sampling is the most fundamental and crucial step in groundwater monitoring, and the quality of the sampling directly determines the accuracy and reliability of the monitoring data. Currently, the industry's main focus is on upgrading sampling equipment (such as submersible pumps) or standardizing sampling procedures, while there is little systematic improvement from the perspectives of improving the operating environment for sampling personnel, enhancing operational convenience, increasing sampling efficiency, and improving human comfort.

[0003] In actual field operations, groundwater sampling faces the following technical challenges:

[0004] (1) Inefficient material management: The consumables required for sampling (such as disposable gloves, disposable droppers, pH test strips, reagent bottles, labels, markers, filter membranes, etc.), testing equipment (such as water quality multi-parameter instruments, vacuum filters, etc.), and various record forms (such as well washing record forms, sampling record forms, sample preservation and inspection record forms, etc.) are numerous and fragmented. Existing technologies mostly use simple storage boxes for storage, which makes the loading and unloading process cumbersome, makes it difficult to find specific consumables on site, makes it easy to miss items, and makes it easy for items to be cross-contaminated due to chaotic storage.

[0005] (2) Harsh working environment: Field sampling often faces harsh weather conditions such as sun exposure, rain showers, and sandstorms. Existing technology lacks effective shielding and protection facilities. At the same time, there is a lack of ergonomic operating tables and chairs. Sampling personnel need to squat or bend over for a long time, which leads to physical discomfort and affects work efficiency and the standardization of sampling operations.

[0006] (3) Inconvenient mobile operation: Each time the sampling point is changed, all items need to be packed, transported and rearranged, which is time-consuming and labor-intensive, seriously affecting the continuity and efficiency of sampling work. In addition, the lack of a stable power supply at the field operation site makes it difficult to charge the equipment, which restricts the field application of modern testing equipment.

[0007] (4) Lagging information interaction: On-site data query, label printing and other operations rely on mobile phones or independent portable printers. The operation process is cumbersome and the mobile phones or printing equipment are easily contaminated by hand pollutants, resulting in low data flow efficiency. The real-time and accuracy of on-site recording and sample labeling are difficult to guarantee.

[0008] Therefore, how to achieve integrated material management, comfortable working environment, intelligent mobility, and smart information interaction in field groundwater sampling operations is a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0009] This invention overcomes the shortcomings of the prior art and provides a smart integrated system and method for groundwater sample sampling assistance and collection management.

[0010] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0011] The first aspect of this invention discloses a smart integrated method for assisting groundwater sample collection and management, comprising the following steps:

[0012] In response to the start of the sampling task, the integrated storage platform is controlled to perform adaptive deployment and dynamically construct a field operation microenvironment that meets the sampling standards based on real-time perceived environmental parameters.

[0013] Identify the electronic tags of materials in each zone of the storage platform, construct a real-time material mapping model for the current sampling point, and verify its conformity with the standard sampling task list;

[0014] The sampling personnel's work instructions are obtained, parsed, and matched with the corresponding sampling procedure data. The procedure data is then integrated with the current operation scenario through an interactive interface to generate enhanced information to guide the sampling execution.

[0015] Collect key node data during the sample processing, evaluate the key node data in real time, and automatically trigger the record generation and label output instructions associated with the corresponding node if the evaluation is passed.

[0016] After the task at the current sampling point is completed, the storage platform is controlled to perform automatic storage, and the mobile platform is driven to autonomously navigate to the target location based on the preset coordinates of the next point.

[0017] Furthermore, in response to the initiation of the sampling task, the integrated storage platform is controlled to perform adaptive deployment, dynamically constructing a field operation microenvironment that conforms to the sampling standards based on real-time perceived environmental parameters, specifically:

[0018] After receiving the sampling task start command, the system collects the temperature, light intensity and wind speed data of the current work site in real time and defines them as the raw environmental parameter set.

[0019] The temperature data in the original environmental parameter set is compared with the preset low temperature threshold and high temperature threshold, the light intensity data is compared with the preset strong light threshold, and the wind speed data is compared with the preset strong wind threshold.

[0020] If the light intensity exceeds the strong light threshold, then the light intensity is defined as the dominant environmental factor; if the wind speed exceeds the strong wind threshold, then the wind speed is defined as the dominant environmental factor; if the temperature is lower than the low temperature threshold or higher than the high temperature threshold, then the temperature is defined as the dominant environmental factor.

[0021] When multiple parameters exceed their corresponding thresholds at the same time, one parameter is selected as the dominant environmental factor from the parameters that exceed the thresholds according to a preset priority rule. The priority rule is ordered as follows: strong wind threshold, strong light threshold, low temperature threshold or high temperature threshold.

[0022] According to the dominant environmental factors, the corresponding microenvironment construction strategy is matched from the preset protection strategy library. The protection strategy library stores the correspondence between different dominant environmental factors and the extension angle of the sunshade and the posture of the side wind deflector.

[0023] Based on the microenvironment construction strategy, extension angle control commands for the sunshade and attitude adjustment commands for the side wind deflectors are generated.

[0024] The extension angle control command and attitude adjustment command are sent to the attitude adjustment mechanism of the storage platform to drive the sunshade and side wind deflector to perform coordinated unfolding action, forming a three-sided physical protective barrier around the sampling station.

[0025] The actual position feedback signal after the attitude adjustment mechanism performs the action is collected and compared with the target position data in the microenvironment construction strategy. Based on the comparison result, it is confirmed that the physical protective barrier has reached the preset airtightness and shielding standards, thereby completing the construction of a field operation microenvironment that meets the sampling specifications.

[0026] Furthermore, the electronic identifiers of materials within each zone of the storage platform are identified, a real-time material mapping model for the current sampling point is constructed, and its conformity is verified against the standard sampling task list. Specifically:

[0027] Activate the RFID reader array deployed in each partition of the storage platform, batch read the electronic tags attached to the materials in each partition, and obtain the unique feature code corresponding to each electronic tag;

[0028] The unique feature code is sequentially input into the embedded material parsing unit. Through the feature code-material mapping table built into the material parsing unit, the name, specifications, current quantity and functional partition attributes of each material are reversely parsed.

[0029] All the parsed materials are classified and collected according to the functional partition attributes, and a real-time material mapping model containing spatial location index is constructed by combining the physical location coordinates of each partition within the storage platform.

[0030] Retrieve a standard sampling task list pre-stored in the local memory. The standard sampling task list defines the baseline name, baseline specifications, baseline quantity, and preset storage partition information of the materials required for this sampling.

[0031] The material details of each partition in the real-time material mapping model are compared item by item with the material requirements of the corresponding partition in the standard sampling task list to identify the missing material items, redundant material items, and quantity deviation items in each partition relative to the list.

[0032] Based on the missing material items, redundant material items, and quantity deviation items identified in each partition, a compliance verification report of the material configuration at the current sampling point is generated, and the verification report is pushed to the interactive interface to guide the sampling personnel to supplement or adjust the materials.

[0033] Furthermore, the sampling personnel's work instructions are obtained, parsed, and matched with the corresponding sampling procedure data. This procedure data is then integrated with the current operational scenario through an interactive interface to generate enhanced information guiding the sampling execution. Specifically:

[0034] The original speech signal containing keywords of the sampling operation is acquired by the sampling personnel through a bone conduction microphone. Endpoint detection and semantic parsing are performed on the original speech signal to extract the target action intention and target sample object corresponding to the current operation.

[0035] The target action intent and target sample object are input into a pre-constructed sampling knowledge graph for vectorized retrieval and association matching. Multimodal sampling procedure data packages that match the current operation scenario are retrieved from the sampling knowledge graph. The multimodal sampling procedure data packages include text-based operation procedures and spatial location guidance identifiers.

[0036] Acquire real-time visual images of the current operation scene, perform instance segmentation and recognition on the real-time visual images, and identify the entity categories and spatial coordinates of sampling bottles, sampling tools and sample containers in the images;

[0037] The spatial location guide identifier in the multimodal sampling procedure data package is aligned with the identified entity spatial coordinates and subjected to perspective projection transformation to generate augmented reality rendering instructions bound to the entity in the current operation scene.

[0038] According to the augmented reality rendering instructions, the text-based operation procedures are overlaid on the corresponding entity on the interactive interface, and the spatial location guidance mark is dynamically projected onto the surface of the corresponding entity in the form of a virtual arrow, forming an augmented information screen to guide the sampling personnel to perform the next operation.

[0039] Furthermore, key node data during sample processing are collected, and the key node data is evaluated in real time. If the evaluation is successful, the record generation and label output instructions associated with the corresponding node are automatically triggered, specifically:

[0040] The sensor array deployed on the operating station and sample container collects real-time data on sample volume changes over time, protective agent drop acceleration rate, and dynamic fluctuation curves of water physicochemical parameters. By performing joint feature extraction on the time-series change data and dynamic fluctuation curves, a real-time process fingerprint characterizing the current sampling operation status is generated.

[0041] The real-time process fingerprint is input into the built-in sampling decision model, which is pre-set with a standard state machine and tolerance domain corresponding to the current sample type and sampling stage.

[0042] Calculate the Mahalanobis distance between the real-time process fingerprint and the standard state machine, and compare the Mahalanobis distance with the threshold of the tolerance domain to generate a process deviation score for the current node operation;

[0043] When the process deviation score falls within the tolerance range, the record generation instruction bound to the current sampling node is automatically triggered, and a time-stamped sampling electronic record containing the current sampling point, sample number and operator information is generated based on the key feature points in the real-time process fingerprint and the preset electronic record template.

[0044] Simultaneously, based on the sample number in the electronic sampling record, the label format data that matches the current sample container specifications and the current sampling stage is retrieved from the pre-stored label template library. The key feature points in the sample number, sampling time, and process deviation score are hashed to generate a label data stream containing a unique check code.

[0045] The data stream of the label to be printed is sent to the portable printer that is connected to the storage platform, driving the portable printer to perform the label output action.

[0046] Furthermore, after the current sampling point task is completed, the storage platform is controlled to perform automatic storage, and the mobile platform is driven to autonomously navigate to the target location based on the preset coordinates of the next point. Specifically:

[0047] The system receives the current sampling point operation completion signal triggered by the sampling personnel through the interactive interface, and retrieves the preset storage strategy library according to the operation completion signal to generate a global storage instruction that includes the material placement order of each zone and the reset posture of the sunshade and side wind deflector.

[0048] The global storage command is sent to the attitude adjustment mechanism and the partition storage drive unit to control the sunshade and the side wind deflector to perform a coordinated folding action. At the same time, the push rod module in each partition is driven to push the external materials into the corresponding storage compartment and lock them until the feedback signals of the door lock status sensors of each partition are all displayed as closed, confirming that the storage platform has been restored to the transport locking configuration.

[0049] After confirming that the storage platform is in the transport locking configuration, the navigation and positioning module that is connected to the mobile platform is activated, the coordinates of the next point are read from the pre-stored task list, and the coordinates of the next point are compared with the current real-time positioning coordinates to generate navigation path planning data containing the start and end points.

[0050] The navigation path planning data is sent to the drive controller of the mobile platform. The drive controller controls the walking motor to perform autonomous driving actions based on the navigation path planning data, so that the mobile platform carrying the storage platform can automatically drive to the next point coordinate.

[0051] Before obtaining the sampling personnel's work instructions, the process also includes collecting the sampling personnel's facial feature information through an interactive interface, comparing it with a pre-stored authorized personnel database for authentication, and activating the voice collection function only after successful authentication.

[0052] The enhanced information screen also includes dynamically adjusting the flashing frequency or color of the virtual arrow based on the changes in the liquid level of the sampling bottle in the real-time visual image, so as to prompt the sampling personnel to control the sampling flow rate.

[0053] The second aspect of this invention discloses a smart integrated system for groundwater sample sampling assistance and collection management, applicable to any of the smart integrated methods for groundwater sample sampling assistance and collection management, comprising:

[0054] An integrated storage platform includes a mobile base and a box set on the mobile base. The box is divided into multiple storage units for classifying and storing sampling materials. Each storage unit is equipped with a pull-out structure or a flip-top structure.

[0055] An ergonomic work surface system includes a work platform that can be folded up on the top of the housing, and a seat assembly that can be folded up on the side or bottom of the housing;

[0056] The comprehensive environmental protection system includes a retractable sunshade installed on the top of the enclosure and a detachable or foldable side wind deflector installed on the side of the enclosure. When the sunshade and the side wind deflector are unfolded, they form a three-sided surrounding protective work space.

[0057] The intelligent information interaction terminal includes a touch screen embedded in the operation panel, a voice acquisition module that is communicatively connected to the touch screen, and a printer module that is communicatively connected to the touch screen.

[0058] The intelligent mobility and energy management system includes a drive wheel assembly disposed within the mobile base, a navigation and positioning module communicatively connected to the drive wheel assembly, a solar power panel disposed on the top of the housing, and an energy storage battery pack disposed inside the housing and electrically connected to the solar power panel.

[0059] This invention addresses the technical deficiencies in the prior art and has the following beneficial effects:

[0060] This invention transforms harsh field environments into microenvironments that meet sampling standards through adaptive deployment and protection mechanisms, fundamentally solving the problem of meteorological interference with sample quality. Through a material mapping model and augmented reality guidance, it achieves precise traceability of sampling materials and zero-error guidance for operational procedures, improving the standardization and traceability of sampling. By automatically generating real-time fingerprint assessments and records of key node data, quality control is embedded in the sampling process, ensuring the authenticity and integrity of the original data. Autonomous storage and navigation enable unmanned sampling and relocation operations, significantly reducing the labor intensity and time costs of fieldwork. This invention transforms the traditional discrete operation mode reliant on human experience into a platform-based, standardized intelligent operation mode, enhancing the digitalization and intelligence level of groundwater environmental monitoring. Attached Figure Description

[0061] To more clearly illustrate the technical solutions in the embodiments of the present invention 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 the present invention. For those skilled in the art, other embodiments can be obtained from these drawings without creative effort.

[0062] Figure 1 Flowchart of a smart integration method for local wastewater sample collection and management;

[0063] Figure 2 Framework diagram of a smart integrated system for local groundwater sample collection and management. Detailed Implementation

[0064] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0065] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0066] like Figure 1 As shown, the first aspect of this invention discloses a smart integrated method for groundwater sample collection assistance and management, comprising the following steps:

[0067] In response to the start of the sampling task, the integrated storage platform is controlled to perform adaptive deployment and dynamically construct a field operation microenvironment that meets the sampling standards based on real-time perceived environmental parameters.

[0068] Identify the electronic tags of materials in each zone of the storage platform, construct a real-time material mapping model for the current sampling point, and verify its conformity with the standard sampling task list;

[0069] The sampling personnel's work instructions are obtained, parsed, and matched with the corresponding sampling procedure data. The procedure data is then integrated with the current operation scenario through an interactive interface to generate enhanced information to guide the sampling execution.

[0070] Collect key node data during the sample processing, evaluate the key node data in real time, and automatically trigger the record generation and label output instructions associated with the corresponding node if the evaluation is passed.

[0071] After the task at the current sampling point is completed, the storage platform is controlled to perform automatic storage, and the mobile platform is driven to autonomously navigate to the target location based on the preset coordinates of the next point.

[0072] Furthermore, in response to the initiation of the sampling task, the integrated storage platform is controlled to perform adaptive deployment, dynamically constructing a field operation microenvironment that conforms to the sampling standards based on real-time perceived environmental parameters, specifically:

[0073] After receiving the sampling task start command, the system collects the temperature, light intensity and wind speed data of the current work site in real time and defines them as the raw environmental parameter set.

[0074] The temperature data in the original environmental parameter set is compared with the preset low temperature threshold and high temperature threshold, the light intensity data is compared with the preset strong light threshold, and the wind speed data is compared with the preset strong wind threshold.

[0075] If the light intensity exceeds the strong light threshold, then the light intensity is defined as the dominant environmental factor; if the wind speed exceeds the strong wind threshold, then the wind speed is defined as the dominant environmental factor; if the temperature is lower than the low temperature threshold or higher than the high temperature threshold, then the temperature is defined as the dominant environmental factor.

[0076] When multiple parameters exceed their corresponding thresholds at the same time, one parameter is selected as the dominant environmental factor from the parameters that exceed the thresholds according to a preset priority rule. The priority rule is ordered as follows: strong wind threshold, strong light threshold, low temperature threshold or high temperature threshold.

[0077] According to the dominant environmental factors, the corresponding microenvironment construction strategy is matched from the preset protection strategy library. The protection strategy library stores the correspondence between different dominant environmental factors and the extension angle of the sunshade and the posture of the side wind deflector.

[0078] Based on the microenvironment construction strategy, extension angle control commands for the sunshade and attitude adjustment commands for the side wind deflectors are generated.

[0079] The extension angle control command and attitude adjustment command are sent to the attitude adjustment mechanism of the storage platform to drive the sunshade and side wind deflector to perform coordinated unfolding action, forming a three-sided physical protective barrier around the sampling station.

[0080] The actual position feedback signal after the attitude adjustment mechanism performs the action is collected and compared with the target position data in the microenvironment construction strategy. Based on the comparison result, it is confirmed that the physical protective barrier has reached the preset airtightness and shielding standards, thereby completing the construction of a field operation microenvironment that meets the sampling specifications.

[0081] Specifically, after acquiring the original environmental parameter set including temperature, light intensity, and wind speed, the built-in data parsing unit fuses these parameters to identify the dominant environmental factors in the current scene. For example, when the wind speed exceeds a preset high wind threshold, wind speed is defined as the dominant factor; conversely, when the light intensity may affect sample stability or the operator's line of sight (i.e., the light intensity exceeds a preset strong light threshold), light intensity becomes the dominant factor.

[0082] After determining the dominant factors, they are input into a pre-defined protection strategy library for matching. It should be noted that the protection strategy library is a pre-generated rule base based on fluid dynamics and ergonomic simulations. The library stores the correspondence between different dominant environmental factors, their magnitude ranges, and physical barrier configurations. For example, for the condition "northwest wind speed 5 m / s," the matching microenvironment construction strategy in the library is "the south side of the sunshade extends to its maximum angle, and the west windbreak rises 45°"; for the condition "light intensity 100,000 lux," the matching strategy is "the sunshade is fully extended, and only the bottom anti-glare layer of the side windbreak rises."

[0083] Based on the matched strategy, specific execution parameters are generated and handed over to the attitude adjustment mechanism to complete the physical construction. By introducing decision-making logic based on hierarchical matching of dominant factors, this scheme makes the response of the environmental protection system more targeted and scientific, avoiding problems such as energy waste or obstruction of vision caused by the full baffle raising.

[0084] Furthermore, the electronic identifiers of materials within each zone of the storage platform are identified, a real-time material mapping model for the current sampling point is constructed, and its conformity is verified against the standard sampling task list. Specifically:

[0085] Activate the RFID reader array deployed in each partition of the storage platform, batch read the electronic tags attached to the materials in each partition, and obtain the unique feature code corresponding to each electronic tag;

[0086] The unique feature code is sequentially input into the embedded material parsing unit. Through the feature code-material mapping table built into the material parsing unit, the name, specifications, current quantity and functional partition attributes of each material are reversely parsed.

[0087] All the parsed materials are classified and collected according to the functional partition attributes, and a real-time material mapping model containing spatial location index is constructed by combining the physical location coordinates of each partition within the storage platform.

[0088] Retrieve a standard sampling task list pre-stored in the local memory. The standard sampling task list defines the baseline name, baseline specifications, baseline quantity, and preset storage partition information of the materials required for this sampling.

[0089] The material details of each partition in the real-time material mapping model are compared item by item with the material requirements of the corresponding partition in the standard sampling task list to identify the missing material items, redundant material items, and quantity deviation items in each partition relative to the list.

[0090] Based on the missing material items, redundant material items, and quantity deviation items identified in each partition, a compliance verification report of the material configuration at the current sampling point is generated, and the verification report is pushed to the interactive interface to guide the sampling personnel to supplement or adjust the materials.

[0091] Specifically, a network scanning system is used by deploying RFID reader arrays within each compartment of the storage platform. It's important to note that this array design overcomes potential signal blind spots that may exist with a single reader within the metal enclosure, ensuring accurate and batch reading of anti-metal interference electronic tags attached to materials in each compartment, thereby obtaining the unique feature code stored in each tag. After obtaining the feature code, it is sent to an embedded material analysis unit. This unit pre-stores a feature code-material mapping table, which is either pre-entered by sampling personnel or synchronously generated by the system backend. This table establishes the correspondence between electronic tag feature codes and specific material attributes (such as product name and specifications). By looking up the table and performing reverse analysis, the tedious feature codes can be transformed into structured material information, clarifying the functional zoning attributes assigned to the material during the initial storage design.

[0092] Next, combining the physical coordinates of each partition within the container, a real-time material mapping model is constructed. This model not only records what materials are available but also where they "should be" and "are currently located." Then, a pre-stored standard sampling task list is retrieved. This list includes not only material requirements but also pre-defined storage partition information based on the sampling route design. By comparing the real-time model with the task list partition by partition, three types of deviations can be accurately identified: first, materials that should be present in a partition are missing (missing items); second, materials that should not be present in a partition appear (redundant items); and third, the quantity of materials is insufficient or excessive (quantity deviation). The resulting verification report provides intuitive guidance for sampling personnel to make precise adjustments before departure or transfer, thereby preventing sampling interruptions or violations due to material errors or omissions from the outset.

[0093] Furthermore, the sampling personnel's work instructions are obtained, parsed, and matched with the corresponding sampling procedure data. This procedure data is then integrated with the current operational scenario through an interactive interface to generate enhanced information guiding the sampling execution. Specifically:

[0094] The original speech signal containing keywords of the sampling operation is acquired by the sampling personnel through a bone conduction microphone. Endpoint detection and semantic parsing are performed on the original speech signal to extract the target action intention and target sample object corresponding to the current operation.

[0095] The target action intent and target sample object are input into a pre-constructed sampling knowledge graph for vectorized retrieval and association matching. Multimodal sampling procedure data packages that match the current operation scenario are retrieved from the sampling knowledge graph. The multimodal sampling procedure data packages include text-based operation procedures and spatial location guidance identifiers.

[0096] Acquire real-time visual images of the current operation scene, perform instance segmentation and recognition on the real-time visual images, and identify the entity categories and spatial coordinates of sampling bottles, sampling tools and sample containers in the images;

[0097] The spatial location guide identifier in the multimodal sampling procedure data package is aligned with the identified entity spatial coordinates and subjected to perspective projection transformation to generate augmented reality rendering instructions bound to the entity in the current operation scene.

[0098] According to the augmented reality rendering instructions, the text-based operation procedures are overlaid on the corresponding entity on the interactive interface, and the spatial location guidance mark is dynamically projected onto the surface of the corresponding entity in the form of a virtual arrow, forming an augmented information screen to guide the sampling personnel to perform the next operation.

[0099] Before obtaining the sampling personnel's work instructions, the process also includes collecting the sampling personnel's facial feature information through an interactive interface, comparing it with a pre-stored authorized personnel database for authentication, and activating the voice collection function only after successful authentication.

[0100] It should be noted that, for the core aspect of human-computer interaction and operation guidance, this solution adopts a multimodal fusion augmented reality interaction mechanism to solve the practical problems of "interference with operation due to consulting manuals" and "hand contamination of interactive devices" in traditional sampling operations.

[0101] First, before the interactive process begins, sampling personnel must complete facial feature recognition through the interface, comparing their facial features with a pre-stored database of authorized personnel. Only after successful authentication can the subsequent voice collection function be activated. The rationale behind this design is that field sampling operations often involve the handling of hazardous chemicals (such as sample preservatives). Identity authentication ensures that operators possess the appropriate qualifications and also prevents unauthorized personnel from accidentally touching or misoperating the equipment.

[0102] During the certification and work instruction acquisition phase, voice signals are collected using a bone conduction microphone. The advantage of bone conduction microphones lies in their strong resistance to environmental noise, enabling them to clearly pick up the voice instructions of sampling personnel even in environments with sandstorms or equipment operating noise. After acquiring the raw voice signal, endpoint detection and semantic parsing are performed to remove environmental noise and extract the core semantic information, namely "what to do" (target action intention, such as "add protective agent") and "what to do" (target sample object, such as "brown sample bottle").

[0103] After obtaining the intent and object, they are input into a pre-constructed sampling knowledge graph for processing. Unlike traditional tree-structured directories or keyword matching libraries, this sampling knowledge graph stores various entities in the sampling procedure (such as sample type, container specifications, and preservative types) and their relationships (e.g., "brown bottle for organic matter detection," "nitric acid added to pH < 2"). Through vectorized retrieval and association matching, multimodal sampling procedure data packages highly relevant to the current operational scenario can be retrieved. These data packages not only contain text-based operational procedures (e.g., "add 3 drops of nitric acid") but also spatial location guidance markers (e.g., "the dropper nozzle should be aligned with the center of the bottle opening").

[0104] At the same time, real-time visual images of the current operating scene are captured by cameras deployed above the operating station, and instance segmentation algorithms are used to identify the images in order to accurately segment entities such as sampling bottles, droppers, and preservative bottles from complex backgrounds and determine their spatial coordinates in the image coordinate system.

[0105] At this point, two sets of data are simultaneously available: one is the abstract "procedural data" from the knowledge graph and its accompanying "ideal location markers"; the other is the concrete "entity location coordinates" from visual recognition. To achieve a "virtual-real fusion" guidance effect, coordinate system alignment and perspective projection transformation are required. Simply put, this involves mapping the abstract "guidance markers" in the knowledge graph to the precise location of the corresponding entity in the current real-time image through mathematical transformation. This process ensures that the virtual arrow can accurately "attach" to the real dropper, rather than floating in mid-air.

[0106] Finally, based on the augmented reality rendering instructions generated after the transformation, the final presentation is achieved on the interactive interface. On one hand, text-based operation procedures are displayed as labels next to the corresponding entities, such as displaying the text "Nitric acid needs to be added" next to the brown bottle; on the other hand, spatial location guidance marks are dynamically projected onto the entity surface in the form of virtual arrows, such as generating a pulsating arrow at the tip of the dropper to guide the sampling personnel to align the dropper with the bottle opening.

[0107] Through the above process, this solution transforms the traditional "people looking up paper manuals" model into an "information revolving around people" model. Sampling personnel do not need to flip through materials or free their hands to operate mobile phones. They can obtain intuitive and accurate step guidance simply by "listening" and "watching" during the operation.

[0108] Furthermore, key node data during sample processing are collected, and the key node data is evaluated in real time. If the evaluation is successful, the record generation and label output instructions associated with the corresponding node are automatically triggered, specifically:

[0109] The sensor array deployed on the operating station and sample container collects real-time data on sample volume changes over time, protective agent drop acceleration rate, and dynamic fluctuation curves of water physicochemical parameters. By performing joint feature extraction on the time-series change data and dynamic fluctuation curves, a real-time process fingerprint characterizing the current sampling operation status is generated.

[0110] The real-time process fingerprint is input into the built-in sampling decision model, which is pre-set with a standard state machine and tolerance domain corresponding to the current sample type and sampling stage.

[0111] Calculate the Mahalanobis distance between the real-time process fingerprint and the standard state machine, and compare the Mahalanobis distance with the threshold of the tolerance domain to generate a process deviation score for the current node operation;

[0112] When the process deviation score falls within the tolerance range, the record generation instruction bound to the current sampling node is automatically triggered, and a time-stamped sampling electronic record containing the current sampling point, sample number and operator information is generated based on the key feature points in the real-time process fingerprint and the preset electronic record template.

[0113] Simultaneously, based on the sample number in the electronic sampling record, the label format data that matches the current sample container specifications and the current sampling stage is retrieved from the pre-stored label template library. The key feature points in the sample number, sampling time, and process deviation score are hashed to generate a label data stream containing a unique check code.

[0114] The data stream of the label to be printed is sent to the portable printer that is connected to the storage platform, driving the portable printer to perform the label output action.

[0115] Specifically, during the sampling process, a series of process data are collected in real time by a sensor array deployed on the operating station and sample container. For example, the time-series data of sample volume change over time is obtained through a miniature camera or liquid level sensor on the station, the droplet flow sensor at the dropper inlet is used to obtain the drop rate data of the preservative, and the dynamic fluctuation curves of water physicochemical parameters (such as pH, dissolved oxygen, etc.) are obtained through a multi-parameter water quality probe.

[0116] After acquiring the raw data, it is fused into a single overall feature vector through joint feature extraction, which this scheme defines as a "real-time process fingerprint." This concept is similar to human biometric fingerprints, and its core lies in transforming multi-dimensional time-series data into a feature identifier that can uniquely represent the current operational state. For example, in a certain sampling operation, the data streams of three dimensions—"the liquid level in the bottle rises uniformly to the 500ml mark," "the preservative is added at a rate of one drop per second," and "the pH value decreases linearly from 7.2 to 2.1"—together constitute the "process fingerprint" of that operation.

[0117] Then, the real-time process fingerprint is input into the built-in sampling decision model. It's important to note that the decision model is not a simple threshold judgment logic, but rather a standard state machine pre-generated based on a large amount of compliant sampling operation data. This state machine is essentially a reference trajectory in a high-dimensional space, representing the ideal data change pattern that a sampling operation (such as adding nitric acid as a preservative) should exhibit. Simultaneously, the model also presets a tolerance domain for each standard state, i.e., the range of allowable fluctuations.

[0118] In the evaluation phase, the Mahalanobis distance between the real-time process fingerprint and the standard state machine is calculated. Mahalanobis distance is used instead of Euclidean distance because it eliminates the influence of differences in data dimensions and takes into account the correlations between variables. For example, an increase in sample volume and an increase in protective dose are physically related, and Mahalanobis distance can more scientifically measure the overall deviation of the current operation from the standard model. The calculated distance value is compared with the tolerance threshold to generate a process deviation score, which intuitively reflects the quality level of the current operation.

[0119] When the process deviation score falls within the tolerance range, indicating that the operation is deemed compliant, the system automatically triggers the subsequent recording and output process. At this point, the system generates a timestamped electronic sampling record containing the current sampling point, sample number, and operator information, based on key feature points in the real-time process fingerprint (such as the final stable pH value and total volume) and a preset electronic record template.

[0120] In the label generation stage, an anti-counterfeiting and traceability design was further introduced. Based on the sample number in the electronic sampling record, label format data matching the current sample container specifications was retrieved. Then, the sample number, sampling time, and key feature points from the process deviation score were hashed together. The one-way and unique nature of the hash operation ensures a strong binding between the generated verification code and the process data of this operation; any subsequent tampering with the sample information will result in a mismatch of the verification code. The final generated label data stream is sent to the portable printer for output, thus achieving a fully automated closed loop from "operation process evaluation" to "record generation" to "anti-counterfeiting label output."

[0121] Furthermore, after the current sampling point task is completed, the storage platform is controlled to perform automatic storage, and the mobile platform is driven to autonomously navigate to the target location based on the preset coordinates of the next point. Specifically:

[0122] The system receives the current sampling point operation completion signal triggered by the sampling personnel through the interactive interface, and retrieves the preset storage strategy library according to the operation completion signal to generate a global storage instruction that includes the material placement order of each zone and the reset posture of the sunshade and side wind deflector.

[0123] The global storage command is sent to the attitude adjustment mechanism and the partition storage drive unit to control the sunshade and the side wind deflector to perform a coordinated folding action. At the same time, the push rod module in each partition is driven to push the external materials into the corresponding storage compartment and lock them until the feedback signals of the door lock status sensors of each partition are all displayed as closed, confirming that the storage platform has been restored to the transport locking configuration.

[0124] After confirming that the storage platform is in the transport locking configuration, the navigation and positioning module that is connected to the mobile platform is activated, the coordinates of the next point are read from the pre-stored task list, and the coordinates of the next point are compared with the current real-time positioning coordinates to generate navigation path planning data containing the start and end points.

[0125] The navigation path planning data is sent to the drive controller of the mobile platform. The drive controller controls the walking motor to perform autonomous driving actions based on the navigation path planning data, so that the mobile platform carrying the storage platform can automatically drive to the next point coordinate.

[0126] It should be noted that when the sampling personnel trigger the completion signal of the current sampling point through the interactive interface, the preset storage strategy library is retrieved. The storage strategy library stores ordered control schemes pre-generated based on the internal structure of the box and the material placement logic. For example, considering that dust may be shaken off when the sunshade is retracted, the strategy library will prioritize the pushing of small items on the table, and then execute the coordinated retraction of the sunshade and wind deflector to avoid secondary contamination.

[0127] After generating a global storage command based on the strategy, it is sent to the attitude adjustment mechanism and the zoned storage drive unit. During this process, the attitude adjustment mechanism is responsible for retracting the sunshade and side wind deflectors, while the miniature pusher modules within each zone are responsible for pushing materials (such as multi-parameter instruments, dropper racks, etc.) removed from the work surface back into their corresponding storage compartments. Only when all sensors return a closing signal does the system confirm that the storage platform has returned to a transport locking configuration with movement safety. This design avoids the risk of materials scattering during movement due to a drawer not being properly closed.

[0128] After confirming the locked configuration, the navigation and positioning module is activated. At this point, the module reads the preset coordinates of the next point from the task list and compares them with the current real-time positioning coordinates to plan a feasible navigation path. Finally, the drive controller controls the walking motor to perform autonomous driving actions based on this path data.

[0129] According to an embodiment of the present invention, the method further includes:

[0130] Collect the real-time process fingerprint constructed at the current sampling point, and retrieve the historical process fingerprint set generated in the previous sampling period at the same point. Segment and reassemble each historical process fingerprint in the historical process fingerprint set according to the sampling stage label to construct the historical operation trajectory cluster of each sampling stage at that point.

[0131] The real-time process fingerprint is segmented into the same stages to obtain the real-time sub-process fingerprints of each sampling stage. Each real-time sub-process fingerprint is then aligned with the historical operation trajectory cluster of the corresponding stage based on phase difference compensation using dynamic time bending. This aligns the time axis of the real-time sub-process fingerprint with the reference time axis of the historical operation trajectory cluster, resulting in the aligned and standardized real-time sub-sequence.

[0132] Based on the aligned historical operation trajectory clusters, the mean and standard deviation of the operation parameters at each time slice in each sampling stage are calculated, and a benchmark operation manifold band composed of multidimensional operation parameters is generated.

[0133] Calculate the Mahalanobis distance between the operation parameters on each time slice in the standardized real-time subsequence and the corresponding slice on the benchmark operation manifold, and extract the slice sequence whose Mahalanobis distance continuously exceeds a preset threshold as operation distortion segments;

[0134] The distorted operation segment is decomposed into spatial vectors to extract its projection components in the dimensions of pressure fluctuation, flow rate change, and liquid level response, thereby generating the distorted operation vector features of the corresponding segment.

[0135] The curl of the operation vector feature is calculated to obtain the vorticity value characterizing the operation distortion morphology. The vorticity value is then jointly encoded with the duration and occurrence stage label of the distortion segment to construct the implicit habit deviation feature tensor of the current operation.

[0136] The implicit habit deviation feature tensor is multiplied by a preset deviation weight matrix to generate a multidimensional deviation index that characterizes the degree to which the current operation deviates from historical operating habits. This multidimensional deviation index is used as a supplementary correction factor for the evaluation of operational standardization.

[0137] It should be noted that in actual sampling operations, due to differences in experience or operating habits among different operators, even if the final sample meets the specifications, subtle deviations in the operation process may affect the representativeness and stability of the sample. Traditional threshold-based evaluation methods can often only identify obvious violations, while deviations from the operator's own habits, such as disordered liquid addition rhythm due to fatigue or fluctuations in the dripping rate of the protective agent due to distraction, are difficult to effectively capture and quantify. To address this issue, this embodiment further mines the operational data in depth based on conventional process fingerprint evaluation. Specifically, after collecting the real-time process fingerprint constructed at the current sampling point, the historical process fingerprint set generated at that point within the previous sampling period is retrieved. This historical fingerprint data actually constitutes the operational experience database of that point under different operators or different working conditions. First, the historical data is segmented and recombined according to the sampling stage to construct the historical operation trajectory clusters of each stage at that point, thereby forming the "normal operating habit baseline" under the specific environment of that point. Subsequently, a dynamic time warping algorithm is used to perform phase difference compensation and alignment between the real-time process fingerprint and the historical trajectory cluster, so as to eliminate the time axis scaling effect caused by the difference in operation speed and ensure that the two are comparable.

[0138] Based on alignment, a baseline operating manifold band composed of multi-dimensional operating parameters is generated by fitting historical trajectory clusters. This manifold band differs from a single threshold line; it is actually a three-dimensional spatial region that allows for fluctuations, reflecting the reasonable range of changes in parameters (such as pressure, flow rate, and liquid level) over time in most historical operations. When real-time operating parameters continuously deviate from this manifold band, the system marks it as an operating distortion segment.

[0139] To gain a deeper understanding of the physical meaning of these distorted segments, spatial vector decomposition is performed to extract their projected components in key dimensions such as pressure fluctuations, flow rate changes, and liquid level response, generating operational vector features. By calculating the curl of these vector features, vorticity values ​​characterizing the operational distortion morphology can be obtained. This index directly reflects whether abnormal oscillations or disturbances exist during the operation. Finally, the vorticity values ​​are jointly encoded with information such as the duration and stage of the distorted segment to construct an implicit habitual deviation feature tensor. This tensor is then multiplied with a preset deviation weight matrix to generate a multidimensional deviation index. This embodiment can identify potentially risky operations that, while not exceeding the standard threshold, significantly deviate from the operator's own habits, and use them as supplementary correction factors to the routine evaluation results. For example, if a liquid addition operation is completed within the allowable time limit, but its flow rate curve exhibits abnormal fluctuations, the system will generate a high multidimensional deviation index, prompting management personnel to pay attention to the operational stability of the sample. This evaluation method achieves a leap from "compliance assessment" to "habitual deviation warning," providing a new technical means for the refined management of sampling quality.

[0140] According to an embodiment of the present invention, the method further includes:

[0141] Acquire multispectral remote sensing images and airborne lidar point cloud data within a preset range of sampling points. Perform endmember extraction on the multispectral remote sensing images to identify the distribution of surface mineral components. At the same time, construct an irregular triangular network from the airborne lidar point cloud data and calculate the surface roughness and runoff accumulation.

[0142] The distribution of surface mineral components is spatially superimposed with the surface roughness and the amount of runoff accumulation to generate a lithology-topography coupled feature vector characterizing the micro-domain hydrogeological conditions at the location.

[0143] Retrieve the final inventory records of the historical points that have been sampled and were triggered by the door lock status sensors of each partition of the storage platform at the end of the sampling task. Extract the name of the materials and the corresponding consumption from each final inventory record. Then, normalize the name and consumption according to the sampling point to construct a historical consumption association dataset containing the point identifier and the material consumption spectrum.

[0144] The lithology-topography coupling feature vector and the historical consumption associated dataset are spatially connected using the location identifier as the key to generate a set of lithology-topography-consumption triples associated with each historical location. Frequent pattern mining is performed on the triple set to extract material combination patterns that appear simultaneously with a frequency exceeding a preset support within different lithology-topography coupling feature vector intervals, which are used as initial co-efficiency rules.

[0145] Redundant rules are pruned on the initial co-efficiency rules to remove derived rules with confidence levels below a preset threshold or those implied by other rules, and the core co-efficiency rules with the highest lift are retained. The frequency of occurrence and lift of material combinations in the core co-efficiency rules are geometrically averaged to generate the material co-efficiency coefficients in the lithology-topography coupling feature vector interval corresponding to the core co-efficiency rules.

[0146] Obtain the inventory set of materials already inventoried in the material mapping model constructed in real time at the current sampling point, determine the target feature vector interval to which the lithology-topography coupling feature vector of the current point belongs, use the target feature vector interval as an index, match the corresponding rule from the core synergistic performance rule, and extract the supplementary material items that are complementary to the missing or insufficient items in the inventory set from the matched rule, and at the same time obtain the material synergistic performance coefficient of each supplementary material in the corresponding rule;

[0147] The items to be replenished are sorted from high to low according to their synergistic efficiency coefficients, and a material allocation preference recommendation list is generated for the current location's lithology-terrain coupling feature vector.

[0148] It should be noted that in actual field sampling work, the demand for consumables often varies implicitly among sampling sites under different geological conditions. For example, in areas with high clay mineral content, the frequency of filter membrane replacement may increase significantly; at sites with high surface roughness, the wear or consumption of sampling equipment may also exhibit different patterns. However, existing technologies typically use a standardized list for material preparation, making it difficult to dynamically adjust according to the actual geological and topographical characteristics of the site. This leads to a mismatch between material preparation and actual needs, potentially resulting in material waste and a shortage of key consumables that could disrupt sampling continuity.

[0149] Therefore, this embodiment first acquires multispectral remote sensing images and airborne lidar point cloud data within a preset range of the sampling points. By extracting endmembers from the multispectral images, the distribution of surface mineral components around the points is identified; an irregular triangular network is constructed from the point cloud data to calculate topographic parameters such as surface roughness and runoff accumulation. By spatially overlaying the aforementioned mineral components and topographic parameters, a lithology-topography coupled feature vector that can characterize the micro-regional hydrogeological conditions of the point is generated. This vector effectively assigns a unique "environmental fingerprint" to each sampling point.

[0150] Simultaneously, the final inventory records triggered at the end of the task for historical sampling points are retrieved. It's important to note that these records are generated by the door lock status sensors in each zone, accurately reflecting the actual names and quantities of materials consumed at each point. By normalizing this consumption data by point, a historical consumption correlation dataset containing point identifiers and material consumption genealogies is constructed, thereby establishing a material consumption archive for historical points.

[0151] Subsequently, using location identifiers as keys, the lithology-topography coupled feature vectors are spatially concatenated with the historical consumption-related dataset to generate a lithology-topography-consumption triplet set. Frequent pattern mining of this set reveals that specific material combinations occur significantly more frequently than in other lithology-topography feature intervals. For example, in the feature interval of "high clay minerals + low runoff accumulation," the combination of "filter membrane + protective agent A" occurs frequently with high confidence. Through redundant rule pruning and lift calculation, the core synergistic energy rules with the highest lift are retained, and a material synergistic energy coefficient is calculated for each rule to quantify the synergistic consumption intensity of the combination under specific geological conditions.

[0152] In the current sampling point preparation phase, the real-time constructed material mapping model is acquired to determine the set of materials in stock at the current location. The target feature vector interval is then identified by calculating its lithology-topography coupled feature vector. Using this interval as an index, core synergistic performance rules are matched, and materials that are missing or insufficient in quantity compared to the currently stocked materials are extracted from the rules. Simultaneously, the synergistic performance coefficient for each material is obtained. Finally, these materials to be supplemented are sorted from high to low according to their synergistic performance coefficients, generating a material configuration preference recommendation list tailored to the characteristics of the current location.

[0153] As can be seen, this embodiment realizes the transformation from an "experience-driven" material preparation mode to a "data-driven" intelligent recommendation mode. Sampling personnel can dynamically adjust the material configuration according to the recommendation list, making the material preparation more in line with the actual geological and topographical conditions of the site. This not only avoids the omission or redundancy of key consumables, but also improves the continuity and adaptability of sampling operations.

[0154] like Figure 2 As shown, the second aspect of this invention discloses a smart integrated system for groundwater sample sampling assistance and collection management, applicable to any of the smart integrated methods for groundwater sample sampling assistance and collection management described in any one of the claims, comprising:

[0155] An integrated storage platform includes a mobile base and a box set on the mobile base. The box is divided into multiple storage units for classifying and storing sampling materials. Each storage unit is equipped with a pull-out structure or a flip-top structure.

[0156] An ergonomic work surface system includes a work platform that can be folded up on the top of the housing, and a seat assembly that can be folded up on the side or bottom of the housing;

[0157] The comprehensive environmental protection system includes a retractable sunshade installed on the top of the enclosure and a detachable or foldable side wind deflector installed on the side of the enclosure. When the sunshade and the side wind deflector are unfolded, they form a three-sided surrounding protective work space.

[0158] The intelligent information interaction terminal includes a touch screen embedded in the operation panel, a voice acquisition module that is communicatively connected to the touch screen, and a printer module that is communicatively connected to the touch screen.

[0159] The intelligent mobility and energy management system includes a drive wheel assembly disposed within the mobile base, a navigation and positioning module communicatively connected to the drive wheel assembly, a solar power panel disposed on the top of the housing, and an energy storage battery pack disposed inside the housing and electrically connected to the solar power panel.

[0160] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A smart integrated method for groundwater sample collection assistance and management, characterized in that, Includes the following steps: In response to the start of the sampling task, the integrated storage platform is controlled to perform adaptive deployment and dynamically construct a field operation microenvironment that meets the sampling standards based on real-time perceived environmental parameters. The electronic tags of materials in each partition of the storage platform are identified, a real-time material mapping model of the current sampling point is constructed, and its conformity is verified with the standard sampling task list. The real-time material mapping model refers to the material information model containing spatial location index constructed by reading the electronic tags attached to the materials through the RFID reader array deployed in each partition, obtaining the unique feature code corresponding to each electronic tag, parsing the unique feature code into the name, specifications, current quantity and functional partition attribute of the material, and combining it with the physical location coordinates of each partition in the storage platform. The sampling personnel's work instructions are obtained, parsed, and matched with the corresponding sampling procedure data. The procedure data is then integrated with the current operation scenario through an interactive interface to generate enhanced information to guide the sampling execution. Key node data are collected during the sample processing process, and the key node data is evaluated in real time. If the evaluation is successful, the record generation and tag output instructions associated with the corresponding node are automatically triggered. The key node data refers to the sample volume time-series change data, protective agent drop acceleration rate data, and dynamic fluctuation curves of water physicochemical parameters, which are sensed in real time by the sensor array deployed on the operating station and sample container. After the task at the current sampling point is completed, the storage platform is controlled to perform automatic storage, and the mobile platform is driven to autonomously navigate to the target location based on the preset coordinates of the next point. Specifically: The system receives the current sampling point operation completion signal triggered by the sampling personnel through the interactive interface, and retrieves the preset storage strategy library according to the operation completion signal to generate a global storage instruction that includes the material placement order of each zone and the reset posture of the sunshade and side wind deflector. The global storage command is sent to the attitude adjustment mechanism and the partition storage drive unit to control the sunshade and the side wind deflector to perform a coordinated folding action. At the same time, the push rod module in each partition is driven to push the external materials into the corresponding storage compartment and lock them until the feedback signals of the door lock status sensors of each partition are all displayed as closed, confirming that the storage platform has been restored to the transport locking configuration. After confirming that the storage platform is in the transport locking configuration, the navigation and positioning module that is connected to the mobile platform is activated, the coordinates of the next point are read from the pre-stored task list, and the coordinates of the next point are compared with the current real-time positioning coordinates to generate navigation path planning data containing the start and end points. The navigation path planning data is sent to the drive controller of the mobile platform. The drive controller controls the walking motor to perform autonomous driving actions based on the navigation path planning data, so that the mobile platform carrying the storage platform can automatically drive to the next point coordinate.

2. The intelligent integrated method for groundwater sample sampling assistance and collection management according to claim 1, characterized in that, In response to the initiation of the sampling task, the integrated storage platform is controlled to perform adaptive deployment, dynamically constructing a field operation microenvironment that conforms to the sampling standards based on real-time perceived environmental parameters, specifically: After receiving the sampling task start command, the system collects the temperature, light intensity and wind speed data of the current work site in real time and defines them as the raw environmental parameter set. The temperature data in the original environmental parameter set is compared with the preset low temperature threshold and high temperature threshold, the light intensity data is compared with the preset strong light threshold, and the wind speed data is compared with the preset strong wind threshold. If the light intensity exceeds the strong light threshold, then the light intensity is defined as the dominant environmental factor; if the wind speed exceeds the strong wind threshold, then the wind speed is defined as the dominant environmental factor; if the temperature is lower than the low temperature threshold or higher than the high temperature threshold, then the temperature is defined as the dominant environmental factor. When multiple parameters exceed their corresponding thresholds at the same time, one parameter is selected as the dominant environmental factor from the parameters that exceed the thresholds according to a preset priority rule. The priority rule is ordered as follows: strong wind threshold, strong light threshold, low temperature threshold or high temperature threshold. According to the dominant environmental factors, the corresponding microenvironment construction strategy is matched from the preset protection strategy library. The protection strategy library stores the correspondence between different dominant environmental factors and the extension angle of the sunshade and the posture of the side wind deflector. Based on the microenvironment construction strategy, extension angle control commands for the sunshade and attitude adjustment commands for the side wind deflectors are generated. The extension angle control command and attitude adjustment command are sent to the attitude adjustment mechanism of the storage platform to drive the sunshade and side wind deflector to perform coordinated unfolding action, forming a three-sided physical protective barrier around the sampling station. The actual position feedback signal after the attitude adjustment mechanism performs the action is collected and compared with the target position data in the microenvironment construction strategy. Based on the comparison result, it is confirmed that the physical protective barrier has reached the preset airtightness and shielding standards, thereby completing the construction of a field operation microenvironment that meets the sampling specifications.

3. The intelligent integrated method for groundwater sample sampling assistance and collection management according to claim 1, characterized in that, The electronic identifiers of materials within each zone of the storage platform are identified, a real-time material mapping model for the current sampling point is constructed, and its conformity is verified against the standard sampling task list. Specifically: Activate the RFID reader array deployed in each partition of the storage platform, batch read the electronic tags attached to the materials in each partition, and obtain the unique feature code corresponding to each electronic tag; The unique feature code is sequentially input into the embedded material parsing unit. Through the feature code-material mapping table built into the material parsing unit, the name, specifications, current quantity and functional partition attributes of each material are reversely parsed. All the parsed materials are classified and collected according to the functional partition attributes, and a real-time material mapping model containing spatial location index is constructed by combining the physical location coordinates of each partition within the storage platform. Retrieve a standard sampling task list pre-stored in the local memory. The standard sampling task list defines the baseline name, baseline specifications, baseline quantity, and preset storage partition information of the materials required for this sampling. The material details of each partition in the real-time material mapping model are compared item by item with the material requirements of the corresponding partition in the standard sampling task list to identify the missing material items, redundant material items, and quantity deviation items in each partition relative to the list. Based on the missing material items, redundant material items, and quantity deviation items identified in each partition, a compliance verification report of the material configuration at the current sampling point is generated, and the verification report is pushed to the interactive interface to guide the sampling personnel to supplement or adjust the materials.

4. The intelligent integrated method for groundwater sample sampling assistance and collection management according to claim 1, characterized in that, The system acquires the sampling personnel's work instructions, parses and matches them with the corresponding sampling procedure data, and then integrates the procedure data with the current operation scenario through an interactive interface to generate enhanced information to guide the sampling execution. Specifically: The original speech signal containing keywords of the sampling operation is acquired by the sampling personnel through a bone conduction microphone. Endpoint detection and semantic parsing are performed on the original speech signal to extract the target action intention and target sample object corresponding to the current operation. The target action intent and target sample object are input into a pre-constructed sampling knowledge graph for vectorized retrieval and association matching. Multimodal sampling procedure data packages that match the current operation scenario are retrieved from the sampling knowledge graph. The multimodal sampling procedure data packages include text-based operation procedures and spatial location guidance identifiers. Acquire real-time visual images of the current operation scene, perform instance segmentation and recognition on the real-time visual images, and identify the entity categories and spatial coordinates of sampling bottles, sampling tools and sample containers in the images; The spatial location guide identifier in the multimodal sampling procedure data package is aligned with the identified entity spatial coordinates and subjected to perspective projection transformation to generate augmented reality rendering instructions bound to the entity in the current operation scene. According to the augmented reality rendering instructions, the text-based operation procedures are overlaid on the corresponding entity on the interactive interface, and the spatial location guidance mark is dynamically projected onto the surface of the corresponding entity in the form of a virtual arrow, forming an augmented information screen to guide the sampling personnel to perform the next operation.

5. The intelligent integrated method for groundwater sample sampling assistance and collection management according to claim 1, characterized in that, Key node data are collected during the sample processing, and the key node data is evaluated in real time. If the evaluation is successful, the record generation and label output instructions associated with the corresponding node are automatically triggered, specifically: The sensor array deployed on the operating station and sample container collects real-time data on sample volume changes over time, protective agent drop acceleration rate, and dynamic fluctuation curves of water physicochemical parameters. By performing joint feature extraction on the time-series change data and dynamic fluctuation curves, a real-time process fingerprint characterizing the current sampling operation status is generated. The real-time process fingerprint is input into the built-in sampling decision model, which is pre-set with a standard state machine and tolerance domain corresponding to the current sample type and sampling stage. Calculate the Mahalanobis distance between the real-time process fingerprint and the standard state machine, and compare the Mahalanobis distance with the threshold of the tolerance domain to generate a process deviation score for the current node operation; When the process deviation score falls within the tolerance range, the record generation instruction bound to the current sampling node is automatically triggered, and a time-stamped sampling electronic record containing the current sampling point, sample number and operator information is generated based on the key feature points in the real-time process fingerprint and the preset electronic record template. Simultaneously, based on the sample number in the electronic sampling record, the label format data that matches the current sample container specifications and the current sampling stage is retrieved from the pre-stored label template library. The key feature points in the sample number, sampling time, and process deviation score are hashed to generate a label data stream containing a unique check code. The data stream of the label to be printed is sent to the portable printer that is connected to the storage platform, driving the portable printer to perform the label output action.

6. The intelligent integrated method for groundwater sample sampling assistance and collection management according to claim 1, characterized in that, Before obtaining the sampling personnel's work instructions, the process also includes collecting the sampling personnel's facial feature information through an interactive interface, comparing it with a pre-stored authorized personnel database for authentication, and activating the voice collection function only after successful authentication.

7. The intelligent integrated method for groundwater sample sampling assistance and collection management according to claim 4, characterized in that, The enhanced information screen also includes dynamically adjusting the flashing frequency or color of the virtual arrow based on the changes in the liquid level of the sampling bottle in the real-time visual image, so as to prompt the sampling personnel to control the sampling flow rate.

8. A smart integrated system for groundwater sample sampling assistance and collection management, applied to the smart integrated method for groundwater sample sampling assistance and collection management as described in any one of claims 1 to 7, characterized in that, include: An integrated storage platform includes a mobile base and a box set on the mobile base. The box is divided into multiple storage units for classifying and storing sampling materials. Each storage unit is equipped with a pull-out structure or a flip-top structure. An ergonomic work surface system includes a work platform that can be folded up on the top of the housing, and a seat assembly that can be folded up on the side or bottom of the housing; The comprehensive environmental protection system includes a retractable sunshade installed on the top of the enclosure and a detachable or foldable side wind deflector installed on the side of the enclosure. When the sunshade and the side wind deflector are unfolded, they form a three-sided surrounding protective work space. The intelligent information interaction terminal includes a touch screen embedded in the operation panel, a voice acquisition module that is communicatively connected to the touch screen, and a printer module that is communicatively connected to the touch screen. The intelligent mobility and energy management system includes a drive wheel assembly disposed within the mobile base, a navigation and positioning module communicatively connected to the drive wheel assembly, a solar power panel disposed on the top of the housing, and an energy storage battery pack disposed inside the housing and electrically connected to the solar power panel.

Images