A multi-dimensional intelligent verification method for building material goods signing

Abstract

The application discloses a multi-dimensional intelligent verification method for building material goods signing and receiving; a "cloud-edge-end" cooperation and "sensing-calculation-storage" integrated architecture is constructed, taking material unique identity coding as the core, a multi-source heterogeneous data cross verification and block chain credible evidence storage mechanism is established, and through six steps, full-process intelligent verification is realized: building an encryption benchmark database, dynamic code assignment of a delivery unit is chained, multi-level identity authentication is started for verification, four-dimensional data acquisition is performed, cloud edge cooperation comparison is performed, and four-party signature reports are chained and stored; the multi-dimensional intelligent verification method for building material goods signing and receiving reduces the quantity error rate to below 0.2%, the specification recognition accuracy rate reaches 99.5%, the bulk material verification efficiency is improved by more than 80%, the dispute rate is reduced by 95%, data islands are broken, the whole life cycle of the material is tracked, the supply chain is digitally upgraded, and the method has remarkable technical innovation and economic benefits.

Description

Technical Field

[0001] This invention belongs to the comprehensive technical field of building material supply chain management, industrial Internet of Things, machine vision, edge computing and blockchain technology, and specifically is a multi-dimensional intelligent verification method for the receipt of building material goods. Background Technology

[0002] Construction material costs typically account for 60%-70% of the total project cost. Material delivery and acceptance are crucial quality control and cost management points connecting the upstream supply chain with construction production. However, for a long time, my country's construction industry has generally relied on a traditional manual method for material acceptance, primarily depending on manual verification of paper purchase contracts, delivery notes, and certificates of conformity, supplemented by simple tools such as measuring tapes and crane scales for random sampling and acceptance of quantity and specifications. This model suffers from the following systemic defects and industry pain points: Low verification accuracy and high error rate: When manually counting bulk materials (such as sand, gravel and steel bars), omissions and errors are common. According to statistics, the error rate of traditional methods can reach 3%-5%. The identification of specifications and models relies on the personal experience of the material handlers. Key parameters such as steel bar grade, cement grade and waterproof membrane thickness are easily confused, which poses a major quality hazard. Low operational efficiency and severe on-site congestion: Weighing bulk materials requires repeated weighing and separation of tare weight and vehicle weight, with each operation taking 30-60 minutes; During peak periods when multiple trucks arrive at the same time, the queue of vehicles at the construction site gate often reaches hundreds of meters, affecting the construction progress and urban traffic. The unclear definition of responsibility leads to frequent disputes: Paper receipts are easily altered, forged, or lost. When problems such as material shortages, non-compliance with specifications, or quality defects occur, the responsibility chain among the supplier, logistics carrier, construction party, and supervisor is unclear, resulting in serious buck-passing and insufficient evidentiary value in legal proceedings. The data silos are severe and the management is opaque: purchase order data is stored in the ERP system, logistics information is scattered across various carrier platforms, and on-site acceptance data is recorded in paper forms or isolated Excel files. The data cannot be interconnected, making it difficult to form a full-chain traceability and unable to support big data analysis, supplier performance evaluation and procurement strategy optimization. Lack of effective anti-cheating mechanisms: Traditional weighbridges are susceptible to cheating methods such as remote control interference and adding water to the water tank, and manual verification is difficult to identify fraudulent activities such as cloned vehicles, swapped goods, and selling inferior goods as superior ones.

[0003] Therefore, there is an urgent need for an integrated, intelligent, and reliable systemic solution to achieve the automation, standardization, and reliability transformation of building material acceptance and verification. Summary of the Invention

[0004] The fundamental purpose of this invention is to provide a multi-dimensional intelligent verification method for the receipt of building materials. By constructing a technical architecture that integrates "cloud-edge-device" collaboration and "sensing-computing-storage", it systematically solves the core technical problems of low accuracy, slow efficiency, difficulty in tracing responsibility, and easy data tampering in the traditional manual receipt mode, and promotes the upgrading of supply chain management in the construction industry towards digitalization, intelligence, and transparency.

[0005] To achieve the above objectives, the present invention adopts the following technical solution, the core innovation of which lies in establishing a multi-source heterogeneous data cross-validation and trusted evidence storage mechanism based on the unique identification code of materials: 1. Baseline Database Construction and Encrypted Storage: During the procurement order generation stage, the system is deeply integrated with the enterprise ERP system or project BIM 5D management platform to automatically extract the unique material codes (following the ISO / IEC 15459 global unique identifier standard), specification geometric parameters, theoretical physical properties (calculated based on the national standard GB / T 208-2014 for material density and the tolerance standard GB / T 702-2017), three-dimensional digital features (PLY format point cloud model constructed through pre-sampling scans), standard values ​​of chemical composition (for key materials such as steel bars and cement), and supplier quality certification information (ISO9001 certificate, production license). After encryption using the national cryptographic SM4 block encryption algorithm (128-bit block length, 128-bit key length), a JSON structured baseline data packet is formed and synchronized to the cloud verification platform via the TLS 1.3 secure transmission protocol. Large-capacity three-dimensional model files are stored in a distributed deduplication manner using the IPFS interplanetary file system, with only the file hash value stored in a relational database.

[0006] 2. Dynamic Identification and On-Chain Logistics Information: Suppliers attach machine-readable tags to quality-inspected shipping units (single pieces, bundles, and truckloads) using industrial-grade coding equipment (such as the Zebra ZT411 printer or Chainway R6 RFID card writer). For harsh outdoor environments, UHF RFID anti-metal tags (such as the Omni-ID Exo 750, IP68 protection rating, and a reading distance of up to 10 meters) or NFC chips (such as the NTAG424 DNA, supporting the SUN identity authentication protocol) are preferred. For ordinary packaging, dynamic encrypted QR codes can be used. Simultaneously, key elements of the shipping event (shipping timestamp accurate to milliseconds, vehicle VIN code, driver's decentralized DID identity, photos captured by the cab camera, and logistics trajectory data) are hashed using SHA-256, and written to the FISCO BCOS consortium blockchain via a smart contract. This generates transaction credentials with block height and timestamps, ensuring the immutability and auditability of logistics information.

[0007] 3. On-site Verification Triggering and Identity Authentication: After the goods arrive at the pre-defined geofence area (supporting circular and polygonal geographical boundaries) at the construction site, verification personnel activate the verification task using a mobile terminal (such as a Huawei Mate 60 RS or a customized industrial tablet) equipped with NFC reading and writing, QR code scanning, and high-precision GPS positioning (supporting RTK differential positioning with centimeter-level accuracy). The system first performs multi-level identity authentication: ① Verification personnel's facial recognition or fingerprint recognition; ② Digital certificate signature challenge; ③ Geographical location validity verification. Only after all three are successful can the system retrieve the corresponding order's baseline data package from the cloud, ensuring the authoritativeness and non-repudiation of the verification operation.

[0008] 4. Multimodal heterogeneous data acquisition: The system synchronously schedules at least three types of heterogeneous sensing devices to form a "field sensing network" to ensure the comprehensiveness and redundancy of data acquisition. Physical quantity dimension: Weight data is acquired through IoT weighbridge (digital sensor interface RS485 / CAN, supports Modbus / TCP protocol, protection level IP67) or Bluetooth 5.2 hook scale (capacity 50 tons, accuracy level OIML C3), and real-time temperature compensation value and zero drift data of weighing sensor are collected to evaluate the reliability of measurement. Geometric dimensions: Deploy an industrial multi-view vision inspection module (including a 12-megapixel RGB camera, a ToF depth camera, and a structured light projector) to capture the stacked state of goods from multiple perspectives and generate point cloud data (PCD format) with color information; or use a handheld laser rangefinder (Leica DISTO D810, accuracy ±1mm) for key dimension spot checks. Chemical quantity dimension: For critical structural safety materials, a portable XRF spectrometer (such as Olympus Vanta) or LIBS laser-induced breakdown spectrometer is used to test the C, Si, Mn, P and S element content of steel bars on site and compare it with the reference chemical composition range to prevent unqualified steel from entering the market. Environmental dimension: Deploy LoRaWAN temperature and humidity sensors (such as RAK7204 from Ruike Huilian, with a 3-year battery life) inside the carriage to continuously monitor the transportation microenvironment. The data is uploaded to the cloud in real time through the gateway to ensure the quality stability of environmentally sensitive materials such as cement and admixtures.

[0009] 5. Cloud-Edge Collaborative Intelligent Comparison: Collected data is uploaded to the cloud via 5G network slicing (URLLC ultra-low latency mode) or Wi-Fi 6. The platform first runs a data quality assessment engine to check data integrity, timestamp consistency, and device calibration status. Once qualified, a multi-dimensional cross-comparison algorithm is activated. Weight verification: Calculate the deviation rate ΔW, with a pass threshold of ±3% and a warning threshold of ±5%; Specification verification: Point cloud registration is performed using an improved ICP algorithm (with RANSAC random sampling consistency to remove outliers), and the RMSE and critical dimension matrix are output. Appearance verification: The YOLOv8n lightweight model deployed on edge computing nodes (such as Huawei Atlas500) performs real-time defect detection, and identifies categories including cracks, corrosion, deformation, contamination, and chipped edges and corners, and outputs defect bounding boxes and confidence scores; Chemical verification: The XRF spectral data is compared with the benchmark value for similarity calculation (cosine similarity ≥ 0.95 is considered acceptable). Environmental verification: Plot temperature and humidity curves during transportation to check if they exceed the material storage standard envelope; Comprehensive decision engine: It adopts a weighted voting mechanism (weight 30%, specifications 30%, appearance 20%, chemistry 10%, environment 10%). If the total score is ≥90 points and the key items (weight and chemistry) are qualified, it will automatically pass.

[0010] 6. Blockchain-based Immutable Evidence Preservation and Multi-Party Signatures: Upon successful verification, the system automatically generates a verification report conforming to the PDF / A-3 long-term preservation standard. The report includes order information, benchmark data, actual collected data, detailed comparison results, equipment calibration certificates, and on-site photos (with digital watermarks). The report undergoes four-party digital signature processing using the national cryptographic algorithm SM2 (supplier quality inspector, carrier driver, construction party material handler, and supervising engineer). The signing process utilizes a layered deterministic key (BIP32 HD Wallet) to derive private keys, ensuring key security. The digital fingerprint (SHA-256 hash) of the signature, along with a timestamp and geographic coordinates, is written to the blockchain via a smart contract, generating a unique evidence preservation number (e.g., BCOS-2025-0119-000123). This number is printed on the electronic receipt and a QR code is generated for subsequent audit verification. The blockchain employs the PBFT (Practical Byzantine Fault Tolerance) consensus mechanism, ensuring consensus is reached even when some nodes fail, with evidence confirmation time less than 3 seconds.

[0011] Compared with the prior art, the beneficial effects of the present invention are: This invention constructs a multimodal heterogeneous data cross-validation system, integrating multi-dimensional detection data such as weight, geometry, chemistry, and environment. Through improved intelligent algorithms such as the ICP algorithm and the YOLOv8n lightweight model, it achieves accurate comparison, reducing the quantity error rate of 3%-5% in the traditional manual mode to below 0.2%, increasing the specification identification accuracy to 99.5%, and the chemical composition misjudgment rate to below 0.1%. For key materials such as steel bars, cement, and precast composite slabs, it can accurately identify quality hazards such as cracks, corrosion, dimensional deviations, and substandard composition, effectively preventing unqualified materials from entering the construction site and ensuring the structural safety and durability of building projects from the source. Compared to the inefficient traditional manual verification method that takes 30-60 minutes per vehicle, this invention, through a cloud-edge-device collaborative architecture and automated sensing equipment, reduces the average verification time for bulk materials to 8 minutes and further shortens the verification time for small-amount, high-frequency emergency materials to within 90 seconds, improving operational efficiency by more than 80%. Simultaneously, automated data collection and intelligent decision-making replace cumbersome processes such as manual document verification and measurement, increasing vehicle turnover at the construction site entrance by 3 times and reducing congestion by 85%, completely resolving the problem of queuing and backlog of material transport vehicles during peak periods and ensuring the orderly progress of construction. This invention, based on the FISCOBCOS consortium blockchain and national cryptographic algorithms, constructs an immutable end-to-end data storage mechanism. Key information such as purchase orders, logistics information, on-site testing data, and four-party digital signatures are all stored on the blockchain, generating a unique storage number and traceable certificate. This mechanism gives the verification report the legal effect recognized by the Electronic Signature Law, clarifying the responsibilities of upstream and downstream suppliers, carriers, construction parties, and supervisors in the supply chain, reducing the dispute rate by 95%, and achieving a 100% evidence acceptance rate in legal proceedings. It effectively solves industry pain points such as the ease of alteration, loss, and shirking of responsibility associated with traditional paper documents. This invention reduces manual labor through automated verification and avoids fraudulent activities such as remote control interference, product swapping, and substandard goods by combining intelligent anti-fraud mechanisms, reducing material waste by 3%-5% annually. At the same time, blockchain evidence storage and digital processes simplify management links such as dispute coordination and audit traceability, reducing labor costs by 70% and overall material management costs by 15%-20%. In addition, the full-chain data accumulated by the system supports the construction of a supplier credit evaluation model, increasing the proportion of procurement from high-quality suppliers by 30%, and realizing dynamic optimization of procurement strategies and precise cost control.

[0012] This invention breaks through the traditional "data silo" dilemma in the construction industry, integrating data from multiple systems such as ERP, BIM5D, and IoT sensing to form data assets covering the entire lifecycle of materials from "procurement to production to logistics to verification to use." This data can support refined management functions such as material consumption prediction (error ±5%), intelligent replenishment reminders (lead time 3-5 days), and dynamic cost analysis, providing a data foundation for project decision-making. At the same time, data systems such as supplier credit scoring and material quality traceability promote the transformation of the building materials supply chain from traditional extensive management to digitalization, intelligence, and transparency. Attached Figure Description

[0013] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.

[0014] In the attached diagram: Figure 1This is a schematic diagram of the multi-dimensional intelligent verification method for signing for building materials goods according to the present invention.

[0015] Figure 2 For the present invention Figure 1 Partial process diagram Figure 1 .

[0016] Figure 3 For the present invention Figure 1 Partial process diagram Figure 2 .

[0017] Figure 4 For the present invention Figure 1 Partial process diagram Figure 3 . Detailed Implementation

[0018] The following will refer to the accompanying drawings in the embodiments of the present invention. Figures 1 to 4 The technical solutions in the embodiments of the present invention will be clearly and completely described.

[0019] Example 1: High-precision verification of bulk delivery of HRB400E rebar for building construction projects Project Background: A high-rise residential project with a building area of ​​180,000 square meters and a total steel reinforcement consumption of approximately 12,000 tons; this batch of procurement involves 300 bundles of HRB400E rebar with a diameter of 25mm and a length of 12m, with a theoretical weight of 60 tons. The contract requires a deviation rate of no more than ±3‰, and the steel reinforcement is supplied by the Class A supplier "Huaxin Steel".

[0020] Baseline Database Construction: When a purchase order is generated on the Glodon BIM5D platform, the system automatically extracts the national standard parameters of the steel bar of that type from the material library: nominal weight 3.85kg / m, theoretical total weight 60,000kg, and a total of 3,896 bars; at the same time, it retrieves the three-dimensional point cloud baseline model of the production batch (file size 47MB, PLY format, containing 15,842 points, accuracy 0.1mm) from the supplier's quality traceability system. The model covers key geometric features of the steel bar, such as the base circle diameter of 25mm, transverse rib height of 1.6mm, longitudinal rib height of 1.2mm, and rib spacing of 12mm, as well as the standard values ​​of chemical composition, such as C content of 0.22% and Si content of 0.55%.

[0021] Coding and Logistics Blockchain Linking: The supplier's quality inspector uses a handheld coding terminal (Xiaobang X8) to print a dynamic encrypted QR code for each bundle of steel bar binding straps. The content of the QR code is encrypted by SM4 and includes the order number PO-XM2025-0119-001, material code GJ-HRB400E-25-12, production furnace number L20250115003, and batch number BT-0119-A. After the vehicle (license plate: Beijing A·K1234, VIN code: LVSHCAME6LE123456) is loaded, it drives over the weighbridge (METTLER TOLEDO POWERCELL GDD, maximum capacity 100 tons, verification scale interval 20 kg) to obtain the gross weight data of 78.35 tons. The weighbridge data is directly connected to the edge gateway through the Modbus TCP protocol, and together with the Beidou positioning trajectory (one point every 30 seconds), it is packaged and called the smart contract to write into the FISCO BCOS consortium blockchain, block height 12345678, transaction hash 0x9a7b...f1c3, and the blockchain linking delay is 2.1 seconds.

[0022] On-site Verification Trigger: At 14:30 on January 19, 2025, the vehicle arrives at the north gate of the construction site, GPS coordinates 116°23'E, 39°54'N, within the project electronic fence (error radius 50 meters). The material handler Zhang San uses a Huawei Mate 60 RS mobile phone to open the "Caizhike" APP. After passing the face recognition, he scans the QR code at the rear of the carriage. The system verifies that the SM4 check code of the QR code is valid, and it takes 1.2 seconds to download the reference data packet. At the same time, the APP interface shows that the credit score of this supplier is 880 points (excellent), and the system automatically fine-tunes the weight threshold to ±3.2%.

[0023] Multi-modal Data Collection: Weight Dimension: The vehicle drives onto the Internet of Things weighbridge at the construction site (model: Jinzhong XK3190-DS10, 6 digital sensors, protection level IP68). The platform automatically obtains the tare weight of 18.40 tons and calculates the net weight of 59.95 tons, ΔW = -0.083%, far lower than the contract requirement of ±3‰. The built-in temperature sensor of the weighbridge shows the current ambient temperature of -5°C, and the system automatically calls the temperature compensation coefficient to correct the weighing value to ensure that the accuracy is not affected in low-temperature environments.

[0024] Image and point cloud dimensions: Upon startup, the tablet's quad-camera module (50MP main camera, 48MP ultra-wide-angle, 12MP telephoto, and 3D depth-sensing lens) automatically captured 12 high-resolution images (8192×6144 resolution, JPEG format, approximately 8MB each) from three preset points: a 45° downward angle at the rear of the vehicle, a 90° upward angle from the side, and a vertical angle at the top. These images were then uploaded to an edge computing node (Huawei Atlas 500 Pro) via 5G network slicing (URLLC mode, 100Mbps uplink bandwidth, 12ms latency). The YOLOv8n model on the node (fine-tuned with historical data from this project, weight file 12MB) performed real-time inference, identifying a total of 300 bundles of steel bars, consistent with the theoretical value. Simultaneously, the structured light camera generated on-site point cloud data. After removing outliers such as the carriage baffle using the RANSAC algorithm, 12,456 effective point clouds were obtained. The ICP registration algorithm (maximum number of iterations 500, error threshold 0.001) was run and compared with the benchmark model, measuring an average diameter of 25.03mm and a transverse rib height of 1.61mm, with deviation rates of +0.12% and +0.625%, respectively, both better than the ±2% threshold.

[0025] Chemical dimension: Three steel bars were randomly selected, and the surface oxide layer was removed by grinding with an angle grinder. Then, a portable XRF spectrometer (Olympus Vanta VMR, 50kV Rh anode X-ray tube) was used for testing. The C content was 0.23% and the Si content was 0.56%, with deviations from the reference values ​​of +0.01% and +0.01%, respectively, which are within the allowable error range.

[0026] Environmental dimension: The three temperature and humidity monitoring points installed in the carriage showed that the temperature during transportation ranged from -8℃ to 5℃ and the humidity ranged from 45% to 65%, which did not exceed the steel bar storage standard (temperature from -20℃ to 40℃, humidity ≤80%).

[0027] Intelligent comparison and decision-making: The cloud-based verification platform runs a comprehensive decision-making engine. The five verification weight scores are as follows: weight 30 points (full marks), specifications 30 points (full marks), appearance 20 points (2 points were deducted for finding 3 bundles of slightly damaged packaging tape, resulting in 18 points), chemistry 10 points (full marks), and environment 10 points (full marks), for a total score of 98 points. Since the score is greater than the threshold of 90 points, the system automatically determines that the verification is "passed" and provides a green light notification on the APP. The entire process takes 7 minutes and 23 seconds, which is 83.7% more efficient than traditional manual verification (approximately 45 minutes).

[0028] Blockchain Evidence Preservation: The material handler Zhang San, the carrier driver Li Si, the supervisor Wang Wu, and the supplier quality inspector Zhao Liu use the APP to perform national secret SM2 electronic signatures. The signature process calls the built-in security chip of the mobile phone (meeting the EAL5+ level), and the private key is never exposed. The verification report PDF file (3.2MB in size) is calculated by SHA-256 to obtain the hash value 0x3c8a...e9f4. Together with the four signature values, the timestamp, and the GPS coordinates, they are packaged into a transaction body, and the smart contract writeReceipt() method is called to write it into the FISCO BCOS chain. The block height is 12345689, and the transaction confirmation time is 2.8 seconds. After the evidence preservation is successful, the system generates an electronic receipt with a QR code, which is printed and posted in the material stacking area. Any subsequent project participants can verify the validity of the evidence preservation by scanning the code.

[0029] Example 2: Verification of the Quality Stability of Bulk Cement Transportation in a Commercial Concrete Station 8>Project Background: For a subway station project, it is required to continuously supply P.O 42.5 low-alkali cement, with extremely high requirements for the soundness, strength, and alkali content of the cement, and it is necessary to prevent moisture absorption and caking during transportation. This batch supplies 60 tons of bulk cement, which is transported by a tanker truck (license plate: Beijing A·98765).

[0030] Specialization of Benchmark Data: As a powdery material, cement cannot be counted bag by bag. The key points of its verification are weight accuracy, chemical properties, and environmental stability. The purchase order includes: theoretical weight 60,000 kg, strength grade 42.5 MPa, alkali content ≤ 0.6%, SO3 content 2.0 - 3.5%, fineness residue on 0.08 mm square hole sieve ≤ 10%. At the same time, the factory inspection report (certificate) PDF file of this batch of cement (already OCR-recognized as structured data) and the 28-day strength prediction curve are embedded in the benchmark database.

[0031] Code Assignment and Environmental Monitoring: The supplier installs a UHF RFID anti-metal tag (Omni-ID Exo 800, reading distance 15 meters, storage capacity 2KB) at the manhole cover of the cement tanker truck. The order number, cement batch, factory time (2025-01-19 08:00:00), and tanker truck volume (35m <T 3 ³) are written into the tag. Three Sensirion SHT40 temperature and humidity sensors (accuracy ±0.2°C, ±1.5%RH) are deployed inside the tanker truck, and data is uploaded to the cloud every 5 minutes through a LoRaWAN gateway to form a transportation environment curve.

[0032] On-site verification innovation: After the tanker truck arrives at the construction site, the material handler does not need to climb the tank. He uses a handheld UHF reader (Chainway C72, Android 11 system, reading power 30dBm) to read RFID tag information in batches from 5 meters away. The APP shows that the highest temperature during the transportation of the tanker truck was 15℃ (affected by ambient temperature) and the humidity was 58%, which did not exceed the standard.

[0033] Weight verification: The tanker truck was first weighed on the weighbridge to obtain a gross weight of 53.2 tons. After unloading into the cement silo on site, it was weighed again to obtain a tare weight of 15.8 tons. The system automatically calculated the net weight to be 37.4 tons. Due to adhesion to the inner wall of the tanker truck and incomplete unloading, the allowable deviation rate was relaxed to ±5%, ΔW=-2.7%, which is qualified.

[0034] Rapid chemical testing: Samples were taken from the tanker unloading port and tested using a portable XRF spectrometer (Brook S1 TITAN 800, detection limit 10ppm). The alkali content was 0.58%, SO3 content was 2.8%, and fine sieve residue was 8.5%, all of which were qualified. The data was directly connected to the APP via Bluetooth to prevent manual entry and tampering.

[0035] Automated sample retention and sealing: When the system triggers a command, the automatic sampler (customized equipment) extracts three 5kg cement samples from the unloading flow, puts them into a standard sample bucket and attaches a tear-resistant NFC tag. The tag is written with the evidence hash value of the main verification report, realizing a strong correlation between the main report and the retained sample, and meeting the requirements of subsequent quality inspection station re-inspection.

[0036] Blockchain-based evidence storage: After four parties sign, the verification report is uploaded to the blockchain. What's special is that the report embeds an SVG graph of temperature and humidity changes throughout the transportation process and a CSV file of raw spectral data from a rapid chemical test. These data are used together in hash calculations to ensure that any minor tampering can be detected. The gas fee for the evidence storage transaction is approximately 0.002 ETH, which is borne by the supplier.

[0037] Example 3: Customized High-Precision Verification of Prefabricated Composite Slabs Project Background: A prefabricated affordable housing project requires the procurement of 50 truss reinforced concrete composite slabs, each measuring 3.0m × 6.0m × 0.2m. The theoretical weight of each slab is 9.02 tons. The components are of high value, large size, and contain many embedded parts, requiring meticulous verification of each slab.

[0038] 3D baseline data generation: During factory production, each composite slab is generated into an actual production point cloud model using a fixed 3D laser scanner (FARO Focus S350, scanning speed 976,000 points / second, accuracy ±1mm) before demolding. After comparison with the BIM theoretical model, the actual model (not the theoretical model) is uploaded as the baseline data. The baseline data includes: external dimensions (length × width × height), 3D coordinates (XYZ) of 8 hoisting pre-embedded nuts, top surface elevation of truss reinforcement, rough surface depth, and unique component identification code (laser engraved on the side wall).

[0039] Transportation process monitoring: The composite slab transport vehicle is a special flatbed truck, with sleepers under each component. Three-axis accelerometers (ADI ADXL345, range ±16g) are installed at the four corners of the truck bed, with a set threshold of ±3g. If sudden braking or sharp turns during transportation cause the impact force to exceed the standard, the sensor will immediately report an early warning and trigger a high-definition camera to capture the road conditions. The data will be uploaded to the blockchain for evidence storage and to clarify responsibility.

[0040] On-site multi-equipment collaboration: After the components arrived, two gantry cranes were used to lift them simultaneously (each with a rated lifting capacity of 10 tons). Wireless force sensors (AVIC Electromechanical Measurement BSL-5t, accuracy 0.05%FS) were installed at the hooks. The data from the two hooks were aggregated to the edge gateway via a Wi-Fi 6 Mesh network. The total weight was 9.08 tons, ΔW=+0.66%, which is qualified.

[0041] Laser tracking measurement: A Leica AT960 laser tracker (accuracy ±0.02mm, measurement radius 80m) was used to perform non-contact measurement on the component's shape. Within 5 minutes, 12 geometric parameters, including the length of the four sides, diagonal difference, and flatness, were acquired and automatically compared with the reference point cloud model. The results showed that the length deviation was +2mm, the width deviation was -1mm, the diagonal difference was 3mm, and the flatness was 2mm / m, all of which were better than the allowable deviations of GB / T 51231.

[0042] Embedded part location verification: The area of ​​the embedded nut is photographed by a structured light camera. The image is identified by AI to identify the three-dimensional coordinates of the nut's center hole and compared with the reference coordinates. The maximum deviation is 6mm, which is within the allowable range of ±10mm. This function ensures that the steel bars can pass through smoothly during on-site hoisting and avoids rework.

[0043] Five-party signatures and quality inspection report: Due to the importance of the component, the signature of the construction unit representative is added to form a five-party certification; in addition to the regular content, the verification report also includes the original measurement data of the laser tracker (LAS format), the acceleration curve of the hoisting process, and the component's "digital twin" QR code. Scanning the code allows you to view the component's full life cycle data from production to hoisting.

[0044] Example 4: Agile Verification Model for Small-Scale, High-Frequency Materials in Municipal Emergency Repair Projects Project Background: Emergency repair of a burst water supply pipeline in a city requires a continuous 72-hour supply of materials such as coarse sand, gravel, cement, and corrugated pipes, with more than 30 truck deliveries per day and a single purchase amount of less than 5,000 yuan. Traditional verification methods cannot meet the timeliness requirements.

[0045] Agile design: The system enables a "small-amount, high-frequency emergency channel," with core adjustments including: Data simplification: Purchase orders only retain material type, estimated amount, and supplier credit authorization limit, eliminating heavy data such as 3D point cloud, and the baseline database is simplified to 9 fields; Simplified labeling: The single-item coding is cancelled and replaced with the vehicle electronic waybill (e-Waybill) mode. The driver's mobile APP automatically generates a dynamic QR code, which includes the trip origin, destination and cargo snapshot. Data collection is simplified: only weight verification and environmental verification are retained, and image collection is simplified to two self-taken photos by the driver (one of the cargo in the truck and one of the empty truck), which must include landmark buildings in the background to prove the authenticity of the location; Simplified signature: The on-site supervisor's signature is cancelled and replaced with the driver's electronic signature + random verification by the construction party. One vehicle out of every 10 vehicles will be randomly selected for standard mode verification, and the results of the random inspection will be used to correct the supplier's credit score. Simplified evidence storage: Blockchain evidence storage only stores the hash value and key metadata of the verification report. Original documents such as original photos and weighing slips are automatically archived to cold storage after 7 days of cloud storage, reducing on-chain storage costs.

[0046] Intelligent anti-cheating measures: To prevent drivers from cheating, the system implements multiple risk control rules: ① GPS trajectory anomaly detection: if a vehicle deviates from the optimal route by more than 20%, it is marked as abnormal; ② Image background AI recognition: to ensure that photos are taken on-site rather than pre-stored photos; ③ Weight deviation clustering analysis: if the same driver has three consecutive negative deviations with similar absolute values, manual inspection is triggered; ④ When the supplier's daily cumulative amount exceeds 20,000 yuan, the system automatically switches back to standard verification mode.

[0047] Implementation Results: Under this model, the verification time for a single vehicle was reduced to 85 seconds, and a total of 89 vehicles were processed in 72 hours. Only one misjudgment occurred due to driver error, which was corrected after manual review. After the dynamic adjustment of supplier credit scores, the sampling rate of high-quality suppliers (score > 750) was reduced to 5%, while the sampling rate of low-quality suppliers was increased to 50%, achieving adaptive risk management.

[0048] Technical effects and quantitative indicators Through the above-described systematic technical solution, this invention has achieved significant technical and economic benefits in multiple pilot projects: Significant improvement in verification accuracy: The multi-dimensional cross-validation mechanism reduces the quantity error rate from 3%-5% in traditional manual verification to below 0.2%, the specification identification accuracy rate reaches 99.5%, the chemical composition misjudgment rate is less than 0.1%, and the overall verification accuracy rate is ≥99.8%. Work efficiency has been significantly improved: the average verification time for bulk materials (steel bars, cement) has been reduced from 40 minutes to 8 minutes, an efficiency increase of 80%; the verification time for small-quantity, high-frequency materials is less than 90 seconds in agile mode, the on-site vehicle turnover rate has increased by 3 times, and the congestion index at the construction site entrance has decreased by 85%. Legally reliable accountability: Blockchain-based evidence storage technology ensures that verification data is tamper-proof, and four-party digital signatures have the legal effect of the Electronic Signature Law, reducing the dispute rate by 95% and achieving a 100% evidence acceptance rate in legal proceedings; Significantly reduced management costs: Material losses due to manual inventory errors are reduced by approximately 3%-5% annually, labor costs for dispute resolution are reduced by approximately 70%, and overall material management costs decrease by 15%-20%; Supply chain collaboration optimization: The supplier credit evaluation model has accumulated more than 100 suppliers and 200,000 verification data points. The credit score has a positive correlation of 0.92 with material quality, supporting the optimization of procurement decisions and increasing the proportion of procurement from high-quality suppliers by 30%. Data asset value enhancement: The entire chain of data is accumulated into digital assets for the project, supporting material consumption prediction (error ±5%), intelligent replenishment reminders (lead time 3-5 days), and dynamic cost analysis, providing a data foundation for refined project management.

[0049] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0050] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A multi-dimensional intelligent verification method for the receipt of building materials, characterized in that, Includes the following steps: step S10: During the purchase order generation stage, based on the BIM model or ERP system, extract the material's unique code, specification parameters (length, width, height, diameter, wall thickness), theoretical weight (calculated based on material density and national standard tolerance), three-dimensional feature data (point cloud model, key dimensions of CAD drawings, surface texture feature vectors), standard range of chemical composition, and supplier quality certification information. After being encrypted using the national cryptographic SM4 algorithm, a structured benchmark data packet is formed and securely transmitted to the cloud verification platform for distributed storage via the TLS1.3 protocol. Step S20: During the shipment process, the supplier attaches a machine-readable identifier to the cargo unit or transport carrier through a coding terminal. The machine-readable identifier is at least one of a dynamic encrypted QR code, a UHF RFID anti-metal tag, or an NFC chip. The identifier embeds the order number, material code, production batch number, encrypted verification code, and shipment time stamp (time, latitude and longitude). At the same time, the shipment details, carrier vehicle information, driver digital identity certificate, and on-the-way monitoring parameters (temperature, humidity, vibration) are written into the blockchain evidence storage system through a smart contract interface. Step S30: After the goods arrive at the designated area of ​​the construction site, the verification personnel use a mobile terminal (industrial-grade rugged tablet or smartphone) with positioning, communication and imaging functions to scan or sense the machine-readable identifier and trigger the verification request. The mobile terminal obtains the corresponding baseline data packet and historical verification record from the cloud verification platform through 5G / 4G or Wi-Fi 6 network, and automatically verifies whether the current GPS coordinates are within the geographical area of ​​the project's electronic fence. Step S40: Simultaneously acquire actual multi-dimensional state data of the goods through no less than three types of heterogeneous sensing and acquisition devices, including: intelligent weighing device to acquire actual weight data, multi-view vision and structured light imaging device to acquire the appearance image and three-dimensional point cloud data of the goods, laser ranging and AR measurement device to acquire on-site specification measurement data, environmental parameter sensor to acquire transportation micro-environment data, and portable spectrometer to acquire material composition rapid detection data. Step S50: Upload the collected multi-dimensional status data to the cloud verification platform. After data preprocessing and quality assessment based on edge computing nodes, the platform runs a multi-dimensional cross-comparison intelligent algorithm. The algorithm includes weight deviation analysis, specification and size registration comparison, appearance defect identification, environmental parameter compliance judgment and component spectrum matching. It calculates the deviation rate and confidence score of each item. When all comparison items meet the preset dynamic threshold range and the comprehensive confidence score is ≥95%, the verification is deemed to be passed. Step S60: After the verification is passed, the system automatically generates an immutable verification report containing a verification timestamp, precise geographical coordinates (longitude, latitude, and elevation), actual status data, detailed comparison results of each dimension, device status log, multi-party electronic signatures, and hash verification code. After the verification report is signed by the national cryptographic SM2 algorithm, its digital fingerprint is permanently written into the FISCO BCO or Hyperledger Fabric consortium blockchain notarization system through a smart contract, generating a unique notarization number and simultaneously pushing it to the digital wallets of each participating party.

2. The multi-dimensional intelligent verification method for signing for receipt of building materials according to claim 1, characterized in that: The three-dimensional feature data in step S10 includes point cloud model data of material standard parts (containing at least 10,000 valid points) pre-acquired by a three-dimensional laser scanner, key dimension parameter vectors (length, width, diagonal, roundness, flatness), surface texture gray-level co-occurrence matrix (GLCM) feature code, and IFC format geometric data exported based on the BIM model. The reference data package uses the IPFS distributed file system to store large-capacity three-dimensional model files.

3. The multi-dimensional intelligent verification method for signing for receipt of building materials according to claim 1, characterized in that: The machine-readable identifier in step S20 is a dynamically encrypted QR code. The QR code uses the QRCode 2005 international standard encoding and has a fault tolerance level of H (30%). Its data content structure includes: header identifier (4 bytes), order number (32 bytes), material code (24 bytes), production batch number (16 bytes), timestamp (8 bytes), random salt (8 bytes), and SM4-CBC mode encryption verification code (32 bytes). The key of the encryption verification code is derived from the supplier's digital certificate private key.

4. The multi-dimensional intelligent verification method for signing for receipt of building materials according to claim 1, characterized in that: In step S40, the intelligent weighing device is an IoT weighbridge (with a digital sensor interface of RS485 or CAN bus, supporting Modbus RTU protocol) or a hook scale with Bluetooth 5.2 data transmission function (measuring range from 0.5 tons to 50 tons, accuracy class OIML C3). The image acquisition device is an industrial-grade multi-view vision inspection module (containing at least one RGB camera, one depth camera and one infrared fill light, with a resolution of not less than 12 million pixels). The parameter recognition device is a handheld laser rangefinder (accuracy ±1mm, measuring range 200m) or a virtual ruler system based on ARKit / ARCore technology.

5. The multi-dimensional intelligent verification method for signing for building materials as described in claim 1, characterized in that: The multi-dimensional cross-comparison intelligent algorithm in step S50 specifically includes: Weight verification module: Calculate the deviation rate between actual weight and theoretical weight ΔW=(W_actual - W_theoretical) / W_theoretical×100%. When |ΔW|≤3%, it is judged as qualified. When 3%<|ΔW|≤5%, it is marked as a warning and triggers secondary sampling. When |ΔW|>5%, it is directly judged as unqualified. Specification verification module: The improved ICP (Iterative Closest Point) algorithm is used to register the 3D point cloud data collected on site with the reference point cloud model, calculate the root mean square error RMSE and the critical dimension deviation vector ΔD. When RMSE ≤ 5mm and all ΔD ≤ 2%, it is judged as qualified. Appearance verification module: It identifies damage, deformation, corrosion and contamination defects in images through a lightweight convolutional neural network model (such as MobileNetV3 or YOLOv8n) and outputs a goodness score S (0-100 points). When S≥85 points, it is judged as qualified. Environmental verification module: For cement, admixtures, and coatings, compare the temperature and humidity data during transportation to see if they exceed the stored standard curve; if they do, the material will be automatically rejected. Composition verification module: For key structural materials, the content of major chemical elements is detected on-site using a portable XRF spectrometer or LIBS laser-induced breakdown spectrometer, and the matching degree with the benchmark composition range is calculated; The system will automatically generate a verification pass instruction only if all five of the above verifications meet the qualification standards or the warning is within an acceptable range; otherwise, it will proceed to the manual arbitration process.

6. The multi-dimensional intelligent verification method for signing for receipt of building materials according to claim 5, characterized in that, When the weight deviation rate enters the warning range (3% < |ΔW| ≤ 5%) or the appearance integrity score is in the critical range (80 ≤ S < 85), the system automatically triggers a secondary refined verification process. The secondary refined verification process includes: automatically starting the weighbridge to weigh repeatedly (at least 3 times and taking the average value), calling up the high-resolution camera to increase the number of image sampling points to no less than 20, expanding the laser ranging sampling range to all six sides of the goods, notifying the purchaser and supervisor to access online review via remote video, and raising the verification confidence threshold to 98%.

7. The multi-dimensional intelligent verification method for signing for receipt of building materials according to claim 1, characterized in that: The multi-party electronic signature in step S60 is generated using hierarchical deterministic key (HD Wallet) technology, including four levels of digital signatures: supplier authorized representative, carrier driver, construction party material clerk, and supervisor witness. Each level of signature uses the national cryptographic SM2 elliptic curve algorithm to generate a 64-byte signature value and attaches a timestamp authorization certificate (TSA). The verification public key of the digital signature is stored in the DID (decentralized identity) registry of the blockchain evidence storage system.

8. The multi-dimensional intelligent verification method for signing for receipt of building materials according to claim 1, characterized in that, Between steps S40 and S50, there is also a data credibility pre-evaluation step S45: acquiring the real-time status parameters of each sensor acquisition device, including the calibration validity period and real-time zero drift value of the intelligent weighing device, the pixel resolution and light intensity of the image acquisition device, the reflected signal intensity of the laser rangefinder, the network communication quality (latency, packet loss rate), and the battery level of the mobile terminal. When any device status parameter is lower than the preset credibility threshold, the corresponding data source is marked as low credibility and a device self-test or manual intervention prompt is triggered.

9. The multi-dimensional intelligent verification method for signing for receipt of building materials according to claim 1, characterized in that: The cloud-based verification platform incorporates a supplier credit evaluation module based on federated learning and dynamic Bayesian networks. This module dynamically calculates supplier credit scores (0-1000 points) using a sliding time window algorithm, based on multiple dimensions of indicators such as historical verification pass rate, average weight deviation trend, appearance defect rate, on-time delivery rate, and data uploading timeliness. In subsequent verifications, the comparison threshold is adaptively adjusted based on the credit score. For high-credit suppliers with a credit score ≥800, the weight deviation threshold is relaxed to ±4%, while for low-credit suppliers with a credit score ≤600, the threshold is tightened to ±2%, and the sampling rate is increased.

10. A multi-dimensional intelligent verification method for signing for receipt of building materials according to any one of claims 1 to 9, characterized in that, Its application covers the full-category verification of nine types of building materials, including steel bars, cement, sand and gravel aggregates, commercial concrete, precast components, waterproof membranes, architectural coatings, curtain wall panels, and pipes and cables. For bagged cement and dry powder mortar, it adds automatic bag breakage rate identification and safe stacking height detection functions. For liquid materials (such as water-reducing agents and coatings), it adds liquid level gauge reading identification and leak prevention detection functions. For valuable decorative materials (such as stone and tiles), it adds color difference comparison and scratch detection functions. Moreover, the cloud verification platform supports automatic association with the project-level BIM model and writes back the status of verified materials to the BIM5D progress management system in real time.

Images