A dangerous chemical spectrum analysis system and management method for preventing replacement

Abstract

This invention relates to the field of hazardous chemical management technology, providing a hazardous chemical spectral analysis system and control method to prevent substitution. The system includes a basic information and physical quantity acquisition module, a dynamic adaptive spectral detection module, and a central control server. The system utilizes zoom control to acquire first background spectral data focused on the outer wall of the container and second mixed spectral data focused on the internal liquid. A differential feature extraction algorithm is used to remove interference from the container material, generating pure reagent spectral feature data and establishing a baseline profile. During verification, the server calculates the weight change difference before and after return, along with spectral similarity values, and performs calibration for natural drift using an environmental attenuation compensation model. A multi-dimensional data anti-counterfeiting judgment model is then used to output a safety control signal. This invention achieves non-destructive verification of component authenticity, effectively solving the problem of identifying malicious substitution by equal weight and deterioration contamination.

Description

Technical Field

[0001] This invention relates to the field of hazardous chemical management technology, specifically to a hazardous chemical spectral analysis system and control method that prevents substitution. Background Technology

[0002] Hazardous chemicals are widely used in scientific research, medical treatment, and industrial production, and their full life-cycle compliance management is a crucial aspect of safety control. Currently, the industry's intelligent management systems for hazardous chemicals mainly adopt a control model that combines identity verification, barcode traceability, and electronic weighing. This model assigns a unique digital identifier to chemical containers and verifies the weight at various stages of circulation, such as warehousing, issuance, and return, thereby recording reagent consumption and updating ledger data.

[0003] However, existing technologies have significant limitations in their underlying logic for practical applications. Existing verification mechanisms can only perform peripheral verification of the packaging containers and macroscopic physical weight of chemicals, lacking non-destructive testing methods for the compositional characteristics of the chemical substances themselves inside the containers.

[0004] This leads to the potential for hidden substitution during the transfer of hazardous chemicals. For example, after receiving a specific reagent, an operator might pour out a portion and refill the original container with another conventional liquid of similar weight and appearance. Because the identification markings on the container remain unchanged and the returned weight conforms to the system's logical verification, the existing management system often misinterprets this as a normal return. This substitution not only results in serious discrepancies between records and actual goods but also greatly increases the risk of illegal loss of controlled chemicals.

[0005] Introducing conventional spectral analysis techniques for substance composition verification without damaging the original packaging presents complex interference factors in engineering practice, making direct application difficult. On one hand, the materials and thicknesses of containers for different chemicals vary significantly, easily leading to severe scattering and distortion of the spectral excitation signal. On the other hand, some chemicals undergo natural volatilization, moisture absorption, or slight degradation during use after opening, resulting in an objective natural drift in their spectral characteristics upon return compared to their initial state. Using conventional absolute value comparison models would produce a high frequency of false alarms.

[0006] In summary, this invention provides a spectral analysis system and control method for hazardous chemicals that prevent substitution, in order to solve the above-mentioned problems. Summary of the Invention

[0007] This invention provides a spectral analysis system and control method for hazardous chemicals that prevents substitution. By introducing dynamic adaptive spectral penetration detection technology and a multi-dimensional data cross-anti-counterfeiting judgment model, it solves the safety hazard problem that existing hazardous chemical management systems can only rely on packaging labels and macroscopic weight for external verification. This makes it impossible to detect the true composition of chemical substances inside the container without damaging the original packaging. As a result, high-risk reagents are easily maliciously replaced by liquids of equal weight and of different qualities, leading to illegal loss.

[0008] The specific technical solution of this invention is as follows:

[0009] A spectral analysis system for hazardous chemicals designed to prevent substitution includes:

[0010] The basic information and physical quantity acquisition module obtains the identification information of hazardous chemical containers and the initial total weight including the container;

[0011] The dynamic adaptive spectral detection module emits a detection beam into the container to acquire first background spectral data focused on the outer wall of the container and second mixed spectral data focused on the liquid inside the container;

[0012] The central control server is communicatively connected to the basic information and physical quantity acquisition module and the dynamic adaptive spectral detection module. The central control server has a built-in differential feature extraction algorithm and a multi-dimensional data anti-counterfeiting judgment model. The differential feature extraction algorithm performs subtraction calculation on the second mixed spectral data based on the first background spectral data to generate pure reagent spectral feature data. The pure reagent spectral feature data is bound to the identity information and the initial total weight to generate the initial digital baseline file of the substance.

[0013] During the verification of hazardous chemical returns, the central control server acquires the current total weight and the current pure reagent spectral characteristic data at the time of return. The multi-dimensional data anti-counterfeiting judgment model calculates the weight change difference between the current total weight and the initial total weight, and calculates the spectral similarity value between the current pure reagent spectral characteristic data and the pure reagent spectral characteristic data in the initial digital baseline file of the substance. Based on the weight change difference and the spectral similarity value, the corresponding safety control alarm signal is output.

[0014] As an improvement of the present invention, the dynamic adaptive spectral detection module includes a container parameter acquisition unit, an optical zoom unit, and a spectral transceiver unit; the container parameter acquisition unit retrieves the distance and thickness parameters of the outer wall of the container; the spectral transceiver unit emits the detection beam and receives the backscattered signal; the optical zoom unit adjusts the focal point of the detection beam according to the distance and thickness parameters, and sequentially locks the focus on a specified depth position within the surface material of the outer wall of the container and the liquid inside the container.

[0015] As an improvement of the present invention, the differential feature extraction algorithm extracts the signal intensity features of the first background spectral data, calculates an adaptive weight scaling factor, multiplies the first background spectral data by the adaptive weight scaling factor, performs differential subtraction calculation from the second mixed spectral data, and outputs the pure reagent spectral feature data.

[0016] As an improvement of the present invention, the system further includes an environmental data sensing module, which collects surface temperature and humidity data of the environment in which the hazardous chemical container is located in real time; the central control server has a built-in environmental attenuation compensation model, which extracts the exposure time parameter of the hazardous chemical after opening, and calculates the baseline offset and peak attenuation of the spectral characteristics in combination with the surface temperature and humidity data, and generates a spectral drift tolerance range that limits the range of natural physicochemical changes.

[0017] As an improvement of the present invention, the multidimensional data anti-counterfeiting judgment model extracts the absorbance values ​​of the current pure reagent spectral feature data and the pure reagent spectral feature data established in the archive at each feature wavenumber sampling point to construct two multidimensional feature vectors, calculates the inner product of the two multidimensional feature vectors, and divides the inner product by the product of the magnitudes of the two multidimensional feature vectors to output the cosine similarity value of the quantized waveform contour difference.

[0018] A method for controlling hazardous chemicals to prevent substitution includes the following steps:

[0019] S1. Obtain the identification information and initial total weight of hazardous chemical containers;

[0020] S2. Control the dynamic adaptive spectral detection module to acquire the first background spectral data focused on the outer wall of the container and the second mixed spectral data focused on the liquid inside the container. Subtract the first background spectral data through the differential feature extraction algorithm to generate pure reagent spectral feature data. Bind the identification information and the initial total weight with the pure reagent spectral feature data to generate the initial digital baseline file of the substance.

[0021] S3. Obtain the identification information of the hazardous chemical container at the time of return, its current total weight, and the current spectral characteristic data of the pure reagent;

[0022] S4. Calculate the weight change difference between the current total weight and the initial total weight, and calculate the spectral similarity value between the current pure reagent spectral feature data and the pure reagent spectral feature data in the corresponding substance's initial digital baseline file;

[0023] S5. Based on the cross-comparison results of the weight change difference and the spectral similarity value, output a safety control signal.

[0024] As an improvement of the present invention, the steps of acquiring spectral data in steps S2 and S3 specifically include:

[0025] Obtain the physical distance and thickness parameters from the probe to the outer wall of the container;

[0026] Based on the physical distance and thickness parameters, the optical zoom unit is driven to lock the first focal plane on the outer wall of the container and emits a spectral detection beam to collect the first background spectral data;

[0027] The optical zoom unit is driven to advance the optical focus of the probe beam into the container to a specified penetration depth, lock the second focal plane, and acquire the second mixed spectral data.

[0028] As an improvement of the present invention, before calculating the spectral similarity value in step S4, the surface temperature data of the container at the time of return and the environmental temperature and humidity data recorded during use are read, and the stored environmental attenuation compensation model is input to perform reverse thermal drift correction and natural attenuation baseline calibration calculation on the current pure reagent spectral characteristic data.

[0029] As an improvement of the present invention, the correspondence logic between the cross-comparison result and the security control signal in step S5 specifically includes:

[0030] When the current total weight is less than the initial total weight and the spectral similarity value is greater than the benchmark similarity threshold, a normal consumption control signal is output.

[0031] When the difference between the current total weight and the initial total weight is approximately zero within the calibrated physical tolerance range and the spectral similarity value is less than the benchmark similarity threshold, a malicious substitution cheating alarm signal is output.

[0032] When the current total weight is greater than the initial total weight and the spectral similarity value is less than the benchmark similarity threshold, an abnormal mixing alarm signal is output.

[0033] As an improvement to the present invention, the logic for corresponding the cross-comparison results with the security control signal further includes:

[0034] When the current total weight is less than the initial total weight and the spectral similarity value is less than the benchmark similarity threshold, the environmental temperature and humidity data and the exposure time parameter after opening the lid are obtained and input into the environmental attenuation compensation model for offset comparison; when the comparison offset is greater than the spectral offset tolerance limit allowed by natural attenuation, an early warning signal for the disposal of reagent contamination failure and deterioration is output.

[0035] In this invention, the basic information and physical quantity acquisition module typically includes barcode scanning devices with identification functions and high-precision electronic weighing devices. The aim is to quickly establish an initial mapping relationship between the unique digital identity of hazardous chemicals and their macroscopic physical mass without damaging the packaging, providing physical benchmark data for subsequent cross-validation.

[0036] In this invention, the optical zoom unit inside the dynamic adaptive spectral detection module can employ an optical lens group driven by a precision micro stepper motor or piezoelectric ceramic. After acquiring the three-dimensional parameters of the external container through the container parameter acquisition unit, the system drives the optical lens group to change its relative position, thereby precisely controlling the focal penetration depth of the excitation beam. This enables the detection beam to achieve spatial transitions from focusing on the outer wall material to focusing on the internal liquid being measured, overcoming the engineering challenge of existing spectrometers being limited by fixed focal lengths and unable to accommodate diverse packaging containers.

[0037] In this invention, the core function of the differential feature extraction algorithm lies in the digital denoising of optical signals. When a light beam penetrates a container, it inevitably excites a strong background spectrum of the container material, masking the weak intrinsic signals of the internal liquid. This algorithm uses the pure background spectral data collected when focusing solely on the outer wall as a reference substrate. After calculating a reasonable scaling ratio based on the signal characteristic intensity, it separates this reference substrate from the mixed internal and external spectral signals. This ensures that the final spectral data used for similarity comparison only reflects the chemical bond vibration characteristics of the internal hazardous chemicals, greatly improving the sensitivity and accuracy of similarity identification.

[0038] In this invention, an environmental degradation compensation model is specifically designed to address the pain point of spectral distortion caused by the natural aging of chemical reagents and environmental interference. Because many hazardous chemicals are volatile or hygroscopic, their concentration undergoes a physical change after opening, leading to a natural shift in the intensity of spectral characteristic peaks or baseline upon return. This model retrieves pre-stored physicochemical property curves of specific chemicals and combines them with dynamically recorded on-site temperature, humidity, and cumulative exposure time. At the algorithm level, it appropriately relaxes or reverses the weighting of the similarity comparison benchmark, ensuring the system has strong adaptability to real-world scenarios and avoiding misjudging normal chemical reagent deterioration as malicious substitution or cheating.

[0039] In this invention, a multi-dimensional data anti-counterfeiting judgment model constructs a rigorous logical cross-loop network. Its innovation lies in abandoning single-dimensional threshold judgment and instead using a matrix-style fusion analysis of physical and chemical dimensions. In actual control logic: normally consumed reagents inevitably experience a decrease in weight while the characteristics of the remaining substance components remain unchanged; however, cheaters attempting to steal high-risk reagents, in order to conceal their actions, usually inject an equal weight of liquid to maintain weight balance. At this point, the spectral fingerprint of the mixture will inevitably undergo a severe abrupt change. This model accurately captures such logical paradoxes, enabling it to instantly output malicious substitution alarms and plug various high-IQ cheating loopholes.

[0040] Compared with the prior art, the present invention has the following beneficial effects:

[0041] 1. This invention achieves non-destructive testing of reagents by penetrating containers of varying thicknesses and materials through the synergistic cooperation of a dynamic adaptive spectral detection module and a differential feature extraction algorithm. By dynamically zooming to acquire two sets of spectra of the container's outer wall and the internal liquid, and performing digital differential processing, it effectively reduces background fluorescence and optical scattering interference from the container wall, obtaining the intrinsic spectral characteristics of the reagent with high purity. This significantly improves upon the technical deficiency of traditional management systems that can only verify the packaging container but not the internal composition, providing a reliable technical means to prevent substitution of reagents at the physical detection level.

[0042] 2. This invention significantly improves the reliability of the system in practical applications by introducing an environmental attenuation compensation model and a multi-dimensional data anti-counterfeiting judgment matrix. The system comprehensively considers the cosine similarity between weight change and spectral profile, and combines environmental parameters such as reagent opening exposure time and on-site temperature and humidity for algorithm correction, effectively mitigating false alarms caused by natural evaporation and moisture absorption of reagents. At the same time, the established cross-judgment logic covers mainstream circulation states such as normal consumption, malicious replacement, abnormal mixing, and deterioration and contamination, realizing accurate traceability of hazardous chemicals throughout their entire life cycle. Attached Figure Description

[0043] Figure 1 This is the logic diagram for the full life cycle management of hazardous chemicals in this invention.

[0044] Figure 2 This is a flowchart of the adaptive spectral detection and signal purification algorithm of the present invention.

[0045] Figure 3 This is a flowchart of the multidimensional data cross-referencing anti-counterfeiting determination matrix of the present invention. Detailed Implementation

[0046] The embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and should not be construed as limiting the scope of the invention.

[0047] like Figure 1-3 As shown, the present invention provides a spectral analysis system for hazardous chemicals that is resistant to substitution, comprising:

[0048] The basic information and physical quantity acquisition module obtains the identification information of hazardous chemical containers and the initial total weight including the container;

[0049] The dynamic adaptive spectral detection module emits a detection beam into the container to acquire first background spectral data focused on the outer wall of the container and second mixed spectral data focused on the liquid inside the container.

[0050] The central control server communicates with the basic information and physical quantity acquisition module and the dynamic adaptive spectral detection module. The central control server has a built-in differential feature extraction algorithm and a multi-dimensional data anti-counterfeiting judgment model. The differential feature extraction algorithm performs subtraction calculation on the second mixed spectral data based on the first background spectral data to generate pure reagent spectral feature data. The pure reagent spectral feature data is bound with the identity information and the initial total weight to generate the initial digital baseline file of the substance.

[0051] During the verification of hazardous chemicals upon return, the central control server obtains the current total weight and the current spectral characteristic data of the purified reagent at the time of return. The multi-dimensional data anti-counterfeiting judgment model calculates the weight change difference between the current total weight and the initial total weight, and calculates the spectral similarity value between the current spectral characteristic data of the purified reagent and the spectral characteristic data of the purified reagent in the initial digital baseline file of the substance. Based on the weight change difference and the spectral similarity value, the corresponding safety control alarm signal is output.

[0052] The dynamic adaptive spectral detection module includes a container parameter acquisition unit, an optical zoom unit, and a spectral transceiver unit. The container parameter acquisition unit retrieves the distance and thickness parameters of the container's outer wall. The spectral transceiver unit emits a detection beam and receives backscattered signals. The optical zoom unit adjusts the focus point of the detection beam according to the distance and thickness parameters, sequentially locking the focus on a specified depth position within the material of the container's outer wall surface and the liquid inside the container.

[0053] The differential feature extraction algorithm extracts the signal intensity features of the first background spectral data, calculates the adaptive weight scaling factor, multiplies the first background spectral data by the adaptive weight scaling factor, performs differential subtraction calculation from the second mixed spectral data, and outputs the pure reagent spectral feature data.

[0054] The system also includes an environmental data sensing module that collects surface temperature and humidity data of the environment in which the hazardous chemical containers are located in real time. The central control server has a built-in environmental attenuation compensation model. The environmental attenuation compensation model extracts the exposure time parameters of the hazardous chemicals after opening and combines them with surface temperature and humidity data to calculate the baseline shift and peak attenuation of spectral characteristics, and generates a spectral drift tolerance range that limits the range of natural physicochemical changes.

[0055] The multidimensional data anti-counterfeiting judgment model extracts the absorbance values ​​of the current pure reagent spectral feature data and the pure reagent spectral feature data established in the archive at each feature wavenumber sampling point to construct two multidimensional feature vectors. The inner product of the two multidimensional feature vectors is calculated, and the inner product is divided by the product of the magnitudes of the two multidimensional feature vectors to output the cosine similarity value of the quantized waveform contour difference.

[0056] This invention also provides a method for controlling hazardous chemicals to prevent substitution, comprising the following steps:

[0057] S1. Obtain the identification information and initial total weight of hazardous chemical containers;

[0058] S2. Control the dynamic adaptive spectral detection module to acquire the first background spectral data focused on the outer wall of the container and the second mixed spectral data focused on the liquid inside the container. Subtract the first background spectral data through the differential feature extraction algorithm to generate pure reagent spectral feature data. Bind the identification information and the initial total weight to the pure reagent spectral feature data to generate the initial digital baseline file of the substance.

[0059] S3. Obtain the identification information of the hazardous chemical container at the time of return, its current total weight, and the current spectral characteristic data of the pure reagent;

[0060] S4. Calculate the weight change difference between the current total weight and the initial total weight, and calculate the spectral similarity value between the current pure reagent spectral characteristic data and the pure reagent spectral characteristic data in the corresponding initial digital baseline file.

[0061] S5. Based on the cross-comparison results of the weight change difference and the spectral similarity value, output the safety management and control signal.

[0062] The steps for obtaining spectral data in steps S2 and S3 specifically include:

[0063] Obtain the physical distance and thickness parameters from the probe to the outer wall of the container;

[0064] Based on physical distance and thickness parameters, the optical zoom unit is driven to lock the first focal plane on the outer wall of the container and emits a spectral detection beam to collect the first background spectral data.

[0065] The optical zoom unit drives the optical focus of the probe beam to a specified penetration depth into the container, locks the second focal plane, and acquires the second mixed spectral data.

[0066] Before calculating the spectral similarity value in step S4, the surface temperature data of the container at the time of return and the environmental temperature and humidity data recorded during use are read. The stored environmental attenuation compensation model is then used to perform reverse thermal drift correction and natural attenuation baseline calibration calculations on the current pure reagent spectral characteristic data.

[0067] The specific logic for the correspondence between the cross-comparison results and the safety management control signals in step S5 includes:

[0068] When the current total weight is less than the initial total weight and the spectral similarity value is greater than the baseline similarity threshold, output a normal consumption control signal;

[0069] When the difference between the current total weight and the initial total weight is approximately zero within the calibrated physical tolerance range and the spectral similarity value is less than the benchmark similarity threshold, a malicious substitution cheating alarm signal is output.

[0070] When the current total weight is greater than the initial total weight and the spectral similarity value is less than the benchmark similarity threshold, an abnormal mixing alarm signal is output.

[0071] The logic for corresponding cross-comparison results with safety management and control signals also includes:

[0072] When the current total weight is less than the initial total weight and the spectral similarity value is less than the benchmark similarity threshold, the environmental temperature and humidity data and the exposure time parameter after opening the lid are input into the environmental attenuation compensation model for offset comparison; when the comparison offset is greater than the spectral offset tolerance limit allowed by natural attenuation, an early warning signal for the disposal of reagent contamination failure and deterioration is output.

[0073] Example 1: The purpose of this example is to verify the system's ability to identify and control the compliant flow and normal consumption scenarios of hazardous chemicals. Taking the flow of anhydrous ethanol, a frequently used substance in the laboratory, as an example. During the warehousing and filing stage, the basic information and physical quantity acquisition module records the batch of reagent information by identifying the container label and measures the initial total weight as 502.4g. The dynamic adaptive spectral detection module then performs a baseline scan to acquire the initial pure reagent spectral characteristic data of the bottle and stores it in the central control server. After the operator has used the reagent for a period of time, they return it, and the electronic weighing device reports the current total weight as 447.8g. The dynamic adaptive spectral detection module responds automatically; its ranging sensor unit measures the wall thickness of the polyester container as 1.18mm, and the optical zoom unit locks the detection focus on the outer wall of the container to acquire the first background spectral data. Then, it drives the lens group to advance the focus 3.2mm into the container to acquire the second mixed spectral data. The differential feature extraction algorithm, after adaptive weight scaling to subtract background interference, extracts the current pure reagent spectral characteristic data. The results showed that the multidimensional data anti-counterfeiting judgment model calculated a weight change difference of 54.6g. The cosine similarity between the current pure reagent spectral characteristic data and the baseline file was 0.988, which is higher than the benchmark similarity threshold of 0.95. Based on the logical conditions that the current total weight is less than the initial total weight and the spectral similarity value is greater than the benchmark similarity threshold, the system judgment matrix determined that the circulation process was a compliant normal consumption, and then output a normal consumption control signal and automatically updated the warehouse ledger.

[0074] Example 2: The purpose of this example is to verify the system's ability to accurately block cheating attempts involving the malicious substitution of liquids of equal weight but different compositions. Detection was performed against the illegal substitution of controlled concentrated nitric acid. During the initial record-keeping phase, the initial total weight of the concentrated nitric acid reagent bottle was recorded as 1000.6g, and the initial pure reagent spectral characteristic data was simultaneously extracted and recorded. After receiving the reagent, the cheater attempted to replace it with an equal weight of a mixture of water and acidic waste liquid to conceal their actions. During the return verification phase, the physical quantity acquisition module obtained a current total weight of 1000.4g. The dynamic adaptive spectral detection module penetrated the 2.1mm thick glass bottle wall to obtain the current pure reagent spectral characteristic data. The results showed that the multidimensional data anti-counterfeiting judgment model calculated that the difference between the current total weight and the initial total weight was approximately zero within the calibrated physical tolerance range. However, due to the qualitative change in the chemical composition inside the bottle, the spectral similarity value of the returned substance plummeted to 0.285, far below the benchmark similarity threshold of 0.95. The system's judgment matrix immediately detected the logical paradox between weight balance and composition shift. Based on the condition that the current total weight is approximately constant and the spectral similarity value is less than the benchmark similarity threshold, it determined that malicious substitution cheating had occurred using liquids of equal weight but different compositions. The system then output a malicious substitution cheating alarm signal, linked the security control module to lock the storage location, and pushed an operation anomaly warning to the management terminal.

[0075] Example 3: The purpose of this example is to verify the crucial corrective role of the environmental attenuation compensation model in avoiding false positives caused by natural physicochemical changes in reagents. Under experimental conditions, a bottle of hygroscopic anhydrous methanol was continuously exposed for 2.5 hours with the cap open at 82% relative humidity and an ambient temperature of 26.4°C. The initial total weight upon warehousing was 401.3g, and during the return phase, the physical quantity acquisition module measured a decrease to 376.5g. Due to the intrusion of environmental moisture, the spectral characteristic data of the current pure reagent extracted by the dynamic adaptive spectral detection module showed a significant shift in the water peak characteristic range. The initially calculated spectral similarity value was only 0.912, lower than the preset baseline similarity threshold of 0.95. At this point, the operation triggered the environmental attenuation compensation model. The system retrieved the temperature and humidity records for that period and the natural hygroscopic characteristic curve of methanol, calculated the spectral drift tolerance range under this condition, and performed reverse thermal drift correction and natural attenuation baseline calibration calculations on the current pure reagent spectral characteristic data. The results showed that after dynamic correction by the model, the spectral similarity value recovered to above the baseline similarity threshold. The multi-dimensional data anti-counterfeiting judgment model, combined with the physical fact that the current total weight is indeed less than the initial total weight, ultimately determines that the spectral shift is a reasonable change within the allowable range of natural attenuation. The system outputs a normal consumption control signal, effectively avoiding false alarms caused by environmental interference.

[0076] Example 4: The purpose of this example is to verify the system's adaptive penetration detection capability for dark, complex containers and its ability to identify and warn of abnormal material contamination. The target was a high-value chemical reagent stored in a dark brown, light-proof glass bottle with a wall thickness of 2.45 mm, with an initial total weight of 601.2 g. During the return process, the dynamic adaptive spectral detection module precisely locked the first focal plane inside the glass material through the optical zoom unit, acquiring the strong background signal generated by the dark container; then, the second focal plane was advanced to a depth of 5.5 mm inside the container to acquire the second mixed spectral data. The differential feature extraction algorithm used the scaled background signal to perform subtraction calculations, successfully removing the strong optical interference from the dark container to extract the spectral feature data of the pure reagent. At the same time, the physical quantity acquisition module weighed the sample and found that the current total weight at the time of return reached 632.8 g. The results show that the spectral similarity value calculated by the multidimensional data anti-counterfeiting judgment model is 0.743. Based on the condition that the current total weight is greater than the initial total weight and the spectral similarity value is less than the benchmark similarity threshold, the system's judgment logic matrix determines that not only has the composition inside the container changed, but additional non-original substances have also been abnormally mixed in. The system then outputs an abnormal mixing alarm signal and generates a waste disposal warning, guiding on-site management personnel to transfer the contaminated reagent to the waste liquid disposal stage for safe isolation.

[0077] The embodiments of the present invention are given for the purposes of illustration and description. Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A spectral analysis system for hazardous chemicals that prevents substitution, characterized in that, include: The basic information and physical quantity acquisition module obtains the identification information of hazardous chemical containers and the initial total weight including the container; The dynamic adaptive spectral detection module emits a detection beam into the container to acquire first background spectral data focused on the outer wall of the container and second mixed spectral data focused on the liquid inside the container; The central control server is communicatively connected to the basic information and physical quantity acquisition module and the dynamic adaptive spectral detection module. The central control server has a built-in differential feature extraction algorithm and a multi-dimensional data anti-counterfeiting judgment model. The differential feature extraction algorithm performs subtraction calculation on the second mixed spectral data based on the first background spectral data to generate pure reagent spectral feature data. The pure reagent spectral feature data is bound to the identity information and the initial total weight to generate the initial digital baseline file of the substance. During the verification of hazardous chemical returns, the central control server acquires the current total weight and the current pure reagent spectral characteristic data at the time of return. The multi-dimensional data anti-counterfeiting judgment model calculates the weight change difference between the current total weight and the initial total weight, and calculates the spectral similarity value between the current pure reagent spectral characteristic data and the pure reagent spectral characteristic data in the initial digital baseline file of the substance. Based on the weight change difference and the spectral similarity value, the corresponding safety control alarm signal is output.

2. The anti-substitution hazardous chemical spectroscopic analysis system according to claim 1, characterized in that, The dynamic adaptive spectral detection module includes a container parameter acquisition unit, an optical zoom unit, and a spectral transceiver unit. The container parameter acquisition unit retrieves the distance and thickness parameters of the outer wall of the container. The spectral transceiver unit emits the detection beam and receives backscattered signals. The optical zoom unit adjusts the focal point of the detection beam according to the distance and thickness parameters, and sequentially locks the focus on a specified depth position within the surface material of the outer wall of the container and the liquid inside the container.

3. The anti-substitution hazardous chemical spectroscopic analysis system according to claim 1, characterized in that, The differential feature extraction algorithm extracts the signal intensity features of the first background spectral data, calculates the adaptive weight scaling factor, multiplies the first background spectral data by the adaptive weight scaling factor, performs differential subtraction calculation from the second mixed spectral data, and outputs the pure reagent spectral feature data.

4. The anti-substitution hazardous chemical spectroscopic analysis system according to claim 1, characterized in that, The system also includes an environmental data sensing module, which collects surface temperature and humidity data of the environment in which the hazardous chemical container is located in real time; the central control server has a built-in environmental attenuation compensation model, which extracts the exposure time parameter of the hazardous chemical after opening, and calculates the baseline shift and peak attenuation of the spectral characteristics in combination with the surface temperature and humidity data, and generates a spectral drift tolerance range that limits the range of natural physicochemical changes.

5. The anti-substitution hazardous chemical spectroscopic analysis system according to claim 1, characterized in that, The multidimensional data anti-counterfeiting judgment model extracts the absorbance values ​​of the current pure reagent spectral feature data and the pure reagent spectral feature data established in the archive at each feature wavenumber sampling point to construct two multidimensional feature vectors. The inner product of the two multidimensional feature vectors is calculated, and the inner product is divided by the product of the magnitudes of the two multidimensional feature vectors to output the cosine similarity value of the quantized waveform contour difference.

6. A method for controlling hazardous chemicals to prevent substitution, characterized in that, Includes the following steps: S1. Obtain the identification information and initial total weight of hazardous chemical containers; S2. Control the dynamic adaptive spectral detection module to acquire the first background spectral data focused on the outer wall of the container and the second mixed spectral data focused on the liquid inside the container. Subtract the first background spectral data through the differential feature extraction algorithm to generate pure reagent spectral feature data. Bind the identification information and the initial total weight with the pure reagent spectral feature data to generate the initial digital baseline file of the substance. S3. Obtain the identification information of the hazardous chemical container at the time of return, its current total weight, and the current spectral characteristic data of the pure reagent; S4. Calculate the weight change difference between the current total weight and the initial total weight, and calculate the spectral similarity value between the current pure reagent spectral feature data and the pure reagent spectral feature data in the corresponding substance's initial digital baseline file; S5. Based on the cross-comparison results of the weight change difference and the spectral similarity value, output a safety control signal.

7. The method for controlling hazardous chemicals to prevent substitution according to claim 6, characterized in that, The steps of obtaining spectral data in steps S2 and S3 specifically include: Obtain the physical distance and thickness parameters from the probe to the outer wall of the container; Based on the physical distance and thickness parameters, the optical zoom unit is driven to lock the first focal plane on the outer wall of the container and emits a spectral detection beam to collect the first background spectral data; The optical zoom unit is driven to advance the optical focus of the probe beam into the container to a specified penetration depth, lock the second focal plane, and acquire the second mixed spectral data.

8. The method for controlling hazardous chemicals to prevent substitution according to claim 6, characterized in that, Before calculating the spectral similarity value in step S4, the surface temperature data of the container at the time of return and the environmental temperature and humidity data recorded during use are read and input into the stored environmental attenuation compensation model to perform reverse thermal drift correction and natural attenuation baseline calibration calculations on the current pure reagent spectral characteristic data.

9. The method for controlling hazardous chemicals to prevent substitution according to claim 6, characterized in that, The specific logic for the correspondence between the cross-comparison results and the security control signals in step S5 includes: When the current total weight is less than the initial total weight and the spectral similarity value is greater than the benchmark similarity threshold, a normal consumption control signal is output. When the difference between the current total weight and the initial total weight is approximately zero within the calibrated physical tolerance range and the spectral similarity value is less than the benchmark similarity threshold, a malicious substitution cheating alarm signal is output. When the current total weight is greater than the initial total weight and the spectral similarity value is less than the benchmark similarity threshold, an abnormal mixing alarm signal is output.

10. The method for controlling hazardous chemicals to prevent substitution according to claim 9, characterized in that, The logic for corresponding the cross-comparison results with the security management and control signals also includes: When the current total weight is less than the initial total weight and the spectral similarity value is less than the benchmark similarity threshold, the environmental temperature and humidity data and the exposure time parameter after opening the lid are obtained and input into the environmental attenuation compensation model for offset comparison; when the comparison offset is greater than the spectral offset tolerance limit allowed by natural attenuation, an early warning signal for the disposal of reagent contamination failure and deterioration is output.

Images