Method for detecting content of salidroside in rhodiola rosea extracting solution based on artificial intelligence

Abstract

The invention discloses a method for detecting the content of salidroside in a rhodiola rosea extracting solution based on artificial intelligence. The method comprises the following steps: S1, performing spectral scanning of different wavebands on an extracting solution sample in a plurality of near-infrared wavelength ranges; s2, performing multi-dimensional interference suppression processing and signal purification; s3, carrying out spectral feature compression calculation to obtain a salidroside feature intensity factor; s4, carrying out salidroside characteristic signal-to-noise ratio enhancement and purity index calculation; s5, predicting the content of salidroside and outputting a content prediction score; and S6, mapping the content prediction score into the final salidroside content. According to the method, spectral signal purification, nonlinear feature extraction, dynamic noise enhancement, artificial intelligence prediction modeling and a secondary recheck mechanism are utilized to realize adaptive analysis of extracting solutions from different sources, so that the detection accuracy and stability are remarkably improved, and rapid and traceable salidroside content detection can be realized under the condition that a large precise instrument is not needed.

Description

Technical Field

[0001] The present invention relates to the field of intelligent detection technology, and in particular to a method for detecting the content of salidroside in a Rhodiola rosea extract based on artificial intelligence. Background Art

[0002] The detection of salidroside content mainly relies on laboratory analysis methods such as high-performance liquid chromatography. Although the accuracy is relatively high, the detection is time-consuming and highly dependent on equipment, making it difficult to meet the needs of rapid on-site detection. Some rapid detection methods are based on the linear fitting relationship between spectral signals and fixed proportional coefficients to estimate the content. They are unable to cope with the nonlinear response characteristics in complex sample backgrounds. At the same time, they lack a dynamic correction mechanism for matrix differences between different batches and different process extracts, resulting in insufficient prediction accuracy and stability. In addition, existing methods generally lack a real-time review link for test results and are unable to perform adaptive corrections on rapid test results, resulting in the problem that result deviations cannot be corrected in a timely manner. Summary of the Invention

[0003] In view of the shortcomings of the existing technology, the present invention provides a method for detecting the content of salidroside in Rhodiola rosea extract based on artificial intelligence to solve the problems raised in the above background technology.

[0004] To achieve the above object, the present invention provides the following technical solutions: In a first aspect, the present invention provides a method for detecting the content of salidroside in a Rhodiola rosea extract based on artificial intelligence, comprising the following steps: S1. Use a multi-band synchronous acquisition device to perform spectral scanning of the extract sample in different bands within multiple near-infrared wavelength ranges, and record the total signal energy and spectral intensity of each band; S2. Based on the obtained total signal energy and the spectral intensity of each band, multi-dimensional interference suppression processing and signal purification are performed to obtain the effective signal intensity after purification; S3, using the purified effective signal intensity as input, and calculating the salidroside characteristic intensity factor through spectral feature compression; S4. Based on the characteristic intensity factor of salidroside, the characteristic signal-to-noise ratio enhancement and purity index of salidroside are calculated; S5. Using the calculated purity index, the prediction score model is used to predict the salidroside content and output a content prediction score; S6. Map the content prediction score to the final salidroside content and quickly output the test results.

[0005] To further optimize the technical solution, in step S1, a signal energy calculation model is established to uniformly measure the energy of spectral signals of different bands; The signal energy calculation model is as follows:

[0006] in, is the total signal energy; is the collected spectrum intensity of the i-th band; is the weight of the i-th band; is the total number of bands; The model integrates multi-band spectral information in a weighted manner to avoid occasional interference from a single band affecting the overall signal characteristics.

[0007] The technical solution is further optimized. In step S2, the extract contains non-target compounds including polysaccharides and flavonoids, which interfere with the spectral signal of salidroside. A dynamic characteristic peak comparison mechanism is introduced to determine the non-salidroside characteristic peaks through real-time comparison with the reference characteristic peak database and perform signal subtraction to purify the effective signal intensity.

[0008] To further optimize this technical solution, the calculation method of the effective signal intensity after purification is as follows:

[0009] in, is the effective signal intensity after purification; is the intensity of the jth interference peak detected; is the suppression coefficient of the j-th interference peak; is the total number of interference peaks; The model achieves pure salidroside spectral feature extraction by calculating the weighted energy of the interfering components and deducting them from the total signal.

[0010] Further optimizing the technical solution, in step S3, a characteristic intensity factor calculation model is constructed, and the nonlinear compression function is used for fusion by considering the nonlinear differences in the responses of different bands to salidroside; The characteristic intensity factor calculation model is as follows:

[0011] in, is the characteristic intensity factor of salidroside; is the characteristic intensity of the reference standard salidroside sample in band i; is the correction factor; The model output It is used to characterize the relative proximity between the purification signal and the standard feature, which directly reflects the potential content level of salidroside.

[0012] Further optimizing the technical solution, in step S4, when performing salidroside characteristic signal-to-noise ratio enhancement, a signal-to-noise ratio enhancement model is constructed to dynamically attenuate the residual low-frequency noise in the purified signal; The signal-to-noise ratio enhancement model is as follows:

[0013] in, To enhance signal strength; is the residual noise energy; is the noise attenuation coefficient.

[0014] To further optimize this technical solution, the signal-to-noise ratio enhancement model outputs the enhanced signal strength and residual noise energy and then calculates the purity index:

[0015] in, It is a purity index used to measure the purity of salidroside signal.

[0016] To further optimize this technical solution, in step S5, the prediction score model is as follows:

[0017] in, Score for content prediction; is the score magnification factor; is the sensitivity correction factor; is the theoretical purity value of the standard salidroside sample; Continuously optimize parameters through artificial intelligence training and , so that it can accurately output the salidroside content prediction score for different production batches of Rhodiola rosea extracts .

[0018] Further optimizing the technical solution, in step S6, a dynamic proportional conversion model is constructed, and a nonlinear mapping relationship is established by referring to the concentration gradient to achieve rapid quantitative analysis; The dynamic scaling model is as follows:

[0019] in, is the final salidroside content; is the average concentration value of the reference standard solution; is the dynamic gain coefficient; By dividing the content prediction score and Combined with the above, the salidroside content of the actual sample can be quickly calculated.

[0020] To further optimize this technical solution, after detecting the salidroside content in the actual sample, an intelligent review mechanism is constructed: By selecting a very small amount of sample from the same batch of extracts, a rapid colorimetric or electrochemical response test is performed using a preset standard rapid reaction test paper or micro biosensor; Cross-comparison is performed with the content result outputted in step S6; When the deviation between the two exceeds the set threshold, the inspector is prompted to conduct a second test or adjust the parameter weights of the prediction score model.

[0021] In a second aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, the steps of the method for detecting the content of salidroside in Rhodiola rosea extract based on artificial intelligence as described in the first aspect of the present invention are implemented.

[0022] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, the steps of the method for detecting the content of salidroside in a Rhodiola rosea extract based on artificial intelligence as described in the first aspect of the present invention are implemented.

[0023] Compared with the prior art, the present invention provides an artificial intelligence-based method for detecting the content of salidroside in Rhodiola rosea extract, which has the following beneficial effects: This AI-based method for rapid detection of salidroside content in Rhodiola rosea extracts addresses the existing technology's inability to balance detection speed and result reliability by constructing a multi-step, layer-by-layer optimized intelligent detection process. This method utilizes spectral signal cleanup, nonlinear feature extraction, dynamic noise enhancement, AI predictive modeling, and a secondary review mechanism to achieve adaptive analysis of extracts from different sources, significantly improving detection accuracy and stability. It enables rapid and traceable salidroside content detection without the need for large-scale precision instrumentation, effectively supporting on-site batch testing and industrial continuous monitoring needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0025] Figure 1 This is a schematic flow chart of the artificial intelligence-based method for detecting the content of salidroside in Rhodiola rosea extract proposed in the present invention. DETAILED DESCRIPTION

[0026] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0027] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0028] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it constitute a separate or selective embodiment that is mutually exclusive with other embodiments.

[0029] Example 1: Reference Figure 1 , which is the first embodiment of the present invention, provides a method for detecting the content of salidroside in Rhodiola rosea extract based on artificial intelligence, comprising the following steps: S1. In the initial stage of detection, it is necessary to perform spectral scanning on the Rhodiola rosea extract. A multi-band synchronous acquisition device (which can be a spectrometer) is used to perform spectral scanning of the extract sample in different bands within multiple near-infrared wavelength ranges, and the total signal energy and spectral intensity of each band are recorded.

[0030] In existing technologies, the detection of salidroside content primarily relies on traditional methods such as high-performance liquid chromatography (HPLC) and ultraviolet spectrophotometry. While these methods offer high accuracy, they suffer from long detection cycles, strict sample pretreatment requirements, complex operational procedures, high equipment costs, and difficulty adapting to rapid on-site testing. Furthermore, existing spectral-based detection methods often employ fixed detection modes within a single or limited number of bands, failing to fully capture the complex spectral information present in rhodiola rosea extracts. This makes the detection signal susceptible to interference from impurities, reducing the accuracy of the results. Furthermore, the signal energy cannot be uniformly quantified, making subsequent data analysis prone to deviations.

[0031] Since the characteristic absorption intensities of salidroside molecules vary across different wavelengths, the present invention significantly improves the integrity of detection information by simultaneously collecting data from multiple wavelengths. A signal energy calculation model is established to uniformly measure the energy of spectral signals from different wavelengths.

[0032] The signal energy calculation model is as follows:

[0033] in, is the total signal energy.

[0034] is the intensity of the spectrum collected for the i-th band. A high-sensitivity spectral sensor scans and collects the Rhodiola rosea extract across a specific wavelength range, and integrated photometry is used to determine the intensity of each band. Spectral data can be directly recorded using a multi-channel photodiode array or CCD array.

[0035] The i-th band weight is obtained from multiple batches of experiments on salidroside standard samples. It is determined by analyzing the contribution of each band signal to the characteristic absorption peak of salidroside. Weights are typically assigned using artificial intelligence algorithms or expert experience to highlight bands sensitive to target component detection. The value ranges from 0 to 1, and the sum of all band weights is typically normalized to 1 to ensure uniformity in total energy calculations.

[0036] is the total number of bands.

[0037] This model integrates multi-band spectral information in a weighted manner, preventing occasional interference from a single band from affecting the overall signal characteristics. It unifies the dimensionality and importance of signals from different bands, reducing the negative impact of high background noise or low-response bands on the overall detection results. Through this model, the detection system can represent the entire spectral signature in the form of quantified energy, providing standardized input for subsequent steps.

[0038] S2. Based on the obtained total signal energy and the spectral intensity of each band, multi-dimensional interference suppression processing and signal purification are performed to obtain the effective signal intensity after purification.

[0039] Existing technologies for addressing interference in spectral signals during salidroside content detection generally rely on simple filtering or threshold denoising methods. These methods often employ fixed frequency band thresholds, making it difficult to accurately identify and subtract dynamic interference from non-target compounds such as polysaccharides and flavonoids in complex extracts. Traditional methods cannot adjust the intensity of interference peak subtraction based on real-time detection conditions, resulting in a weakened target signal or residual interference when the signal-to-noise ratio is high, ultimately affecting the accuracy and stability of test results.

[0040] Because extracts often contain non-target compounds such as polysaccharides and flavonoids, they can interfere with the spectral signal of salidroside. The present invention introduces a dynamic characteristic peak comparison mechanism. Unlike traditional fixed-threshold filtering, this mechanism compares detected spectral peaks one by one against a database of salidroside characteristic peaks, automatically identifying and distinguishing interfering peaks from target peaks. This step simultaneously addresses two main types of interference: chemical matrix interference from the sample itself (such as characteristic peak interference caused by non-target compounds such as polysaccharides and flavonoids), and physical noise generated by the external environment or equipment (such as ambient light, scattered light, and detector dark noise). This not only removes spectral overlap interference caused by sample components but also suppresses external physical noise, ensuring the comprehensiveness of the purified signal. By setting suppression coefficients for different interfering peaks, the model precisely controls the proportion of total signal energy subtracted by each type of interference, avoiding the accidental deletion of valid signals. This not only achieves higher target component signal purity in complex extracts, but also automatically adapts to the interference characteristics of samples from different batches and sources, making the detection process more intelligent and adaptive.

[0041] The calculation method of the effective signal intensity after purification is as follows:

[0042] in, is the effective signal intensity after purification.

[0043] is the intensity of the jth interference peak detected. The noise signal intensity value measured by the spectrometer in the region without characteristic absorption peaks or in the reference blank sample is used to characterize the signals of non-target components such as ambient light and scattered light.

[0044] is the suppression coefficient for the jth interfering peak. The average influence coefficient for each type of noise is calculated by analyzing the background signal of multiple batches of Rhodiola rosea extract samples, using blank solvent measurement results or a noise database fit. The value typically ranges from 0 to 1, reflecting the ratio of noise intensity to the original signal.

[0045] is the total number of interference peaks.

[0046] The model achieves pure salidroside spectral feature extraction by calculating the weighted energy of the interfering components and deducting them from the total signal.

[0047] During actual use, the system will automatically load the interference peak database, compare the spectra in real time, calculate the energy of each interference peak, and dynamically adjust the deduction ratio to achieve accurate removal of noise and non-target component signals, significantly improving the accuracy and reliability of subsequent salidroside content detection.

[0048] In this step, the process of identifying and distinguishing interference peaks from target peaks is as follows: first, the multi-band spectral data is compared one by one with the salidroside characteristic peak database to lock the characteristic peak position and relative intensity range that should appear in theory; second, if the actual peak detected deviates significantly from the characteristic peak database record in position, shape or intensity ratio, it is determined to be an interference peak; third, the interference intensity of the signal determined to be an interference peak is calculated. , and according to the inhibition coefficient obtained by fitting Perform weighted subtraction on it; finally, retain the target peak signal determined by matching to form the purified effective signal .

[0049] S3. After signal purification is completed, in order to reduce redundant information and highlight the characteristics of salidroside, the purified effective signal intensity is used as input, and the salidroside characteristic intensity factor is obtained through spectral feature compression calculation.

[0050] In the prior art, when processing spectral signals, linear regression or simple peak area integration methods are usually used to estimate the characteristic intensity of the target component. However, the multi-band spectral response in Rhodiola rosea extracts often exhibits nonlinear characteristics. Especially in the high concentration range, the signal is prone to saturation or nonlinear amplification, resulting in the linear model being unable to accurately reflect the true content. In addition, the existing technology for processing multi-band information often lacks effective compression strategies, and is prone to problems such as excessively high input data dimensions, noise accumulation, and low computational efficiency, which limits the application effect of rapid detection.

[0051] In the present invention, a characteristic intensity factor calculation model is constructed, and by considering the nonlinear differences in the responses of different bands to salidroside, a nonlinear compression function is used for fusion to reduce the saturation distortion effect of high-concentration signals.

[0052] The characteristic intensity factor calculation model is as follows:

[0053] in, is the characteristic intensity factor of salidroside.

[0054] The characteristic intensity of the reference standard salidroside sample in band i. The spectral response signal values ​​under each calibration condition, recorded in the standard sample database, are obtained through historical experimental data or online real-time reference testing. In this step, the test object is a high-purity, known-concentration salidroside standard solution. This represents the spectral intensity of the standard sample in the same band, which is the characteristic absorption or emission position of salidroside, hence the name "characteristic intensity." To facilitate the distinction between "unknown test data" and "standard reference data," the terms "spectral intensity" and "characteristic intensity" are used in the nomenclature.

[0055] is the correction factor. By performing multiple spectral measurements on a standard salidroside solution, the influence of the spectral response on concentration prediction under each reference condition (such as different temperature, pH, or solvent conditions) is calculated. The AI ​​algorithm quantifies this influence. The value typically ranges from 0.1 to 5, with larger values ​​indicating a greater influence of the reference condition on the final content prediction.

[0056] The model output This factor characterizes the relative proximity of the purified signal to the standard signature, directly reflecting the potential level of salidroside. A logarithmic function reduces errors caused by over-amplification of high-concentration sample signals. The system uses this factor as an input parameter for the next AI prediction model, significantly reducing redundant data dimensions and improving computational efficiency while ensuring stable and reliable salidroside content detection results even in complex sample backgrounds.

[0057] S4. Based on the characteristic intensity factor of salidroside, the characteristic signal-to-noise ratio enhancement and purity index of salidroside are calculated.

[0058] Most existing methods for analyzing spectral signals from complex plant extracts rely solely on fixed filtering parameters or simple noise subtraction techniques to improve the signal-to-noise ratio. These methods often ignore variations in noise distribution caused by differences in the sample matrix itself, making it impossible to dynamically adjust the noise suppression process. When the target of detection is Rhodiola rosea extract, non-target components (such as polysaccharides, amino acids, and pigments) introduce random noise of varying frequencies. Traditional processing methods are unable to effectively distinguish and dynamically adapt to these residual interfering signals, which can easily lead to insufficient purity of the target salidroside signal, thereby affecting the accuracy of subsequent concentration predictions.

[0059] Step S3 utilizes some "noise patterns" in the original signal as auxiliary features for fitting, while step S4, after the mapping is complete, further removes random noise that hinders result presentation and subsequent storage, thereby striking a balance between preserving useful feature information and improving signal purity. This sequence is not simply noise reduction, but a strategy of feature extraction followed by signal purification, ensuring that the final output is both accurate and easy to analyze and apply.

[0060] When performing salidroside characteristic signal-to-noise ratio enhancement, a signal-to-noise ratio enhancement model is constructed to dynamically attenuate the residual low-frequency noise in the purified signal; The signal-to-noise ratio enhancement model is as follows:

[0061] in, To enhance signal strength.

[0062] The residual noise energy is measured by statistically analyzing the residual background signal intensity in the non-characteristic peak region during spectral analysis or by using noise detection technology to measure the residual noise energy in real time.

[0063] is the noise attenuation coefficient. By comparing the impact of residual noise on prediction results through multiple batches of experiments, a fitting method or an AI adaptive algorithm is used to determine the optimal attenuation coefficient. The value range is typically between 0.01 and 5, with larger values ​​indicating greater noise suppression of the signal.

[0064] After the signal-to-noise ratio enhancement model outputs the enhanced signal strength and residual noise energy, the purity index is calculated:

[0065] in, This is a purity index used to measure the purity of the salidroside signal. Its value ranges from 0 to 1, with a value close to 1 indicating a clean spectral signal with minimal noise interference.

[0066] This step can adaptively adjust the signal enhancement strength according to the actual noise level of each detection, so that the signal purification results are more consistent with the actual sample characteristics, thereby improving the distinction between the salidroside signal and background noise, and significantly improving the input data quality of downstream artificial intelligence content prediction.

[0067] During actual use, the detection system automatically adjusts according to the real-time noise level The enhancement effect is more pronounced in samples with higher noise levels, while over-amplification errors are avoided in samples with lower noise levels, resulting in more accurate signal processing. This method overcomes the problem of insufficient noise suppression in complex extracts by traditional technologies, providing a high-purity characteristic input signal for subsequent AI-based content prediction.

[0068] S5. Using the purity index obtained through calculation, the salidroside content is predicted through a prediction score model and a content prediction score is output.

[0069] Salidroside content prediction often relies on traditional modeling approaches based on linear fitting or multiple regression, estimating content by simply comparing the difference between the input signal purity parameter and the standard reference concentration curve. In practical applications, such methods are susceptible to factors such as sample batch differences, noise fluctuations, and spectral nonlinearity. This results in poor model generalization, the need for frequent parameter recalibration for different extract samples, low detection efficiency, and difficulty ensuring the accuracy and stability of prediction scores, making them unable to meet the needs of rapid testing and industrial batch testing.

[0070] In the present invention, the prediction score model constructed is as follows:

[0071] in, The score for content prediction. A higher score indicates that the test result is closer to the standard reference conditions and the prediction accuracy is higher.

[0072] This is the score amplification factor. This factor is obtained through training with historical detection data and represents the weight of signal purity's contribution to the prediction result. The algorithm automatically determines the optimal value based on fitting. The value range is typically between 0.5 and 5, with larger values ​​indicating a greater influence of signal purity on the prediction score.

[0073] This is the sensitivity correction factor. Derived through a machine learning optimization process, it adjusts the sensitivity of the predicted score to deviations from the standard purity signal, ensuring that the results are more consistent with the standard reference conditions. It is generally between 0.1 and 10, with larger values ​​penalizing deviations more severely.

[0074] The theoretical purity value of a standard salidroside sample. This reference purity value, obtained through multiple experiments on standard salidroside solutions, serves as a benchmark input parameter for the prediction score model. This value generally ranges from 0.8 to 1, based on standard laboratory conditions.

[0075] Continuously optimize parameters through artificial intelligence training and , so that it can accurately output the salidroside content prediction score for different production batches of Rhodiola rosea extracts .

[0076] This step not only utilizes spectral purity metrics but also dynamically corrects the predicted value based on deviations from the theoretical purity value of standard salidroside, enabling rapid adaptation to extracts from different sources and under different process conditions. Compared to traditional fixed-parameter linear models, this method possesses self-learning and adaptive capabilities, optimizing parameter settings through artificial intelligence algorithms. This significantly improves the accuracy and robustness of content predictions, reduces manual parameter adjustments, and enhances the overall level of automated testing.

[0077] S6. Map the content prediction score to the final salidroside content and quickly output the test results.

[0078] In existing technologies, quantitative analysis of salidroside content often relies on laboratory testing methods such as high-performance liquid chromatography (HPLC). Although these methods offer high accuracy, they are time-consuming and require complex pre-treatment steps and expensive instrumentation. Even with the limited number of rapid detection methods available, these methods often directly convert spectral signal intensity or predicted scores into content using a fixed proportionality factor, lacking the ability to dynamically correct for differences between sample batches and fluctuations in the extraction process. Consequently, for extract samples from different sources and with different concentration gradients, the quantitative results exhibit large errors, and the reliability and universality of rapid detection are insufficient.

[0079] The present invention constructs a dynamic proportional conversion model, establishes a nonlinear mapping relationship by referring to the concentration gradient, and realizes rapid quantitative analysis.

[0080] The dynamic scaling model is as follows:

[0081] in, is the final salidroside content.

[0082] The average concentration of the reference standard solution is set by the known concentration of the standard salidroside solution, usually measured by a high-purity standard solution prepared in the laboratory and stored in the database.

[0083] The dynamic gain coefficient is obtained by fitting a large number of sample test data with the actual measured concentration values. It is used to reflect the sensitivity of the predicted score to concentration correction. It is usually between 0.01 and 1. The larger the value, the more significant the impact of the predicted score on the final concentration result.

[0084] By dividing the content prediction score and Combined with the above, the salidroside content of the actual sample can be quickly calculated.

[0085] Compared with traditional methods, this step can adapt to various uncertain factors such as fluctuations in raw material concentration and differences in extraction processes, achieving more flexible and adaptive quantitative calculations, and greatly improving the accuracy and versatility of rapid detection methods in different production environments.

[0086] In this example, although the rapid detection results of salidroside content are given, due to differences in raw material origin, extraction solvent ratio, temperature fluctuation, etc. between different extract batches, there may be slight deviations in the detection results caused by matrix effects.

[0087] Therefore, after detecting the salidroside content in the actual sample, this method constructs an intelligent review mechanism: By selecting a very small amount of sample from the same batch of extracts, a quick colorimetric or electrochemical response test is performed using a preset standard rapid reaction test paper or micro-biosensor, and the verification signal is used as a "soft calibration" reference value; Cross-comparison is performed between the reference value and the content result outputted in step S6; When the deviation between the two exceeds the set threshold, the tester is prompted to conduct a second test or adjust the parameter weights of the prediction score model to adapt to the new matrix characteristics.

[0088] Compared with existing technologies, this step significantly improves the stability and reliability of the rapid detection method. At the same time, it does not require the introduction of large instruments or the extension of detection time. It can be achieved by relying solely on low-cost, rapid biochemical sensors to assist in judgment, significantly improving the robustness of the detection system. It is particularly suitable for continuous detection on production lines and instant verification of large batches of samples, ensuring that the results are more reliable and traceable.

[0089] Example 2: This embodiment also provides a computer device, which is suitable for the method for detecting the content of salidroside in Rhodiola rosea extract based on artificial intelligence, and includes a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the method for detecting the content of salidroside in Rhodiola rosea extract based on artificial intelligence proposed in the above embodiment.

[0090] This embodiment also provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for detecting the content of salidroside in a Rhodiola rosea extract based on artificial intelligence as proposed in the above embodiment.

[0091] The computer device may be a terminal, comprising a processor, memory, a communication interface, a display, and an input device connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system and computer programs. The internal memory provides an environment for the operating system and computer programs stored in the non-volatile storage media. The communication interface of the computer device is used to communicate with external terminals via wired or wireless communication. Wireless communication may be achieved via Wi-Fi, a carrier network, NFC (near-field communication), or other technologies. The display of the computer device may be a liquid crystal display or an electronic ink display. The input device may be a touchscreen overlay on the display, buttons, a trackball, or a touchpad on the computer device housing, or an external keyboard, touchpad, or mouse.

[0092] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the method of the present invention. The aforementioned storage medium includes various media that can store program code, such as USB flash drives, mobile hard drives, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0093] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0094] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting, or processing it in another suitable manner as necessary, and then storing it in a computer memory.

[0095] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having logic gate circuits for implementing logic functions on data signals, an application-specific integrated circuit having suitable combinational logic gate circuits, a programmable gate array (PGA), a field-programmable gate array (FPGA), etc.

[0096] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for detecting the content of salidroside in Rhodiola rosea extract based on artificial intelligence, characterized in that: The following steps are involved: S1. Use a multi-band synchronous acquisition device to perform spectral scanning of the extract sample in different bands within multiple near-infrared wavelength ranges, and record the total signal energy and spectral intensity of each band; S2. Based on the obtained total signal energy and the spectral intensity of each band, multi-dimensional interference suppression processing and signal purification are performed to obtain the effective signal intensity after purification; S3, using the purified effective signal intensity as input, and calculating the salidroside characteristic intensity factor through spectral feature compression; S4. Based on the characteristic intensity factor of salidroside, the characteristic signal-to-noise ratio enhancement and purity index of salidroside are calculated; S5. Using the calculated purity index, the prediction score model is used to predict the salidroside content and output a content prediction score; S6. Map the content prediction score to the final salidroside content and quickly output the test results.

2. The method for detecting the content of salidroside in Rhodiola rosea extract based on artificial intelligence according to claim 1, wherein: In step S1, a signal energy calculation model is established to uniformly measure the energy of spectral signals of different bands; The signal energy calculation model is as follows: ； in, is the total signal energy; is the collected spectrum intensity of the i-th band; is the weight of the i-th band; is the total number of bands; The model integrates multi-band spectral information in a weighted manner to avoid occasional interference from a single band affecting the overall signal characteristics.

3. The method for detecting the content of salidroside in Rhodiola rosea extract based on artificial intelligence according to claim 1, wherein: In step S2, a dynamic characteristic peak comparison mechanism is introduced to determine non-salidroside characteristic peaks by comparing with the reference characteristic peak database in real time and performing signal subtraction to purify the effective signal intensity.

4. The method for detecting the content of salidroside in Rhodiola rosea extract based on artificial intelligence according to claim 3, wherein: The calculation method of the effective signal intensity after purification is as follows: ； in, is the effective signal intensity after purification; is the intensity of the jth interference peak detected; is the suppression coefficient of the j-th interference peak; is the total number of interference peaks; The model achieves pure salidroside spectral feature extraction by calculating the weighted energy of the interfering components and deducting them from the total signal.

5. The method for detecting the content of salidroside in Rhodiola rosea extract based on artificial intelligence according to claim 1, wherein: In step S3, a characteristic intensity factor calculation model is constructed, and nonlinear compression function is used to perform fusion by considering the nonlinear differences in the responses of different bands to salidroside; The characteristic intensity factor calculation model is as follows: ； in, is the characteristic intensity factor of salidroside; is the characteristic intensity of the reference standard salidroside sample in band i; is the correction factor; The model output It is used to characterize the relative proximity between the purification signal and the standard feature, which directly reflects the potential content level of salidroside.

6. The method for detecting the content of salidroside in Rhodiola rosea extract based on artificial intelligence according to claim 1, wherein: In step S4, when performing salidroside characteristic signal-to-noise ratio enhancement, a signal-to-noise ratio enhancement model is constructed to dynamically attenuate the residual low-frequency noise in the purified signal; The signal-to-noise ratio enhancement model is as follows: ； in, To enhance signal strength; is the residual noise energy; is the noise attenuation coefficient.

7. The method for detecting the content of salidroside in Rhodiola rosea extract based on artificial intelligence according to claim 6, characterized in that: After the signal-to-noise ratio enhancement model outputs the enhanced signal strength and residual noise energy, the purity index is calculated: ； in, It is a purity index used to measure the purity of salidroside signal.

8. The method for detecting the content of salidroside in Rhodiola rosea extract based on artificial intelligence according to claim 1, wherein: In step S5, the prediction score model is as follows: ； in, Score for content prediction; is the score magnification factor; is the sensitivity correction factor; is the theoretical purity value of the standard salidroside sample; Continuously optimize parameters through artificial intelligence training and , so that it can accurately output the salidroside content prediction score for different production batches of Rhodiola rosea extracts .

9. The method for detecting the content of salidroside in Rhodiola rosea extract based on artificial intelligence according to claim 1, wherein: In step S6, a dynamic proportional conversion model is constructed, and a nonlinear mapping relationship is established by referring to the concentration gradient to achieve rapid quantitative analysis; The dynamic scaling model is as follows: ； in, is the final salidroside content; is the average concentration value of the reference standard solution; is the dynamic gain coefficient; By dividing the content prediction score and Combined with the above, the salidroside content of the actual sample can be quickly calculated.

10. The method for detecting the content of salidroside in Rhodiola rosea extract based on artificial intelligence according to claim 9, characterized in that: After detecting the salidroside content in the actual sample, this method constructs an intelligent review mechanism: By selecting a very small amount of sample from the same batch of extracts, a rapid colorimetric or electrochemical response test is performed using a preset standard rapid reaction test paper or micro biosensor; Cross-comparison is performed with the content result outputted in step S6; When the deviation between the two exceeds the set threshold, the inspector is prompted to conduct a second test or adjust the parameter weights of the prediction score model.

Images