Preparation method and application of green and environment-friendly olfactory color-sensitive sensor

Abstract

This invention relates to the fields of nanomaterial synthesis technology, olfactory color-sensitive sensing and detection, and non-destructive testing of food, specifically to a method for preparing and applying a green and environmentally friendly olfactory color-sensitive sensor. First, CL-CD-MOF nanomaterials are obtained, and then CL-CD-MOF@color-sensitive dyes are prepared. Next, a CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor is prepared. Finally, a food quality evaluation model is constructed based on the sensor to achieve rapid evaluation of food quality. This invention fully utilizes the advantages of CL-CD-MOF nanomaterials, such as high biocompatibility, large surface area, and high porosity, overcoming the technical challenge of controllable preparation of nanoscale materials. The prepared sensor can efficiently capture volatile components, showing a significant improvement in sensitivity compared to ordinary olfactory color-sensitive sensors. Furthermore, the materials used are safe and environmentally friendly, offering broader and safer application prospects in the field of food testing.

Description

Technical Field

[0001] This invention relates to the fields of nanomaterial synthesis technology, olfactory color-sensitive sensing and detection, and non-destructive testing technology for food, specifically to a method for preparing and applying a green and environmentally friendly olfactory color-sensitive sensor. Background Technology

[0002] Food quality is a core attribute determining a product's acceptability and market value, and its evaluation involves multiple indicators such as appearance, texture, aroma, taste, and nutritional components. Among these, flavor quality, as a key characteristic of food quality, directly impacts consumer decisions and supply chain efficiency. Current flavor evaluation primarily relies on two methods: sensory evaluation and instrumental analysis. While subjective sensory evaluation methods can directly reflect human perception, they are easily influenced by the evaluator's physiological state, experience differences, and environmental factors, exhibiting inherent defects such as strong subjectivity and poor reproducibility. Instrumental analysis (such as gas chromatography-mass spectrometry) can objectively detect volatile flavor substances, but requires complex pretreatment, specialized operation, and expensive equipment, resulting in limitations such as long processing times, cumbersome procedures, and difficulty in rapid on-site response. Faced with the food industry's urgent need for efficient, standardized, and real-time quality control, existing methods struggle to balance speed, objectivity, and practicality. Therefore, developing a novel sensing technology capable of capturing real-time changes in volatile components of food, converting odor signals into intuitive and readable responses, and suitable for rapid on-site detection has become a key direction for overcoming industry bottlenecks.

[0003] Color-sensitive sensing technology, characterized by its fast response, simple operation, and portability, has become a promising technology. Odor-based color-sensitive sensing technology is widely used in plant disease detection, food safety factor detection, food flavor analysis, and human lung cancer monitoring, among other applications. However, olfactory color-sensitive sensing technology suffers from low detection sensitivity, limiting its application in food testing. Metal-organic frameworks (MOFs), novel nanomaterials self-assembled from metal ions and organic ligands, have been widely used in the fabrication of gas sensors. γ-cyclodextrin (γ-CD) and potassium ions (K+)... + This invention utilizes a self-assembly process to form a green and usable cyclodextrin metal-organic framework (CD-MOF) nanoporous material. This material can effectively enrich volatile gas components, improving the sensitivity of gas sensors, while also exhibiting good biocompatibility, expanding its applications in food. Therefore, this invention improves upon existing CD-MOF synthesis processes, successfully preparing nanoscale CD-MOFs and applying them to color-sensitive sensing technology, significantly enhancing detection sensitivity. Furthermore, by incorporating a deep learning model, it provides a precise and efficient solution for rapid on-site detection of food flavor and quality. Summary of the Invention

[0004] This invention controllably synthesizes nanoscale cross-linked cyclodextrin metal-organic frameworks (CL-CD-MOF), provides a method for preparing olfactory nanocolor-sensitive sensors based on CL-CD-MOF@color-sensitive dyes, and proposes a strategy combining a bidirectional neural network model (Bi-LSTM) for food quality evaluation to solve the problems of low detection sensitivity, complex detection process, and low detection accuracy in existing technologies.

[0005] To achieve the above objectives, the present invention provides the following solution: Firstly, a method for preparing a green and environmentally friendly olfactory color-sensitive sensor (CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor) is provided, which specifically includes the following steps: Step 1, Preparation of CD-MOF and CL-CD-MOF nanomaterials: (1) Weigh potassium hydroxide (KOH) and cyclodextrin (CD) and dissolve them in deionized water. After ultrasonic treatment, KOH-CD mixed solution is obtained. Weigh PEG-4000 and dissolve it in methanol to obtain PEG-4000 methanol solution. Then mix the PEG-4000 methanol solution with the KOH-CD mixed solution, collect the precipitate by centrifugation after water bath incubation, wash the precipitate with methanol and centrifuge, and finally dry the washed precipitate to constant weight to obtain CD-MOF nanomaterial. The energy input provided by ultrasonic treatment enables the system to quickly overcome the nucleation energy barrier and accelerate the nucleation rate. The PEG-4000 methanol solution acts as a reaction control agent to inhibit subsequent crystal growth, thereby obtaining small-sized CD-MOF. (2) Weigh the CD-MOF nanomaterials and diphenyl carbonate from step (1) and dissolve them in N,N-dimethylformamide solution. Add triethylamine as a catalyst and carry out cross-linking reaction under stirring conditions. After the reaction, collect the precipitate and wash it with ethanol. After washing, centrifuge and dry it to obtain the dried product, which is CL-CD-MOF. Preferably, in step (1), the ratio of potassium hydroxide, cyclodextrin, and deionized water is 112 mg: 324 mg: 8 mL; the ultrasonic power of the ultrasonic treatment is 300 W, and the ultrasonic duration is 10-50 min; the ratio of PEG-4000 to methanol is 50 mg-250 mg: 8 mL; the volume ratio of PEG-4000 methanol solution to KOH-CD mixed solution is 1:1; the water bath temperature is 30-50 ℃, and the holding time is 1 h; the number of times methanol is used for washing is 3-5 times.

[0006] Preferably, in step (2) of step one, the ratio of CD-MOF nanomaterial, diphenyl carbonate, N,N-dimethylformamide and triethylamine is 0.5 g: 0.49 g: 6.5 mL: 0.2-0.4 mL; the temperature of the crosslinking reaction is 60-80℃, the time is 12-24 h, and the number of times ethanol is used for washing is 3-5 times.

[0007] Step 2, Preparation of CL-CD-MOF@color-sensitive dyes: The color-sensitive dye and the CL-CD-MOF obtained in step one were added to methanol, stirred at room temperature, and the precipitate was collected by centrifugation. The precipitate was washed with methanol, centrifuged and dried to obtain CL-CD-MOF@color-sensitive dye powder. Preferably, in step two, the ratio of the color-sensitive dye, CL-CD-MOF, and methanol is 20 mg: 80 mg: 10 mL, and the stirring time is 24 h; the washing with methanol is performed 2-3 times; the color-sensitive dye includes one or more of 8-(4-nitrophenyl)-4,4-difluoro-2,6-dibromoborondipyrrolemethane, 5,10,15,20-tetraphenyl-21H,23H-porphyrin manganese(II), bromothymol blue, bromocresol green, methyl red, bromophenol blue, cresol red, or crystal violet.

[0008] Step 3, Fabrication of CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor: S1. First, prepare a series of standard test solutions of color-sensitive dyes at various concentrations; then, use a UV-Vis spectrophotometer to perform spectral scanning on the standard color-sensitive dye solutions of each concentration and record the absorbance values ​​at their characteristic absorption peaks; according to Lambert-Beer's law, i.e. A = εbc , in A Absorbance ε The molar absorptivity is 1. b For optical path, c For concentration, under fixed wavelength and cuvette optical path length conditions, absorbance shows a linear relationship with concentration; ultimately, for concentration... c and absorbance A Perform linear fitting to establish a standard curve; S2. Dissolve the CL-CD-MOF@color-sensitive dye powder obtained in step two in a solvent to obtain a mixed solution; the solvent includes N,N-dimethylformamide and / or dichloromethane and / or anhydrous ethanol; use a UV-Vis spectrophotometer to perform a spectral scan on the mixed solution and measure its absorbance A0 at a characteristic wavelength; substitute the absorbance value A0 into the standard curve established in step S1 to obtain the concentration of color-sensitive dye in the solution, and then calculate the loading of color-sensitive dye in CL-CD-MOF@color-sensitive dye; S3. Based on the loading of color-sensitive dye in CL-CD-MOF@ color-sensitive dye calculated in S2, weigh out the corresponding CL-CD-MOF@ color-sensitive dye powder and dissolve it in a solvent, wherein the solvent includes N,N-dimethylformamide and / or dichloromethane and / or anhydrous ethanol; obtain a shaken solution; then add PEG-300 to the shaken solution to obtain a CL-CD-MOF@ color-sensitive dye solution; fix the CL-CD-MOF@ color-sensitive dye solution on a substrate to prepare a CL-CD-MOF@ color-sensitive dye olfactory color-sensitive sensor; and store the prepared olfactory nano-color-sensitive sensor in a sealed bag for later use. Preferably, the concentration range of the color-sensitive dye detection solution in step S1 of step three is 1×10⁻⁶. -4 ~1×10 -1 mg / mL; X types of color-sensitive dye solutions are prepared, where X is a positive integer, and the X types of color-sensitive solutions include any one or more of solution A, solution B, or solution C; wherein solution A is an N,N-dimethylformamide solution of fluoroborophyrrole, solution B is a dichloromethane solution of metalloporphyrin, and solution C is an ethanol solution of pH indicator; wherein the ratio of fluoroborophyrrole to N,N-dimethylformamide in solution A is 1 mg:1 mL, the ratio of metalloporphyrin to dichloromethane in solution B is 1 mg:1 mL, and the ratio of pH indicator to ethanol in solution C is 1 mg:1 mL; The fluoroboropyrrole comprises 8-(4-nitrophenyl)-4,4-difluoro-2,6-dibromoborondipyrrolemethane; the metalloporphyrin comprises 5,10,15,20-tetraphenyl-21H,23H-porphyrin manganese(II); the pH indicator comprises bromothymol blue, bromocresol green, methyl red, bromophenol blue, cresol red, and crystal violet. Preferably, in step S2 of step three, the loading of the color-sensitive dye in the CL-CD-MOF@color-sensitive dye is calculated as follows: the mass of the CL-CD-MOF@color-sensitive dye is M0 (unit: mg), the volume of the solvent is V (unit: mL), and the concentration of the CL-CD-MOF@color-sensitive dye solution is C. 复 (Unit: mg / mL); the solvent includes N, N - Dimethylformamide and / or dichloromethane and / or anhydrous ethanol, C 复 = M0 / V; the concentration-absorbance linear relationship formula is... A = k C + b, where k is the slope, representing the change in absorbance caused by a unit change in concentration, and is a real number; b is the intercept, representing the background absorbance value when the concentration is 0, and is a real number; C For concentration, AThe absorbance is A; the absorbance value of the CL-CD-MOF@color-sensitive dye solution obtained using a UV-Vis spectrophotometer is A. 测 Substitute A = k C In the + b relationship, the concentration of the color-sensitive dye in the CL-CD-MOF@color-sensitive dye is calculated to be C. 测 mg / mL, C 测 = (A) 测 -b) / k, further yielding that each 1 mg of CL-CD-MOF@color-sensitive dye contains L mg of color-sensitive dye, where L = C 测 / C 复 = (A 测 -b)×V / (k×M0; Preferably, in step S3, the ratio of CL-CD-MOF@color-sensitive dye powder to solvent is 1 / L mg: 1 mL; the shaking treatment time is 15-30 min; the volume ratio of the shaken solution to the PEG-300 solution is 1 mL: 0.2 mL; the amount of CL-CD-MOF@color-sensitive dye solution fixed on the substrate is 1.5-2 μL; the substrate includes cellulose filter paper, silica gel plate, PVDF membrane, and mixed cellulose ester.

[0009] The application of the CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor prepared according to this invention to food quality evaluation follows these steps: (1) Construction of food quality evaluation model: S1. Sample and Grade Classification: Samples with different processing levels are selected. Different processing levels correspond to different volatile odor substances, which will cause the CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor to produce different color changes. The samples include agricultural raw materials, deep-processed agricultural and tea products, aquatic products, etc. The different quality grades include the degree of processing, freshness grade, or different sensory quality grades of the samples. S2. Feature extraction and model training First, an image of the CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor before reaction is acquired using a camera. Then, the sample is placed in a reaction container, and the CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor and the sample are stored in the same sealed reaction container, allowing the volatile odor substances of the color-sensitive sensor and the sample to react fully for a period of time. After the reaction, an image of the olfactory nano-color-sensitive sensor after the reaction is acquired using a camera, and the obtained image is saved to a computer. The computer is used to locate the position of the color-sensitive units in the images before and after the reaction, extract their color features, and calculate the difference in the mean grayscale values ​​of each color-sensitive unit before and after the reaction. The difference in the mean grayscale values ​​is the characteristic variable of the color-sensitive unit. The feature variables of each color-sensitive unit in all samples are concatenated into a feature matrix. This matrix is ​​used as input, and the corresponding sample quality level is used as output. The model is trained using a bidirectional long short-term memory recurrent neural network (Bi-LSTM) to obtain the final food quality evaluation model.

[0010] (2) Rapid evaluation of food quality: The sample to be tested is reacted according to the method described in (1) to obtain the characteristic variables; the characteristic variables of the sample to be tested are substituted into the Bi-LSTM model constructed in (1) to obtain the quality grade information of the sample to be tested and realize the rapid evaluation of food quality.

[0011] Preferably, in step (1) S2, when the sample is a solid sample, the amount used is 0.5-10 g, and when the sample is a liquid sample, the amount used is 5-10 mL; the reaction time is 10-30 min; the CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor is fixed at the top of the reaction container; Preferably, in step (1) S2, the feature variable extraction is performed according to the following steps: using a computer to locate the coordinate region of each color-sensitive unit on the sensor; extracting the color-sensitive sensor images before and after the reaction, and decomposing them into grayscale images of the R, G, and B channels; simultaneously calculating the values ​​of each color-sensitive unit in hue (H), saturation (S), brightness (V), lightness (L), red-green component (a), and yellow-blue component (b); subtracting the channel values ​​after the reaction from those before the reaction to obtain ΔR, ΔG, ΔB, ΔH, ΔS, ΔV, ΔL, Δa, and Δb in sequence; and according to ED = (ΔR... 2 +ΔG 2 +ΔB 2 ) 1 / 2 Further calculate the Euclidean distance; △R, △G, △B, △H, △S, △V, △L, △a, △b and ED are the characteristic variables of a color-sensitive unit. After the sensor array composed of M color-sensitive units is acquired, a total of N-dimensional characteristic variables are derived, where N=10M.

[0012] Preferably, in step (1) S2, the size of the sample set used to train the food quality assessment model is set to E, which spans e discrete levels, and each level is equipped with f parallel samples, i.e., E=e×f; where e is a positive integer not less than 2, and f and E are both positive integers.

[0013] The preferred method for constructing a Bi-LSTM food quality evaluation model is as follows: Let the feature matrix be S∈R E×N , where E is the number of samples and N is the total number of characteristic variables; The feature matrix S, as the input sequence, is first fed into a bidirectional long short-term memory network for bidirectional temporal modeling, capturing the forward and backward sequence dependencies respectively, and outputting a fused hidden state matrix H∈R. E×D D is the dimension of the hidden state space; subsequently, the fused hidden state matrix H undergoes dimensional transformation and feature recombination through a linear projection layer to obtain the reconstructed output matrix Z: Z = ReLU(W proj ×H + b proj ) Where ReLU is the rectified linear activation function, W proj Let b be the weight parameter matrix of the linear projection layer. proj The bias vector is used; a Dropout random deactivation layer is cascaded after the linear projection layer, randomly zeroing the activation values ​​of some neurons with a probability ratio of 0.15-0.25, thereby enhancing the model's generalization ability and suppressing overfitting; then the output of the Dropout layer is fed into a Softmax normalized exponential layer, outputting the probability matrix P∈R for each category. E×t t represents the number of food quality grade categories; During model training, class cross-entropy is used as the loss metric, and the Adam optimizer is used for end-to-end parameter updates. The parameters to be optimized include: the weight matrices and bias vectors of each gate in the Bi-LSTM, and the W of the linear projection layer. proj With b proj And the classification weights of the Softmax layer; the true quality level label matrix T∈R encoded in one-hot encoding. E×t As a supervisory benchmark, the cross-entropy deviation between the predicted outputs Z and T is minimized, thereby constructing the Bi-LSTM model for food quality assessment. Preferably, the rapid evaluation steps for food quality in step (2) are as follows: Based on the signal acquisition and feature extraction process described in step (1), extract the total number of N-dimensional feature variables from the F samples to be tested, and construct the feature matrix to be evaluated R∈R F ×N The Bi-LSTM model for food quality assessment from step (1) is invoked, with the feature matrix R used as the input for forward inference. The model outputs a probability prediction matrix Q∈R.F×t The probability vector in each row of matrix Q is processed by argmax, and the food quality grade of each sample is determined based on the maximum probability index, thus completing a rapid and automated evaluation of food quality.

[0014] The present invention discloses the following technical effects: (1) The method described in this invention utilizes the nanoporous structure of CL-CD-MOF nanomaterials as a carrier for color-sensitive dyes to prepare a method for preparing a CL-CD-MOF@color-sensitive dye olfactory nanocolor-sensitive sensor, and discloses its rapid evaluation method for food quality. The main focus is on breaking through the small size and good biocompatibility of the CL-CD-MOF@color-sensitive dye olfactory nanocolor-sensitive sensor, and realizing rapid evaluation of food quality by acquiring information on volatile components of food.

[0015] (2) The method described in this invention combines hydrothermal synthesis and ultrasonic-assisted synthesis. The energy input provided by ultrasonic treatment enables the system to quickly overcome the nucleation energy barrier and accelerate the nucleation rate. The methanol solution of PEG-4000 acts as a reaction control agent to inhibit subsequent crystal growth. By optimizing the ultrasonic time, PEG addition amount and heating temperature in the CD-MOF synthesis method, nanoscale CD-MOF materials are prepared, and high water stability CL-CD-MOF nanomaterials are prepared by crosslinking diphenyl carbonate, effectively controlling the size of the synthesized CL-CD-MOF nanomaterials.

[0016] (3) The raw material selected in this invention is low-toxicity K + It is formed by self-assembly of CD-MOF with γ-cyclodextrin through coordination bonds. The raw materials can be obtained from food ingredients. The synthesized CD-MOF has the characteristics of low toxicity and high biocompatibility.

[0017] (4) Compared with traditional color-sensitive sensors based on color-sensitive dyes, the present invention prepares a CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor by coupling color-sensitive dyes (fluoroboron pyrrole, metalloporphyrin and pH indicator) with CL-CD-MOF. This sensor can effectively enrich volatile organic compounds and improve the detection sensitivity of the sensor.

[0018] (5) In order to highlight the beneficial effects of the olfactory nano-color-sensitive sensor, the present invention proposes a method for calculating the effective loading of color-sensitive dye in CL-CD-MOF@color-sensitive dye. By controlling a certain concentration of color-sensitive dye, the comparability of the olfactory nano-color-sensitive sensor with traditional color-sensitive sensors is enhanced.

[0019] (6) The bidirectional long short-term memory recurrent neural network (Bi-LSTM) model for rapid evaluation of food quality established in this invention is a deep learning model. The model is configured with ReLU activation layer and Dropout layer to effectively prevent overfitting. When the training sample batches are sufficiently extensive and the sample size is sufficient, the evaluation model has good universality and can improve the efficiency of quality evaluation. It is of great significance for the evaluation of food quality during the processing. Attached Figure Description

[0020] Figure 1 This is a particle size distribution diagram of the CD-MOF nanomaterial in Example 1. Figure 2 This is a scanning electron microscope image of the CD-MOF nanomaterial in Example 1; Figure 3 Fourier near-infrared images of CD-MOF and CL-CD-MOF nanomaterials in Example 1; Figure 4 The X-ray diffraction patterns of CD-MOF and CL-CD-MOF nanomaterials in Example 1 are shown below. Figure 5 This is a comparison diagram of the water stability of CD-MOF and CL-CD-MOF nanomaterials in Example 1; Figure 6 The concentration-absorbance curve of the CL-CD-MOF@color-sensitive dye in Example 1 is shown. Figure 7 This is a color response difference diagram of the olfactory nanocolor sensor based on CL-CD-MOF@color-sensitive dye in Example 2 to gas responses; Figure 8 This is a comparison chart of the gas response of the color-sensitive sensor based on color-sensitive dye and the CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor in Example 2. Figure 9 The results of the olfactory nanocolor sensor based on CL-CD-MOF@color-sensitive dye in Example 3 are used to identify the degree of tea wilting. LSTM model recognition results, training set (a) and prediction set (b); Bi-LSTM model recognition results, training set (c) and prediction set (d). Detailed Implementation

[0021] Various exemplary embodiments of the present invention will now be described in detail. This detailed description should not be considered as a limitation of the present invention, but rather as a more detailed description of certain aspects, features, and embodiments of the present invention.

[0022] It should be understood that the terminology used in this invention is merely for describing particular embodiments and is not intended to limit the invention. Unless otherwise stated, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art.

[0023] While only preferred methods and materials have been described in this invention, any methods and materials similar or equivalent to those described herein may be used in the implementation or testing of this invention. All references to this specification are incorporated by way of citation to disclose and describe the methods and / or materials associated with those references. In the event of any conflict with any incorporated reference, the content of this specification shall prevail.

[0024] Various improvements and variations can be made to the specific embodiments described in this specification without departing from the scope or spirit of the invention, which will be obvious to those skilled in the art.

[0025] Other embodiments derived from this specification will be apparent to those skilled in the art. The specification and embodiments are merely exemplary. Example 1:

[0026] Step 1: Condition optimization of CD-MOF nanomaterials (1) Amount of PEG-4000 added First, 112 mg of potassium hydroxide (KOH) and 324 mg of γ-cyclodextrin powder (CD) were weighed and dissolved in 8 mL of deionized water to obtain a KOH-CD mixed solution. Then, the KOH-CD mixed solution was sonicated in a 300 W ultrasonic device for 20 min. Next, 50, 100, 150, 200, and 250 mg of PEG-4000 were weighed and dissolved in 8 mL of methanol solution to obtain PEG-4000 methanol solutions. The 8 mL PEG-4000 methanol solution was mixed with 8 mL of the KOH-CD mixed solution and incubated in a 30 ℃ water bath for 1 h. The mixture was then washed with methanol and centrifuged. Finally, the washed precipitate was dried (80 ℃) to constant weight to obtain CD-MOF nanomaterials. Figure 1 The attached image in the first row; (2) Temperature of water bath insulation First, 112 mg of potassium hydroxide (KOH) and 324 mg of γ-cyclodextrin powder (CD) were weighed and dissolved in 8 mL of deionized water to obtain a KOH-CD mixed solution. Then, the KOH-CD mixed solution was sonicated in a 300 W ultrasonic device for 20 min. Next, 200 mg of PEG-4000 was weighed and dissolved in 8 mL of methanol solution to obtain a PEG-4000 methanol solution. The 8 mL of PEG-4000 methanol solution was mixed with 8 mL of the KOH-CD mixed solution and incubated in a water bath at 30, 40, 50, 55, and 60 ℃ for 1 h. The mixture was then washed with methanol and centrifuged. Finally, the washed precipitate was dried (80 ℃) to constant weight to obtain CD-MOF nanomaterials. Figure 1 The attached image in the second row; (3) Ultrasound time First, 112 mg of potassium hydroxide (KOH) and 324 mg of γ-cyclodextrin powder (CD) were weighed and dissolved in 8 mL of deionized water to obtain a KOH-CD mixed solution. Then, the KOH-CD mixed solution was sonicated in a 300 W ultrasonic device for 10, 20, 30, 40, and 50 min. Next, 200 mg of PEG-4000 was weighed and dissolved in 8 mL of methanol solution to obtain a PEG-4000 methanol solution. The 8 mL of PEG-4000 methanol solution was mixed with 8 mL of the KOH-CD mixed solution and incubated in a 50 ℃ water bath for 1 h. The mixture was then washed with methanol and centrifuged. Finally, the washed precipitate was dried (80 ℃) to constant weight to obtain CD-MOF nanomaterials. Figure 1 The attached image in the third row of the middle section; Figure 1 These are particle size distribution diagrams of CD-MOF nanomaterials under different synthesis parameters; through Figure 1 It can be seen that the ultrasonic time, the water bath temperature, and the amount of PEG-4000 added all affect the average particle size of CD-MOF nanomaterials. Among them, the CD-MOF nanomaterials synthesized with an ultrasonic time of 20 min, an PEG-4000 addition of 200 mg, and a holding temperature of 50 ℃ have the smallest average particle size of 182.16 nm. The CD-MOF nanomaterials obtained under these optimal conditions are used as materials for subsequent steps.

[0027] Figure 2 These are scanning electron microscope (SEM) images of CD-MOF nanomaterials obtained under optimal conditions; through... Figure 1 It can be seen that the synthesized CD-MOF nanomaterials exhibit a smooth cubic structure, indicating the successful synthesis of CD-MOF nanomaterials.

[0028] Step 2, Preparation of CL-CD-MOF: 500 mg of CD-MOF nanomaterials and 490 mg of diphenyl carbonate were weighed and dissolved in 6.5 mL of N,N-dimethylformamide solution, and 0.4 mL of triethylamine was added as a catalyst. The cross-linking reaction was completed by stirring at 80 °C for 12 h. The mixture was washed three times with ethanol, dried, and the precipitate was collected to obtain CL-CD-MOF. Figure 3 These are Fourier near-infrared images of CD-MOF nanomaterials and CL-CD-MOF; 3340 cm⁻¹ -1 The characteristic absorption peaks of OH vibrational stretching of γ-CD were observed at 1022, 1152, and 1635 cm⁻¹. -1 Characteristic peaks caused by COC, CO, and HOH vibrations were observed at 1746 cm⁻¹; while a peak at 1746 cm⁻¹ was observed in the infrared spectrum of CL-CD-MOF. -1 The presence of the characteristic C=O peak confirms the successful crosslinking of diphenyl carbonate and CD-MOF, indicating the successful synthesis of CL-CD-MOF.

[0029] Figure 4 The images show the X-ray diffraction patterns of CD-MOF nanomaterials and CL-CD-MOF nanomaterials. The characteristic peak positions of the synthesized CD-MOF nanomaterials are consistent with those of the simulated CD-MOF, indicating that the synthesized CD-MOF nanomaterials have high purity and crystallinity. The CL-CD-MOF nanomaterials, while retaining all the characteristic peak positions of CD-MOF, show a significant decrease in diffraction peak intensity, confirming that the crosslinking process did not destroy the crystal framework structure of the MOF. The decrease in diffraction peaks is mainly due to the shielding effect of the crosslinking agent on the crystal diffraction signal.

[0030] Figure 5 This is a comparison of the water stability of CD-MOF nanomaterials and CL-CD-MOF. When 20 mg of CD-MOF and CL-CD-MOF were placed in water and shaken, CD-MOF dissolved while CL-CD-MOF did not, indicating that the water stability of CL-CD-MOF prepared by cross-linking diphenyl carbonate is enhanced.

[0031] Step 3, Preparation of CL-CD-MOF@color-sensitive dyes: First, weigh 20 mg of color-sensitive dye and 80 mg of CL-CD-MOF, and add 10 mL of methanol solution. Stir at room temperature for 24 h. The color-sensitive dyes are color-sensitive dye A (8-(4-nitrophenyl)-4,4-difluoro-2,6-dibromoborondipyrrolemethane, NO2Br2BDP), color-sensitive dye B (5,10,15,20-tetraphenyl-21H,23H-porphyrin manganese(II), TPPMn), color-sensitive dye C (bromothymol blue), color-sensitive dye D (bromocresol green), color-sensitive dye E (methyl red), color-sensitive dye F (bromophenol blue), color-sensitive dye G (cresol red), and color-sensitive dye H (crystal violet). After stirring, centrifuge to collect the precipitate, wash the precipitate three times with methanol, centrifuge and dry (80 ℃, 12 h) to obtain CL-CD-MOF@color-sensitive dye; Step 4, Fabrication of CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor: Based on the principle of chemical sensing using molecular recognition mechanisms, and relying on the intermolecular interactions between different volatile flavor substances and dyes, a multi-channel color-sensitive response array covering multiple recognition modes is constructed to achieve high-dimensional information encoding of complex volatile components. By leveraging the strong electron-withdrawing effect of bromine atoms (-Br) and nitro groups (-NO2) on the fluoroboron pyrrole skeleton, halogen bonds are formed with electron-rich volatile components (such as alkenes, sulfur-containing compounds, and oxygen-containing compounds). The response is achieved by changing the energy level difference between the ground state and excited state of the dye through n→σ* electronic transitions. This scheme selects 8-(4-nitrophenyl)-4,4-difluoro-2,6-dibromoborondipyrrolemethane (NO2Br2BDP) as a representative of this type of dye, and dissolves it in... N,N Preparing a color-sensitive dye solution in dimethylformamide, NO2Br2BDP and... N,N The ratio of dimethylformamide used is 1 mg: 1 mL, denoted as color-sensitive solution A; The empty orbitals of the central metal ion in metalloporphyrins undergo axial coordination with volatile nitrogen-, sulfur-, and oxygen-containing compounds, forming coordination bonds. This induces electron cloud rearrangement in the macrocyclic conjugated system of porphyrins, resulting in characteristic shifts in the Soret and Q band absorption spectra. This scheme selects 5,10,15,20-tetraphenyl-21H,23H-porphyrin manganese(II) (TPPMn) as the representative of this type of dye. It is dissolved in dichloromethane to prepare a color-sensitive dye solution. The ratio of TPPMn to dichloromethane is 1 mg:1 mL, which is denoted as color-sensitive solution B. Based on the hydrogen bonding association and proton transfer reaction between proton donors such as phenolic hydroxyl / carboxyl groups and volatile organic acids / bases, the absorption peak position and absorbance of the colorimetric system are significantly changed. This protocol selects the following six pH indicators, which are dissolved in anhydrous ethanol to prepare stock solutions. The ratio of each dye to ethanol is 1 mg: 1 mL: Bromothymol blue was selected as the dye to obtain color-sensitive solution C. The dye chosen was bromocresol green, resulting in color-sensitive solution D. The dye chosen is methyl red, resulting in color-sensitive solution E; The dye chosen was bromophenol blue, resulting in color-sensitive solution F; The dye chosen was cresol red, resulting in color-sensitive solution G. Crystal violet was selected as the dye, and color-sensitive solution H was obtained; Subsequently, gradient dilutions were performed using a color-sensitive solution to obtain a series of color-sensitive dye detection solutions of different concentrations (1×10⁻⁶). -3 1.25×10 -3 , 2×10 -3 2.5×10 -3 5×10 -3 and 1×10 -2 (mg / mL); then, the color-sensitive dye solution was spectrally scanned and its UV absorbance was measured using a UV-Vis spectrophotometer, and a concentration-absorbance relationship curve was established through linear fitting; Figure 6 The curves show the concentration-absorbance relationship of eight color-sensitive dyes. The coefficient of determination for the linear relationship between the concentration and absorbance of each color-sensitive dye is greater than 0.9411, indicating a good linear relationship.

[0032] 1 mg of CL-CD-MOF@color-sensitive dye A was dissolved in 5 mL of N,N-dimethylformamide solution; 1 mg of CL-CD-MOF@color-sensitive dye B was dissolved in 5 mL of dichloromethane; and 1 mg of CL-CD-MOF@color-sensitive dye C, CL-CD-MOF@color-sensitive dye D, CL-CD-MOF@color-sensitive dye E, CL-CD-MOF@color-sensitive dye F, CL-CD-MOF@color-sensitive dye G, and CL-CD-MOF@color-sensitive dye H powder were dissolved in 5 mL of anhydrous ethanol and mixed using a vortex mixer for 20 min. Subsequently, the CL-CD-MOF@color-sensitive dye solution was spectrally scanned and its UV absorbance was measured using a UV-Vis spectrophotometer. The absorbance values ​​were substituted into the concentration-absorbance curve to calculate the loading of color-sensitive dyes in the CL-CD-MOF@color-sensitive dye solution.

[0033] Calculate the mass of CL-CD-MOF@color-sensitive dye required to contain 1 mg of dye according to the packaging load, and weigh out the CL-CD-MOF@color-sensitive dye powder. Then dissolve it in 1 mL of the corresponding solvent (N,N-dimethylformamide for CL-CD-MOF@color-sensitive dye A; dichloromethane for CL-CD-MOF@color-sensitive dye B; and anhydrous ethanol for CL-CD-MOF@color-sensitive dye C, CL-CD-MOF@color-sensitive dye D, CL-CD-MOF@color-sensitive dye E, CL-CD-MOF@color-sensitive dye F, CL-CD-MOF@color-sensitive dye G, and CL-CD-MOF@color-sensitive dye H), and shake for 20 min. Add 0.2 mL of the shaken solution to the solution. PEG-300 was used to prepare a CL-CD-MOF@color-sensitive dye color-sensitive solution; eight CL-CD-MOF@color-sensitive dyes were fixed on a substrate to prepare a CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor; and the prepared olfactory nano-color-sensitive sensor was stored in a sealed bag for later use.

[0034] If a 54-fold diluted CL-CD-MOF@color-sensitive dye A solution has an absorbance of 0.293 au at 562 nm, according to the concentration-absorbance linear relationship formula: y(absorbance) = 162.51x + 0.0773, x = 0.001327 mg / mL. The concentration of color-sensitive dye A in the original CL-CD-MOF@color-sensitive dye A solution is 0.071674 mg / mL, while the concentration of the original CL-CD-MOF@color-sensitive dye A solution is 0.24 mg / mL. Therefore, 1 mg of CL-CD-MOF@color-sensitive dye A contains 0.299 mg of dye. That is, the mass of CL-CD-MOF@color-sensitive dye required to obtain 1 mg of dye is approximately 3.344 mg. Example 2:

[0035] The sensitivity evaluation of gas detection based on the CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor mainly involves the following steps: Step 1: Sensor fabrication: First, a 1 mg / mL color-sensitive dye solution was prepared, and eight color-sensitive dyes were fixed onto filter paper to create a conventional color-sensitive sensor. Then, a CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor was prepared according to step four of Example 1. The prepared conventional olfactory color-sensitive sensor and olfactory nano-color-sensitive sensor were stored in sealed bags for later use. The CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor prepared based on color-sensitive solution A was designated S1, the CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor prepared based on color-sensitive solution B was designated S2, and the CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor prepared based on color-sensitive solution C was designated S2. The color-sensitive dye olfactory nano-color-sensitive sensor is designated as S3, the CL-CD-MOF color-sensitive dye olfactory nano-color-sensitive sensor prepared based on color-sensitive solution D is designated as S4, the CL-CD-MOF color-sensitive dye olfactory nano-color-sensitive sensor prepared based on color-sensitive solution E is designated as S5, the CL-CD-MOF color-sensitive dye olfactory nano-color-sensitive sensor prepared based on color-sensitive solution F is designated as S6, the CL-CD-MOF color-sensitive dye olfactory nano-color-sensitive sensor prepared based on color-sensitive solution E is designated as S7, and the CL-CD-MOF color-sensitive dye olfactory nano-color-sensitive sensor prepared based on color-sensitive solution F is designated as S8.

[0036] Step Two: Gas Detection Jasmone, nerolidol, and (3Z)-3-hexene benzoate were prepared at concentrations of 0.1, 1, 10, 50, and 100 μg / mL using 1% ethanol. Color information from a conventional color-sensitive sensor and a CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor was recorded before the reaction. The sensors were then placed in reaction chambers with different concentrations of jasmone, nerolidol, and (3Z)-3-hexene benzoate, and reacted for 12 min. Color information from the conventional color-sensitive sensor and the CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor was recorded after the reaction. The position of each color-sensitive unit of the color-sensitive sensor was located using a computer. The images of the color-sensitive sensor before and after the reaction were decomposed into R-channel, G-channel, and B-channel grayscale images. The differences in R, G, and B values ​​of each sensitive unit before and after the reaction were calculated to obtain ΔR, ΔG, and ΔB. The values ​​were then calculated according to ED = (ΔR / ΔG) / ΔB. 2 +ΔG 2 +ΔB 2 ) 1 / 2 Calculate the Euclidean distance; use the Euclidean distance to represent the sensor's response signal value to evaluate the sensor's sensitivity.

[0037] Figure 7The image shows the response characteristics of the CL-CD-MOF@color-sensitive dye olfactory nano-color sensor to three types of gases. The color difference images before and after gas detection reveal differences in the sensor's color response to different gases. This is because the three types of gas molecules differ in polarity, molecular weight, and spatial configuration, resulting in significant differences in their interaction strength and adsorption capacity with the CL-CD-MOF@color-sensitive dye composite material. Consequently, the sensor exhibits varying degrees of color response signals to different gases.

[0038] Therefore, the color changes of the eight color-sensitive points can be combined to form a fingerprint spectrum for each gas, thus distinguishing different gases.

[0039] Figure 8 This chart compares the responses of a conventional olfactory colorimetric sensor based on color-sensitive dyes and a CL-CD-MOF@color-sensitive dye olfactory nano-colorimetric sensor to the same gas concentration. It shows that when detecting the same gas concentration, the CL-CD-MOF@color-sensitive dye olfactory nano-colorimetric sensor exhibits a higher response value, demonstrating higher sensitivity. By calculating the coating amount of the color-sensitive dye in the CL-CD-MOF@color-sensitive dye, and using the same effective color-sensitive dye concentration in both sensors (the conventional olfactory colorimetric sensor and the CL-CD-MOF@color-sensitive dye olfactory nano-colorimetric sensor) for the detection of the same gas concentration, the comparability of the results is effectively enhanced. Example 3:

[0040] The application of CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor for evaluating the degree of withering in black tea mainly includes the following steps: Step 1: Sample Selection The samples selected were tea leaves with different withering times, grown in Jurong City, Jiangsu Province. Based on the moisture content of the withered samples, the degree of withering was divided into insufficient withering, moderate withering, and excessive withering. Withering degree description: The withering degree of black tea is graded according to the national standard GB / T 3580-2018 for black tea processing technology. Black tea with different withering degrees has different moisture content and is accompanied by certain aroma and appearance characteristics. Insufficient wilting: The moisture content of wilted leaves is higher than 64%, the leaf surface has a glossy appearance, the leaves are bright green, and they have a grassy smell; the leaves are brittle, the stems break easily, they clump together when tightly squeezed, and they quickly fall apart when released. Proper wilting: The moisture content of wilted leaves is 60-64%, the leaf surface loses its luster, the leaves turn dark green, and the grassy smell diminishes; the leaves are soft, the stems do not break when broken, they can be tightly clenched into a ball, and slowly fall apart when released; Over-wilted: The moisture content of wilted leaves is less than 60%, the leaf surface loses its luster, the leaves turn dark green, and there is no grassy smell; the leaves are soft, the stems do not break when broken, they are tightly clenched into a ball, and they cannot be separated when released. Step 2: First, use a camera to acquire an image of the CL-CD-MOF@color-sensitive dye olfactory nanocolor-sensitive sensor before the reaction; Subsequently, 0.6 g of each wilted sample was weighed and placed together with the CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor in a reaction vessel. The prepared CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor was fixed at the top of the reaction vessel, and the CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor and the volatile odor substances of the wilted sample were allowed to react fully for 12 min at 25 ℃. Finally, an image of the CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor after the reaction was acquired using a camera, and the obtained image was saved to a computer.

[0041] Step 3: Use a computer to locate the position of each color-sensitive unit of the color sensor; decompose the images of the color sensor before and after the reaction acquired by the camera into grayscale images of the R, G, and B channels, and extract the hue (H), saturation (S), value (V), color brightness (L), red-green value (a), and yellow-blue value (b) of the image; calculate the difference between R, G, B, H, S, V, L, a, and b of each sensitive unit before and after the reaction to obtain ΔR, ΔG, ΔB, ΔH, ΔS, ΔV, ΔL, Δa, and Δb, and then calculate ED = (ΔR / ΔB) * (ΔG ... 2 +ΔG 2 +ΔB 2 ) 1 / 2 Calculate the Euclidean distance. △R, △G, △B, △H, △S, △V, △L, △a, △b, and ED are the feature variables of one color-sensitive unit; 8 color-sensitive units yielded 80 feature variables; 3 samples representing withering degree, 3 batch processing samples, for a total of 200 samples. The 80 feature variables from these 200 samples are combined to obtain the feature matrix S; using the feature matrix S as input and the sample-corresponding quality grade matrix T as output, LSTM and Bi-LSTM evaluation models for the withering degree of black tea are constructed respectively. Figure 9 Figure (a) shows an LSTM model for evaluating the quality of withered tea leaves, built based on sample information collected by a CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor. The training accuracy of the evaluation model is 92.5%. Figure 9(c) is a Bi-LSTM model for evaluating the quality of withered tea leaves, built based on sample information collected by the CL-CD-MOF@color-sensitive dye olfactory nanocolor-sensitive sensor. The training accuracy of the evaluation model is 97%. Compared with the LSTM model, the recognition accuracy of the Bi-LSTM model is improved by 4.5%.

[0042] Step 4: Take 100 withered samples with unknown withering degree, and obtain 80 feature variables of the 100 samples according to the methods described in Steps 2 and 3, forming a feature variable matrix R (R is a 100*80 matrix); call the Bi-LSTM evaluation model constructed in Step 3, take the feature matrix R as the input value, and output matrix Q, which is the quality grade information corresponding to the 100 samples, realizing the rapid evaluation of the withering degree of black tea.

[0043] Figure 9 (b) shows the results of the LSTM evaluation model constructed in step three predicting 100 wilted samples, with a prediction accuracy of 88%. Figure 9 (d) shows the results of predicting 100 withered samples using the Bi-LSTM evaluation model constructed in step three, with a prediction accuracy of 93%. Compared with the LSTM model, the prediction accuracy is improved by 5%. This result confirms that the CL-CD-MOF@color-sensitive dye olfactory nanocolor-sensitive sensor constructed in this invention can achieve rapid evaluation of the degree of withering in black tea, and has better recognition effect when combined with Bi-LSTM.

[0044] Note: The above embodiments are only used to illustrate the present invention and are not intended to limit the technical solutions described in the present invention. Therefore, although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the present invention. All technical solutions and improvements that do not depart from the spirit and scope of the present invention should be covered within the scope of the claims of the present invention.

Claims

1. A method for preparing a green and environmentally friendly olfactory color-sensitive sensor, characterized in that, The steps are as follows: Step 1, Preparation of CD-MOF and CL-CD-MOF nanomaterials: (1) Weigh potassium hydroxide and cyclodextrin and dissolve them in deionized water. After ultrasonic treatment, KOH-CD mixed solution is obtained. Weigh PEG-4000 and dissolve it in methanol to obtain PEG-4000 methanol solution. Then mix the PEG-4000 methanol solution with KOH-CD mixed solution, keep warm in water bath and centrifuge to collect the precipitate. Wash the precipitate with methanol and centrifuge. Finally, dry the washed precipitate to constant weight to obtain CD-MOF nanomaterials. (2) Weigh the CD-MOF nanomaterials and diphenyl carbonate from step (1) and dissolve them in N,N-dimethylformamide solution. Add triethylamine as a catalyst and carry out cross-linking reaction under stirring conditions. After the reaction, collect the precipitate and wash it with ethanol. After washing, centrifuge and dry it to obtain the dried product, which is CL-CD-MOF. Step 2, Preparation of CL-CD-MOF@color-sensitive dyes: The color-sensitive dye and the CL-CD-MOF obtained in step one were added to methanol, stirred at room temperature, and the precipitate was collected by centrifugation. The precipitate was washed with methanol, centrifuged and dried to obtain CL-CD-MOF@color-sensitive dye. Step 3, Fabrication of CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor: S1. First, prepare a series of standard test solutions of color-sensitive dyes at various concentrations; then, use a UV-Vis spectrophotometer to perform spectral scanning on the standard color-sensitive dye solutions of each concentration and record the absorbance values ​​at their characteristic absorption peaks; according to Lambert-Beer's law, i.e. A = εbc , in A Absorbance ε The molar absorptivity is 1. b For optical path, c For concentration, under fixed wavelength and cuvette optical path length conditions, absorbance shows a linear relationship with concentration; ultimately, for concentration... c and absorbance A Perform linear fitting to establish a standard curve; S2. Dissolve the CL-CD-MOF@color-sensitive dye powder obtained in step two in a solvent to obtain a mixed solution; the solvent includes N,N-dimethylformamide and / or dichloromethane and / or anhydrous ethanol; use a UV-Vis spectrophotometer to perform a spectral scan on the mixed solution and measure its absorbance A0 at a characteristic wavelength; substitute the absorbance value A0 into the standard curve established in step S1 to obtain the concentration of color-sensitive dye in the solution, and then calculate the loading of color-sensitive dye in CL-CD-MOF@color-sensitive dye; S3. Based on the loading of color-sensitive dye in CL-CD-MOF@ color-sensitive dye calculated in S2, weigh out the corresponding CL-CD-MOF@ color-sensitive dye powder and dissolve it in a solvent, wherein the solvent includes N,N-dimethylformamide and / or dichloromethane and / or anhydrous ethanol; obtain a shaken solution; then add PEG-300 to the shaken solution to obtain a CL-CD-MOF@ color-sensitive dye solution; fix the CL-CD-MOF@ color-sensitive dye solution on a substrate to fabricate a CL-CD-MOF@ color-sensitive dye olfactory color-sensitive sensor.

2. The method for preparing a CL-CD-MOF@dye olfactory nanocolor-sensitive sensor according to claim 1, characterized in that, In step one (1), the ratio of potassium hydroxide, cyclodextrin, and deionized water is 112 mg: 324 mg: 8 mL; the ultrasonic power of the ultrasonic treatment is 300 W, and the ultrasonic duration is 10-50 min; the ratio of PEG-4000 to methanol is 50 mg-250 mg: 8 mL; the volume ratio of PEG-4000 methanol solution to KOH-CD mixed solution is 1:1; the water bath temperature is 30-50 ℃, and the holding time is 1 h; the number of times methanol is used for washing is 3-5 times.

3. The method for preparing a green and environmentally friendly olfactory color-sensitive sensor according to claim 1, characterized in that, In step (2) of step one, the ratio of CD-MOF nanomaterials, diphenyl carbonate, N,N-dimethylformamide and triethylamine is 0.5 g: 0.49 g: 6.5 mL: 0.2-0.4 mL; the crosslinking reaction is carried out at a temperature of 60-80 °C for 12-24 h, and the washing is performed with ethanol 3-5 times.

4. The method for preparing a green and environmentally friendly olfactory color-sensitive sensor according to claim 1, characterized in that, In step two, the ratio of the color-sensitive dye, CL-CD-MOF, and methanol is 20 mg: 80 mg: 10 mL, and the stirring time is 24 h; the washing with methanol is performed 2-3 times; the color-sensitive dye includes one or more of 8-(4-nitrophenyl)-4,4-difluoro-2,6-dibromoborondipyrrolemethane, 5,10,15,20-tetraphenyl-21H,23H-porphyrin manganese(II), bromothymol blue, bromocresol green, methyl red, bromophenol blue, cresol red, or crystal violet.

5. The method for preparing a green and environmentally friendly olfactory color-sensitive sensor according to claim 1, characterized in that, In step S1 of step three, the concentration range of the color-sensitive dye detection solution is 1×10⁻⁶. -4 ~1×10 -1 mg / mL; X types of color-sensitive dye solutions are prepared, where X is a positive integer, and the X types of color-sensitive solutions include any one or more of solution A, solution B, or solution C; wherein solution A is an N,N-dimethylformamide solution of fluoroborophyrrole, solution B is a dichloromethane solution of metalloporphyrin, and solution C is an ethanol solution of pH indicator; wherein the ratio of fluoroborophyrrole to N,N-dimethylformamide in solution A is 1 mg:1 mL, the ratio of metalloporphyrin to dichloromethane in solution B is 1 mg:1 mL, and the ratio of pH indicator to ethanol in solution C is 1 mg:1 mL; The fluoroboropyrrole comprises 8-(4-nitrophenyl)-4,4-difluoro-2,6-dibromoborondipyrrolemethane; the metalloporphyrin comprises 5,10,15,20-tetraphenyl-21H,23H-porphyrin manganese(II); and the pH indicator comprises bromothymol blue, bromocresol green, methyl red, bromophenol blue, cresol red, and crystal violet.

6. The method for preparing a green and environmentally friendly olfactory color-sensitive sensor according to claim 1, characterized in that, The calculation method for the loading of color-sensitive dye in CL-CD-MOF@color-sensitive dye in step S2 of step 3 is as follows: the mass of CL-CD-MOF@color-sensitive dye is M0, unit: mg; the volume of solvent is V, unit: mL; the concentration of CL-CD-MOF@color-sensitive dye solution is C. 复 Unit: mg / mL; the solvent includes N, N - Dimethylformamide and / or dichloromethane and / or anhydrous ethanol, C 复 = M0 / V; the concentration-absorbance linear relationship formula is... A = k C + b, where k is the slope, representing the change in absorbance caused by a unit change in concentration, and is a real number; b is the intercept, representing the background absorbance value when the concentration is 0, and is a real number; C For concentration, A Absorbance; The absorbance value of the CL-CD-MOF@color-sensitive dye solution was obtained using a UV-Vis spectrophotometer. 测 Substitute A =k C In the + b relationship, the concentration of the color-sensitive dye in the CL-CD-MOF@color-sensitive dye is calculated to be C. 测 mg / mL, C 测 = (A) 测 -b) / k, further yielding that each 1 mg of CL-CD-MOF@color-sensitive dye contains L mg of color-sensitive dye, where L = C 测 / C 复 = (A 测 -b)×V / (k×M0).

7. The method for preparing a green and environmentally friendly olfactory color-sensitive sensor according to claim 1, characterized in that, In step S3, the ratio of CL-CD-MOF@color-sensitive dye powder to solvent is 1 / L mg:1mL; the shaking treatment time is 15-30 min; the volume ratio of the shaken solution to PEG-300 solution is 1 mL:0.2 mL; the amount of CL-CD-MOF@color-sensitive dye solution fixed on the substrate is 1.5-2 μL; the substrate includes cellulose filter paper, silica gel plate, PVDF membrane, and mixed cellulose ester.

8. The use of the green and environmentally friendly olfactory color-sensitive sensor prepared according to any one of claims 1-7 for food quality monitoring, characterized in that, The steps are as follows: (1) Construction of food quality evaluation model: S1. Sample and Grade Classification: Samples with different processing levels are selected. Different processing levels correspond to different volatile odor substances, which will cause the CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor to produce different color changes. The samples include agricultural raw materials, deep-processed agricultural and tea products, aquatic products, etc. The different quality grades include the degree of processing, freshness grade, or different sensory quality grades of the samples. S2. Feature extraction and model training First, an image of the CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor before the reaction was acquired using a camera. Then, the sample was placed in a reaction container, and the CL-CD-MOF@color-sensitive dye olfactory nano-color-sensitive sensor and the sample were stored together in the same sealed reaction container, allowing the volatile odor substances of the color-sensitive sensor and the sample to react fully for a period of time. After the reaction, an image of the olfactory nano-color-sensitive sensor was acquired using a camera, and the obtained images were saved to a computer. The computer was then used to locate the positions of the color-sensitive units in the images before and after the reaction, and their color features were extracted. The difference between the mean gray values ​​of each color-sensitive unit before and after the reaction is obtained, and the difference in the mean gray values ​​is the characteristic variable of the color-sensitive unit. The feature variables of each color-sensitive unit in all samples are concatenated into a feature matrix. This matrix is ​​used as input and the corresponding sample quality level is used as output. The model is trained using a bidirectional long short-term memory recurrent neural network (Bi-LSTM) to obtain the final food quality evaluation model. (2) Rapid evaluation of food quality: The sample to be tested is reacted according to the method described in (1) to obtain the characteristic variables; the characteristic variables of the sample to be tested are substituted into the Bi-LSTM model constructed in (1) to obtain the quality grade information of the sample to be tested, thereby realizing the rapid evaluation of food quality.

9. The use according to claim 8, characterized in that, In step (1) S2, the feature variables are extracted according to the following steps: the coordinate region of each color-sensitive unit on the sensor is located using a computer; the color-sensitive sensor images before and after the reaction are extracted and decomposed into grayscale images of the R, G, and B channels; the values ​​of each color-sensitive unit in hue, saturation, brightness, lightness, red-green components, and yellow-blue components are calculated simultaneously; the channel values ​​after the reaction are subtracted from those before the reaction to obtain ΔR, ΔG, ΔB, ΔH, ΔS, ΔV, ΔL, Δa, and Δb in sequence; and according to ED = (ΔR 2 +ΔG 2 +ΔB 2 ) 1 / 2 Further calculate the Euclidean distance; △R, △G, △B, △H, △S, △V, △L, △a, △b and ED are the characteristic variables of a color-sensitive unit. After the sensor array composed of M color-sensitive units is acquired, a total of N-dimensional characteristic variables are derived, where N=10M; The sample set used to train the food quality assessment model is set to size E, which spans e discretized levels, with f parallel samples under each level, i.e., E = e × f; where e is a positive integer not less than 2, and f and E are both positive integers. The construction process of the Bi-LSTM food quality evaluation model is as follows: Let the feature matrix be S∈R E×N , where E is the number of samples and N is the total number of characteristic variables; The feature matrix S, as the input sequence, is first fed into a bidirectional long short-term memory network for bidirectional temporal modeling, capturing the forward and backward sequence dependencies respectively, and outputting a fused hidden state matrix H∈R. E×D D is the dimension of the hidden state space; subsequently, the fused hidden state matrix H undergoes dimensional transformation and feature recombination through a linear projection layer to obtain the reconstructed output matrix Z: Z = ReLU(W proj ×H + b proj ) Where ReLU is the rectified linear activation function, W proj Let b be the weight parameter matrix of the linear projection layer. proj It is the bias vector; A Dropout random deactivation layer is cascaded after the linear projection layer, randomly zeroing the activation values ​​of some neurons with a probability ratio of 0.15-0.25, thereby enhancing the model's generalization ability and suppressing overfitting. The output of the Dropout layer is then fed into a Softmax normalization exponential layer, outputting the probability matrices P∈R for each category. E×t t represents the number of food quality grade categories; During model training, class cross-entropy is used as the loss metric, and the Adam optimizer is used for end-to-end parameter updates. The parameters to be optimized include: the weight matrices and bias vectors of each gate in the Bi-LSTM, and the W of the linear projection layer. proj With b proj And the classification weights of the Softmax layer; The true quality grade label matrix T∈R is encoded using one-hot encoding. E×t As a supervisory benchmark, the cross-entropy deviation between the predicted outputs Z and T is minimized to construct the Bi-LSTM model for food quality assessment.

10. The use according to claim 8, characterized in that, The rapid evaluation steps for food quality in step (2) are as follows: Based on the signal acquisition and feature extraction process described in step (1), extract the total number of N-dimensional feature variables from the F samples to be tested, and construct the feature matrix R∈R to be evaluated. F×N The Bi-LSTM model for food quality assessment from step (1) is invoked, with the feature matrix R used as the input for forward inference. The model outputs a probability prediction matrix Q∈R. F×t The probability vector in each row of matrix Q is processed by argmax, and the food quality grade of each sample is determined based on the maximum probability index, thus completing a rapid and automated evaluation of food quality.

Images