A small molecule probe based on fluorescent sensing and application thereof

Abstract

The present application relates to a kind of small molecule probe TPE-J based on fluorescent sensing and application.The intelligent real-time monitoring system of nitro explosive of the present application utilizes PPYOLO algorithm, under the assistance of optical camera, captures fluorescence colorimetric image.Fluorescent probe TPE-J is designed and synthesized, for nitro explosive dose sensitivity and visual detection.Electron transfer between PA and probe causes specific response, rapidly quenches original blue fluorescence to non-emission within 5 seconds, detection limit is as low as 1mg / mL.Color change can be integrated into optical camera capture and quantification, the image data generated is automatically processed by deep learning algorithm platform.The sensing system helps to efficiently monitor and high-sensitivity detect PA in various scenarios.Combined with the fluorescent sensing detection platform of deep learning, a new perspective is provided for the efficient portable detection of explosives.

Description

Technical Field

[0001] This invention relates to the field of fluorescent probe technology, and in particular to a small molecule probe based on fluorescence sensing and its applications. Background Technology

[0002] While fluorescence sensing technology offers advantages such as fast response, ease of operation, and visualization, making it well-suited for mobile detection needs (especially in civilian and police scenarios), its reliance on sophisticated optical equipment and highly trained technicians limits its application in mobile detection environments. Furthermore, the complex spectral information further increases detection time, hindering the rapid acquisition of accurate trace quantitative results. Compared to the complex spectral data monitoring process, especially the lengthy real-time recording and analysis, converting spectral data into image information output facilitates precise and rapid data visualization. However, achieving rapid data visualization while reducing manpower consumption and improving detection speed, accuracy, and portability limits the further development of fluorescence sensing technology for nitro explosives.

[0003] Deep learning, which enables machines to learn from massive amounts of data to replicate human intelligence, has proven to be one of the most effective tools for image data processing. The detection of explosives using fluorescence sensing technology generates a large amount of fluorescence image data, which traditional RGB value reading methods process inefficiently and with significantly reduced accuracy. In contrast, deep learning, especially convolutional neural networks (CNNs), excels in automatically extracting image features, recognizing complex patterns and structures in images, and exhibiting superior performance in both speed and accuracy during image processing. Summary of the Invention

[0004] The purpose of this application is to provide a solution that addresses the problems existing in the prior art.

[0005] This application provides a structure for the small molecule probe as shown below:

[0006]

[0007] The small molecule probe is synthesized by degassing a mixture of 3,4-dibromothiophene, (4-(1,2,2-triphenylvinyl)phenyl)boronic acid, toluene, potassium carbonate, distilled water, tetra(triphenylphosphine)palladium and anhydrous ethanol, and stirring and refluxing under nitrogen at 90°C for 24 h.

[0008] The molar ratio of 3,4-dibromothiophene, (4-(1,2,2-triphenylvinyl)phenyl)boronic acid, and potassium carbonate was 1:3:8; the volume ratio of toluene, distilled water, and anhydrous ethanol added was 6:4:3; the amount of toluene used was 0.75 mL per millimole of potassium carbonate; and the equivalent of tetra(triphenylphosphine)palladium added was 0.02 times the equivalent of 3,4-dibromothiophene.

[0009] Application of small molecule probes in detecting explosives containing picric acids.

[0010] A fluorescent sensor, wherein the small molecule probe is immobilized on the fluorescent sensor.

[0011] Furthermore, the fluorescence sensor is a paper-based fluorescence sensor or a hydrogel fluorescence sensor film.

[0012] A portable explosive detection platform includes a laptop computer, a sealed box, a UV light source, an optical camera, and sample vials containing a fluorescence sensor or a small molecule probe.

[0013] A method for quantifying picric acid content based on fluorescence image spectroscopy is proposed. This method involves using a portable explosives detection platform, the aforementioned fluorescence sensor, or the aforementioned small molecule probe to bind picric acid to an image captured by an optical camera. The RGBV values ​​in the image are extracted, and a linear relationship is established between the RGB and HSV values ​​and the picric acid concentration. The qualitative analysis and concentration of picric acid can be determined based on the fluorescence image using this linear relationship formula.

[0014] Furthermore, the image processing employs the PPYOLO model, which receives the input image and uses a deep convolutional neural network for feature extraction. A feature pyramid network and a path aggregation network are utilized to enhance the representation of multi-scale features. Subsequently, the network processes the feature maps to predict the class probability and bounding box coordinates for each region. Anchor boxes are used to match targets, and non-maximum suppression is applied to eliminate overlapping predictions. Finally, post-processing steps are performed to filter and fine-tune the results, yielding the final object detection result.

[0015] A cloud-based intelligent visualization detection system is obtained by deploying the PPYOLO model on a cloud platform.

[0016] The beneficial effects of this invention are as follows: This intelligent real-time monitoring system for nitro explosives utilizes the PPYOLO algorithm to capture fluorescence colorimetric images with the assistance of an optical camera. A fluorescent probe, TPE-J, was designed and synthesized for dose-sensitive and visual detection of nitro explosives (PA). Electron transfer between the PA and the probe induces a specific response, rapidly quenching the original blue fluorescence to non-luminescence within 5 seconds, with a detection limit as low as 1 mg / mL. Color changes can be integrated into the optical camera capture and quantification, and the resulting image data is automatically processed by a deep learning algorithm platform. This sensing system facilitates efficient real-time monitoring and high-sensitivity detection of PAs in various scenarios. The fluorescence sensing detection platform combined with deep learning provides a new perspective for the efficient and portable detection of explosives. Attached Figure Description

[0017] Figure 1 The hydrogen spectrum of the small molecule probe TPE-J is shown.

[0018] Figure 2 This is the carbon spectrum of the small molecule probe TPE-J.

[0019] Figure 3 The synthetic route for the small molecule probe TPE-J is shown.

[0020] Figure 4 The Stern-Volmer plot of I0 / I-1 and picric acid concentration at point TPE-J for the small molecule probe.

[0021] Figure 5 The graph shows the relationship between the fluorescence intensity and wavelength of the small molecule probe TPE-J and picric acid.

[0022] Figure 6 The graph shows the relationship between the quenching intensity of the small molecule probe TPE-J and picric acid and time.

[0023] Figure 7 This is a darkroom image of the fluorescence quenching reaction between the small molecule probe TPE-J and picric acid under ultraviolet light irradiation.

[0024] Figure 8 The structure diagram of the small molecule probe TPE-J and picric acid calculated by DFT theory is shown.

[0025] Figure 9 The reaction mechanism of the small molecule probe TPE-J with picric acid.

[0026] Figure 10 This is a flowchart of a fluorescence image processing method based on the PPYOLO model.

[0027] Figure 11 The image shows the RGBV values ​​obtained using image processing methods at different picric acid concentrations.

[0028] Figure 12 This is a comparison chart showing the concentration values ​​of the analyte detected by image processing methods at different picric acid concentrations, and the actual concentration values ​​of the analyte.

[0029] Figure 13 This is a linear relationship graph of RGBV values ​​obtained using image processing methods at different picric acid concentrations. a, b, c, and d represent the linear relationships between picric acid concentration and the R, G, B, and V values ​​of the photon fluorescence channels, respectively. e represents the functional relationship between picric acid concentration and the photon fluorescence channel value ΔE. f represents a comparison between the analyte concentration value detected using RGBVE and the actual analyte concentration value. Detailed Implementation

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] Unless otherwise specified, the "water" used in the following examples is deionized water.

[0032] In the following tests of this invention, proton nuclear magnetic resonance (NMR) spectra were performed on an AVANCE NEO 400MHz NMR spectrometer from Bruker GmbH, Germany. The solvent used was deuterated chloroform (CDCl3), and the instrument was calibrated with tetramethylsilane (TMS).

[0033] Example 1: Preparation method of small molecule probe TPE-J and its fluorescence quenching response to picric acid.

[0034] 1. The synthesis methods of the small molecule probe TPE-J include:

[0035] A mixture of 3,4-dibromothiophene (483.86 mg, 2 mmol), (4-(1,2,2-triphenylvinyl)phenyl)boronic acid (2.75 g, 6 mmol), toluene (12 mL), potassium carbonate (2.21 g, 16 mmol), distilled water (8 mL), tetra(triphenylphosphine)palladium (92.40 mg, 0.04 mmol), and anhydrous ethanol (6 mL) was degassed and refluxed under nitrogen at 90 °C for 24 h to prepare a small molecule probe with fluorescence response. The probe was then analyzed by proton NMR spectroscopy (1H NMR). Figure 1 ) and carbon spectrum ( Figure 2 The structure of the small molecule probe (TPE-J) obtained by detection is as follows:

[0036]

[0037] Synthesis path as follows Figure 3 As shown. The molar ratio of 3,4-dibromothiophene, (4-(1,2,2-triphenylvinyl)phenyl)boronic acid, and potassium carbonate is 1:3:8; the solvent used is toluene, distilled water, and anhydrous ethanol in a volume ratio of 6:4:3, with 0.75 mL of toluene required per millimole of potassium carbonate; the equivalent of tetra(triphenylphosphine)palladium added is 0.02 times the equivalent of 3,4-dibromothiophene.

[0038] 2. Study on the fluorescence quenching response of the small molecule probe TPE-J to picric acid

[0039] The synthesized probe TPE-J was diluted to 1 μM with THF tetrahydrofuran and subjected to fluorescence spectroscopy experiments. The time-dependent UV-Vis spectroscopic study of TPE-J and picric acid was carried out in an aqueous solution of tetrahydrofuran (tetrahydrofuran:water (volume ratio) = 1:9). The UV-Vis absorption spectra were measured on a Hitachi UV-3900 UV-Vis spectrophotometer.

[0040] The results are as follows: The linear relationship between fluorescence quenching intensity and picric acid concentration is as follows: Figure 4 As shown, the concentration is 1 μM, I represents the peak intensity, and I0 represents the peak intensity without PA. The relationship between fluorescence intensity and wavelength is as follows: Figure 5 As shown, the emission wavelength of TPE-J was determined. Meanwhile, Figure 6 As shown, kinetic measurements revealed that the reaction between TPE-J and picric acid at a final concentration of 40 μg / mL reached near equilibrium within 30 seconds. Fluorescence quenching was visualized using darkroom imaging; the fluorescence image under 365 nm UV light illumination in a darkroom is shown below. Figure 7 As shown.

[0041] Density functional theory (DFT) calculations were performed on the probe TPE-J and picric acid PA using the Gaussian 16 package. Without a CPCM solvation model, the energy level calculation method using B3LYP / 6-31G(d,p) was employed, and the results are as follows. Figure 8As shown, molecular orbital (MO) diagrams and MO levels were calculated at the same theoretical level. In the geometric optimization of the two molecules, the HOMO-LUMO gaps for TPE-J and picric acid were 3.93 eV and 5.45 eV, respectively. Furthermore, PA exhibits a wider band gap and a lower LUMO than TPE-J. Generally, nitro compounds with relatively low LUMO levels can accept excited-state electrons from AIE molecules, thereby quenching the fluorescence of AIE molecules. The LUMO gap between the AIE sensor and the nitro compound is considered the "driving force." The electrostatic potential surface (EPS) of the probe was studied using the natural bond orbital method to evaluate its preferred reaction site for PA. The probe's extrema were observed to be located in the thiophene ring and the adjacent benzene ring portion of the molecule (red region), indicating a strong attraction of this region to PA. Many polynitro-substituted aromatic explosives, being inherently electron-deficient, can bind to electron-rich substances through donor-acceptor interactions.

[0042] like Figure 9 The reaction mechanism of the probe TPE-J with picric acid is as follows: Fluorescence response is achieved through a photoinduced electron transfer reaction between the electron-rich group of the probe (TPE-J) and the electron-deficient group of picric acid. The rapid reaction between the probe and picric acid, accompanied by the release of picric acid solution, allows for a more complete chemical reaction. During photoinduced electron transfer (PET), the excited-state fluorophore donor provides an electron acceptor to the LUMO (luminous induction molecular mass). The electron-rich TPE-J interacts with the electron-deficient nitro aromatic molecule through electrostatic interactions, which is beneficial for highly sensitive explosive detection.

[0043] Example 2: Application of the small molecule probe TPE-J in a fluorescence sensor

[0044] By immobilizing the small molecule probe TPE-J on a support, fluorescent sensors responsive to picric acid-based explosives can be fabricated. These include paper-based fluorescent sensors and hydrogel fluorescent sensor films.

[0045] The paper-based fluorescent sensor is prepared by loading the fluorescently responsive small molecule probe TPE-J onto a paper-based support via solution deposition. The hydrogel fluorescent sensor film is prepared by combining the fluorescently responsive small molecule probe TPE-J with agarose to obtain a hydrogel with a fluorescent probe. The three-dimensional porous structure of the hydrogel is used to adsorb and detect picric acid solid particles, and it also has good mechanical properties.

[0046] The structural formula of agarose is as follows:

[0047]

[0048] A method for preparing a hydrogel film with a fluorescent small molecule probe includes: mixing a fluorescent small molecule probe TPE-J with agarose hydrogel at a mass ratio of 1:200, reacting in water for 5 min, and obtaining a hydrogel with a fluorescent probe.

[0049] Example 3: Portable PA Detection Platform for Nitro Explosives Using the Small Molecule Probe TPE-J

[0050] The portable detection platform includes a laptop computer, a sealed container, a UV light source, an optical camera, and sample vials. The small molecule probe TPE-J or a fluorescence sensor composed of the small molecule probe TPE-J is immobilized in the sample vial. The reaction takes place in the sealed container, which eliminates the influence of external light sources on the fluorescence sensing. A substance containing picric acid is added to the sample vial and photographed under a UV light source. Fluorescence images are obtained by the optical camera, and information is analyzed from the image colors and spectral information to obtain accurate identification and quantitative detection of picric acid-based explosives.

[0051] The color information includes red (R), green (G), blue (B), hue (H), saturation (S), and brightness (V). The captured image is segmented and extracted according to these six image signal channels. The relationship between RGB and HSV values ​​and PA concentration in the sample is established, thus allowing the PA concentration in the sample to be determined from the RGB and HSV values ​​of the image. This part uses the PP-YOLO model.

[0052] like Figure 10 The image processing procedure is as follows: The PPYOLO model receives the input image and uses a deep convolutional neural network for feature extraction. It utilizes a Feature Pyramid Network (FPN) and a Path Aggregation Network (PANet) to enhance the representation of multi-scale features. Subsequently, the network processes the feature maps to predict the class probability and bounding box coordinates for each region. Anchor boxes are used to match targets, and non-maximum suppression (NMS) is applied to eliminate overlapping predictions. Finally, post-processing steps are performed to filter and fine-tune the results, yielding the final object detection result.

[0053] Specifically, it includes the following parts:

[0054] 1. Input processing: Image preprocessing includes resizing the image to a fixed size and normalizing it to suit the network's needs; during the training phase, data augmentation techniques such as random cropping, rotation, and flipping are applied to improve the model's generalization ability.

[0055] 2. Backbone Network: PPYOLO uses a ResNet50 or similar backbone network as a feature extractor. These networks typically include multiple convolutional and pooling layers to capture optical features in the image.

[0056] Residual connections, such as the skip connections introduced in ResNet, can help solve the vanishing gradient problem in deep networks and promote information flow. The backbone uses ResNet50-vd, which has powerful feature extraction capabilities, to extract feature maps at different scales. Furthermore, ResNet50-vd introduces a deformable convolutional network (DCN), which can better handle target deformation and pose changes, helping to improve the accuracy of target recognition and enhance the model's adaptability to targets. This improvement allows PPYOLO to achieve high accuracy while maintaining high efficiency, making it suitable for target detection tasks in diverse scenarios.

[0057] 3. Feature Pyramid Network (FPN): FPN constructs multi-scale feature representations by upsampling feature maps at different levels and fusing information from adjacent levels. It can detect different targets simultaneously. FPN can effectively extract multi-scale features to adapt to the requirements of target detection across different scales. After obtaining multi-level feature maps from the backbone, three resolution feature levels are formed. Images of different scales are input through a 1*1 convolutional layer. First, a top-down path is constructed. An Upsample Block is used in the backbone network to upsample high-level (shallower but semantically richer) feature maps to match the size of low-level (deeper but spatially richer) feature maps. Horizontal connections are introduced, and 1*1 convolutional layers are used to fuse the upsampled features with the corresponding low-level feature maps to match the channel size. The fused feature map is output for the next step of detection and prediction. The fusion of high-level semantic features and low-level spatial features finally yields a multi-scale feature map, which is output through Depthwise Separable Convolution.

[0058] 4. Path Aggregation Network (PANet): PANet further enhances the effect of FPN by adding additional bottom-up paths to strengthen the information transfer from low to high layers.

[0059] 5. Head prediction: Add classification and regression heads to feature maps of different scales generated by FPN / PANet.

[0060] Classification head: Responsible for predicting the probability of the object category present at each location.

[0061] The regression head provides the positional offset and aspect ratio change relative to the anchor boxes for precise target localization. PPYOLO's detection head consists of 3×3 and 1×1 convolutional layers to improve feature extraction and adjust the number of channels for the final prediction. Each final prediction outputs 3(K+5) channels, where K is the number of classes. Each location on each final prediction map is associated with three distinct anchors (described in the next section). For each anchor, the first is the detection head's prediction of the class probability for each location; the probability prediction for the Kth class is output by the first K channels of the convolutional layer. The second part's four channels are responsible for bounding box position prediction, accomplished through the four channels in the middle of the convolutional layer. The last channel is the objective score prediction, typically using the model to predict quantifiable results based on objective criteria. The accuracy of these predictions can be verified by comparing them with the true values.

[0062] 6. Loss Functions: The PPYOLO model uses different types of loss functions to train its classification and localization tasks. To classify objects and train the model to improve the accuracy of class prediction, Cross Entropy Loss measures the difference between the model's predicted class probability distribution and the true class; L1 Loss is sensitive to the absolute value of distance and measures the distance between the predicted bounding box and the true bounding box, accurately locating the target; Objectness Loss monitors whether the model recognizes the presence of the target object. Objectness Loss is usually calculated using binary cross-entropy loss, which quantifies the difference between the model's prediction of object presence and the actual scene. This multi-task learning approach enables the model to complete object detection tasks quickly and accurately.

[0063] 7. Post-processing: threshold filtering, bounding box adjustment, class probability assignment, score reordering, output formatting (img / csv), and making the results visible.

[0064] 8. Output Results: Final Output: The PPYOLO model optimizes the final test results through postprocessing, selecting predictions with higher confidence. The algorithm formats the output results into the desired format (CSV or JPG) and visualizes the detection results (labeling the image with bounding boxes and class labels). Through post-processing steps, PPYOLO ensures that the model can provide high-quality object detection results, suitable for various practical application scenarios.

[0065] Through the above process, PPYOLO can achieve rapid target detection while maintaining high accuracy. This design is particularly suitable for applications requiring real-time performance, such as video surveillance and autonomous driving. In this embodiment, the color change of the probe solution in the sample vial is monitored at a rate of 0.1 seconds per frame. The operator manually selects the target recognition area in the image, and the color extraction function outputs RGB / HSV values. In addition, image data can be automatically recorded and stored according to the timeline to construct a complete dataset, which can then be used as input for subsequent deep learning platforms.

[0066] PA samples at different concentrations (0, 10, 20, 30 μg / mL) were input into the PA detection platform for accuracy verification. The results are as follows: Figure 11 and Figure 12 As shown, the platform can extract R, G, B, V, and E values ​​at various concentrations, and compare the analyte concentration values ​​detected by the platform with the actual analyte concentration values, with an accuracy exceeding 95%. Further analysis of the fluorescence quenching concentration curve of PA established the relationship between image information and target concentration. In addition to deriving RGB signal values ​​from the fluorescence image output, HSV signal values ​​were generated by adding a signal output channel, improving image processing accuracy.

[0067] After the above steps, the image can retrieve the same color data from multiple image processing platforms, such as mobile applications, computer software (Photoshop CC 2018), and cameras, confirming that the extraction of RGBV values ​​is not affected by differences between different image processing tools. Even in scenarios without a dedicated camera, smartphone cameras can still perform the recognition, thus meeting a wider range of application needs.

[0068] The relationship between RGB and HSV values ​​and PA concentration was determined by calibrating a standard curve. The linear relationship between RGB and HSV signal values ​​and target concentration was analyzed experimentally, and the results are shown in Table 1. Figure 13 As shown, the R, G, B, and V values ​​all exhibit a good linear correlation with PA concentration (R0, G1, B2, and V3). 2 >99). This indicates that the detection platform can accurately determine the RGBV values ​​through image analysis, thereby determining the PA concentration. Furthermore, it was found that the correlation between color information and target concentration exhibits two distinct phases. Initially, when the PA concentration is within the range of 0–10 μmg / mL, the fluorescence quenching is significant, leading to a marked change in the RGBV values ​​of the image. Subsequently, as the PA concentration continues to increase outside this range, the change in fluorescence quenching becomes gradual, and the relationship between RGBV values ​​and PA concentration presents a different linear relationship.

[0069] Furthermore, PA concentration can be detected through visual color perception. Changes in visual color perception can be quantified using the following formula:

[0070]

[0071] Where Rn, Bn, Gn, and Vn represent the R, B, G, and V values ​​after adding quenching agent, respectively, and R0, B0, G0, and V0 represent the R, B, G, and V values ​​without adding quenching agent, respectively. The test results are as follows: Figure 13 As shown in e.

[0072] Table 1 Evaluation results of PA at different concentrations

[0073]

[0074] Example 4: Cloud-based Intelligent Visualization Detection System

[0075] The method for color analysis of fluorescence images in the PA detection platform of Example 3 is deployed on a cloud platform. This allows for the coverage of multiple detection platforms and locations through the cloud platform's open interface, standardizing and unifying multiple detection endpoints and enabling real-time data transmission and sharing. This facilitates the establishment of a large database, providing valuable resources for future model training and updates. Furthermore, a cloud website can be established to allow multiple mobile devices to access the cloud platform, increasing the ways to detect explosives. The PPYOLO model is trained and automatically updated based on datasets collected in the cloud, improving target detection efficiency and accuracy.

[0076] In this embodiment, PaddlePaddle is chosen as the open-source platform. The PPYOLO target detection model based on the PaddlePaddle deep learning platform represents an enhanced version of the YOLO (You Only Look Once) algorithm. The YOLO algorithm is widely praised for its speed and performance, and is particularly suitable for real-time processing scenarios, effectively solving the practical needs of PA detection. Based on YOLO, PPYOLO has been optimized and upgraded, improving detection accuracy and speed. This embodiment utilizes the PPYOLO algorithm in the intelligent real-time monitoring system for nitro explosives, capturing fluorescence colorimetric images with the assistance of an optical camera. A fluorescent probe (TPE-J) composed of TPE fragments and thiophene [3,4-b] groups was designed and synthesized for dose-sensitive and visual detection of nitro explosives (PA). Electron transfer between PA and the probe induces a specific response, rapidly quenching the original blue fluorescence to non-luminescence within 5 seconds, with a detection limit as low as 1 mg / mL. The competitive reaction mechanism between the detector and the fluorescence sensor was explored using density functional theory (DFT). Color changes can be integrated into the optical camera capture and quantization, and the generated image data is automatically processed by the deep learning algorithm platform. This sensing system facilitates efficient real-time monitoring and high-sensitivity detection of explosives (PAs) in various scenarios. The portable fluorescence sensing platform, combined with deep learning, offers a new perspective for efficient portable detection of explosives.

[0077] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and not restrictive.

Claims

1. A small molecule probe based on fluorescence sensing, characterized in that, The structure of the small molecule probe is shown below: 。 2. A method for synthesizing a small molecule probe as described in claim 1, characterized in that, The small molecule probe is synthesized by degassing a mixture of 3,6-dibromothiophene[3,2-b]thiophene, (4-(1,2,2-triphenylvinyl)phenyl)boronic acid, toluene, potassium carbonate, distilled water, tetra(triphenylphosphine)palladium and anhydrous ethanol, and stirring and refluxing under nitrogen at 90°C for 24 h.

3. The synthesis method according to claim 2, characterized in that, The molar ratio of 3,6-dibromothiopheno[3,2-b]thiophene, (4-(1,2,2-triphenylvinyl)phenyl)boronic acid, and potassium carbonate was 1:3:8; the volume ratio of toluene, distilled water, and anhydrous ethanol was 6:4:3; the amount of toluene used was 0.75 mL per millimole of potassium carbonate; and the equivalent of tetra(triphenylphosphine)palladium added was 0.02 times the equivalent of 3,6-dibromothiopheno[3,2-b]thiophene.

4. The application of the small molecule probe as described in claim 1 in detecting explosives containing picric acid.

5. A fluorescence sensor, characterized in that, The fluorescent sensor is equipped with the small molecule probe as described in claim 1.

6. The fluorescence sensor according to claim 5, characterized in that, The fluorescence sensor is a paper-based fluorescence sensor or a hydrogel fluorescence sensor film.

7. A portable explosives detection platform, characterized in that, It includes a laptop computer, a sealed box, a UV light source, an optical camera, and a sample vial, wherein the sample vial contains a fluorescence sensor as described in claim 5 or a small molecule probe as described in claim 1.

8. A method for quantifying picric acid content based on fluorescence image spectroscopy, characterized in that, The portable explosive detection platform as described in claim 7, the fluorescence sensor as described in claim 5, or the small molecule probe as described in claim 1 are used to capture images after binding with picric acid using an optical camera. The RGBV values ​​in the images are extracted, and a linear relationship between the RGB and HSV values ​​and the concentration of picric acid is established. The qualitative and concentration detection of picric acid can be calculated based on the fluorescence image using the linear relationship formula.

9. The method according to claim 8, characterized in that, Image processing uses the PPYOLO model, which receives the input image and uses a deep convolutional neural network for feature extraction. We utilize feature pyramid networks and path aggregation networks to enhance the representation capabilities of multi-scale features; Subsequently, the network processes the feature maps to predict the class probability and bounding box coordinates for each region; anchor boxes are used to match targets, and non-maximum suppression is applied to eliminate overlapping predictions; finally, post-processing steps are performed to filter and fine-tune the results to obtain the final object detection results.

Images