In the cultivation of mushroom (Agaricus bisporus), the factory-style environment with constant temperature and humidity, along with mechanized production equipment, is adopted to enhance the efficient production and supply of mushrooms throughout the four seasons. However, the harvesting process still relies heavily on manual labor, which severely restricts industrial development and the improvement of enterprise efficiency. Therefore, there is an urgent need to develop automatic mushroom harvesting machines. This paper conducts systematic research and develops a bioinspired mushroom harvesting robot inspired by the manual mushroom-picking movements. First, the configuration optimization of the mushroom harvesting robot was carried out, based on the requirements of being able to walk between mushroom shelves, reach mushroom beds at different heights, and achieve low weight, low energy consumption, and high efficiency. Secondly, a measurement and control system for the robot was developed to achieve the detection of position and target information, as well as dynamic and precise control of different actuators. Then, a fast algorithm for mushroom recognition, size measurement, and positioning based on YOLOv8n, as well as a compliant control algorithm for mushroom-picking grippers, were successively researched and implemented. Finally, field harvesting experiments were conducted in a mushroom factory. The results showed that the robot's recognition and positioning accuracy was less than 1 cm, the harvesting accuracy reached 99.3%, the mushroom damage rate was less than 3.2%, and the average harvesting efficiency was 20.8 kg per hour-equivalent to the efficiency of one worker. The experimental results verify the rationality of the robot design, the accuracy of the visual algorithm and the robustness of the control, and can effectively complete the automated harvesting of multi-layer mushrooms.

Autonomous navigation for orchard tractors faces significant challenges in robust path planning and reliable operation in unstructured orchard environments. To address these issues, this paper proposes a human-robot collaborative driving navigation method for tractors, based on UAV photogrammetry and safe reinforcement learning. First, rapid semantic segmentation and global path planning are achieved using UAV photogrammetry and an Attention U-Net model. Second, a composite control strategy utilizing APF-MPC-Safe RL-HIL is designed. This approach incorporates an Artificial Potential Field (APF) into Model Predictive Control (MPC) to provide safety-aware baseline path tracking. It further introduces the Twin Delayed Deep Deterministic Policy Gradient (TD3) algorithm to learn residual steering corrections for the MPC controller in complex terrain conditions. Additionally, a Human-in-the-Loop (HIL) safety intervention mechanism is introduced to provide expert corrective actions during unsafe exploration, thereby accelerating reinforcement learning convergence and improving operational safety during the initial exploration phase. The system underwent pre-training and parameter iteration within a Gazebo simulation environment, followed by deployment and validation in real-world orchard scenarios. Experimental results demonstrate that the proposed method exhibits robustness in typical operations, maintaining an average lateral deviation within 0.13 m. This study provides a practical solution for the rapid deployment of autonomous navigation in orchard environments, with both theoretical significance and practical engineering value.

Identification of ancient wood is crucial for historical research and the preservation of cultural heritage. This study focuses on identifying three ancient wood samples from the Ming Dynasty of China using terahertz (THz) technology and a convolutional neural network (CNN). A comparative analysis was conducted to identify these unknown ancient woods by comparing them with three types of modern pine and three types of modern fir. The terahertz absorption coefficients of nine types of wood samples were first calculated, followed by an analysis of their frequency characteristics within the 0.01-2.5 THz spectral range. The THz spectra were then preprocessed using wavelet denoising (WD) and low-pass filtering (LPF). Dimensionality reduction was subsequently applied to the spectral data based on a cumulative variance contribution threshold of 95%. Finally, the CNN model was developed to identify ancient wood species by minimizing root mean square error (RMSE). Results demonstrate that ancient wood samples THKM7 and NTGSMO1 share five characteristic frequencies with modern pine and fir and a consistent upward trend in absorption coefficients. In addition, the absorption coefficient of ancient wood NTSO3 shows significant deviations in frequency points and amplitudes. Furthermore, the CNN prediction results reveal the minimal RMSE values between the ancient wood THKM7 sample and modern pine XS sample (RMSE = 0.0143) and between the ancient wood NTGSMO1 sample and modern fir LS sample (RMSE = 0.0265). Finally, the accuracy of the prediction results was verified by Generalized Regression Neural Network (GRNN) and Random Forest (RF) classifiers. The study integrating THz technology with deep learning provides a research idea for ancient wood identification and can advance scientific research in cultural heritage conservation.

Brown mushrooms are widely consumed globally due to their low calorie content, high nutritional value, and suitability for periodic growth in industrial mushroom houses, offering significant commercial value. Most robotic grippers pick mushrooms based on precise force control, which requires a high-precision force sensor, increasing production costs and potential failure rates. This study presents a fully soft gripper, as the body made of silicon rubber and driven by cable. Its inherent softness, offering a more natural solution for safely picking mushrooms by relying only on simple position control of the servo. Finite element analysis was employed to optimize the cable-driven displacement. Additionally, the gripper can measure mushroom diameters during picking using rough thin-film force sensors and bending sensors attached to the fingers, based on mathematical derivation. Field experiments were conducted with the proposed gripper mounted on a homemade mushroom-harvesting robot to pick medium-sized and large-sized mushrooms. The results demonstrated non-destructive harvesting, an average measurement accuracy of 96.6% for medium mushrooms and 96.1% for large mushrooms, and an average harvesting time of 7.5 s per mushroom. Compared to force-controlled grippers, the proposed cable-driven gripper features a simpler structural design and more efficient control logic, making it highly suitable for industrial applications.

Variable fertilizer application technology is widely used in precision agriculture due to the sustained rapid development of global intelligent agriculture. Centrifugal variable fertilizer-spreading has become increasingly common in large-scale modern agricultural production because of its simple structure, wide spreading capability, high efficiency, and low cost. To improve the accuracy and uniformity of centrifugal fertilizer application, enhance the detection performance of the fertilizer discharge flow sensor, and facilitate effective real-time feedback and online adjustment of application rates, this study introduces an improved lightweight YOLOv5sseg network for a centrifugal fertilizer discharge flow detection system. The proposed method replaces the original YOLOv5s backbone network with a lightweight HGNetV2 network, effectively boosting detection speed and efficiency. Additionally, the model optimizes feature extraction for small granular fertilizers by refining the detection head, significantly reducing the number of model parameters and expanding the sensing field. By introducing an adaptable AKConv structure, the model becomes more versatile, accommodating fertilizers of varying sizes. The inclusion of the iRMB further enhances the model's sensitivity to granular fertilizer shapes, improving segmentation performance. Experimental results showed that the improved YOLOv5s-seg network achieved an FPS of 45 frames/s, a mAP of 95.9 %, a computation of 17.8 G, and a model size of 6.66 MB. Compared to other network models, the proposed model reduced size and computation volume while improving detection accuracy and speed, achieving optimal segmentation effects for granular fertilizers, making it easy to deploy on mobile devices. Validation experiments determined that TIT was set to 35 ms, and the fertilizer discharge port opening ranged from 17 to 25 mm, the total mass of particles counted per second exhibited a linear correlation, indicating stable fertilizer discharge. The maximum detection error between the total particle mass measured by the system and the actual fertilizer discharged was 5.07 %, with a flow rate measurement range of 15.13-54.83 g/s and a correlation coefficient of 0.9984. These results provide a new method for precise detection of centrifugal fertilizer discharge flow, offering a theoretical reference for real-time feedback control in variable fertilizer application technology within precision agriculture.

Robotic harvesting is a growing trend in modern cash crop production. Mechanical damage is important for determining the quality of robotic harvesting, and the manipulator (end-effector) gripping force is a key parameter governing the occurrence of mechanical damage. To realize rapid and accurate estimation of the viscoelastic parameters of asparagus during grasping, obtain real-time optimization of the gripping force, and reduce mechanical damage caused by the end-effector, this study, a prediction model was established for the viscoelastic parameters of asparagus based on a back-propagation neural network. The gripping force was then optimized based on the viscoelastic model. The deformation of asparagus during harvesting was obtained indirectly by analyzing the movement of the end-effector. The creep behavior of asparagus was described using the Burgers model, and the corresponding viscoelastic parameters were obtained. The topological structure and transfer function of the prediction model were determined through repeated training, and the gripping force range was determined based on conditions producing no damage and stable gripping. The model performance results showed that the loading and maximum inference times of the model were 0.06 and 0.09 s, respectively. The accuracy, robustness, and efficiency of the model were good. In clamping tests, the clamping success rate was 90 % and the damage rate was 7.5 %, indicating that the determined gripping force range could meet the requirements for stable and non-destructive harvesting of asparagus. These results provide a theoretical basis for avoiding mechanical damage during grasping of asparagus and can facilitate the high-quality and efficient operation of asparagus harvesting robots.

This article discusses the salt stress in strawberry seedlings under greenhouse conditions in summer. Spectral acquisition equipment was used to obtain spectral data, and the ambient and leaf temperatures were combined to model and analyze the relative chlorophyll content in the strawberry seedling leaves. Four different salt gradients were employed to culture the strawberry seedings: S1 (0 mmol/L NaCl), S2 (50 mmol/L NaCl), S3 (100 mmol/L NaCl), and S4 (150 mmol/L NaCl). The results indicated that the spectral curves of the strawberry seedlings in groups S3 and S4 began to differentiate after day 3 (D3), and their average canopy temperature increased by 2.5 °C and 3.1 °C, respectively. The performance of traditional machine learning models integrating leaf temperature improved by more than 80%. Under each stress treatment, the one-dimensional ResNet model integrated with leaf temperature performed the best, with root mean square and mean absolute errors below 1.7 and 1.5, respectively. These results highlight the potential of incorporating temperature as an additional factor to improve the accuracy of plant stress assessments. By integrating temperature with spectral data, the model enhances the ability to monitor plant health dynamically and provides a more comprehensive understanding of how environmental factors influence plant physiology.

Root phenotype detection is the basis for screening of dominant root and improving of seed breeding. The efficiency of root phenotype detection is restricted by the concealment and complexity of crop roots. In this paper, a new method based on terahertz imaging technology was proposed to detect the phenotypic characteristics of rice root and quantitatively predict root nitrogen content. First, the terahertz imaging spectral data of rice root were preprocessed with de-overlapping, reconstruction, enhancement, and segmentation. Then, the phenotypic characteristics of root length, diameter and surface area of rice roots were extracted according to the refinement algorithm. In addition, three linear regression models were fitted between root phenotype and root nitrogen content. Finally, Convolutional Neural Network (CNN), Genetic Algorithm-Back Propagation Neural Network (GA-BPNN), and Sparrow Search Algorithm-Support Vector Regression (SSA-SVR) were used to predict nitrogen content in rice roots by training and testing terahertz time domain data. Compared with two kinds of root phenotypic analysis software, the average errors of root length, diameter and root surface area calculated in this paper were 10.68 %, 7.29 % and 3.49 %, respectively. The fitting determination coefficients between root length, root surface area and the root nitrogen content were 0.88 and 0.87 after removing three groups of high-nitrogen data and one group of contrast data. The prediction accuracy of SSA-SVR model was 0.99 and the root mean square error was 0.05. Studies show that terahertz imaging technique can provide an effective analytical way for qualitative analysis of root phenotype.

Soybean seeds are susceptible to damage from the Riptortus pedestris, which is a significant factor affecting the quality of soybean seeds. Currently, manual screening methods for soybean seeds are limited to visual inspection, making it difficult to identify seeds that are phenotypically defect-free but have been punctured by stink bugs on the sub-surface. To facilitate the convenient and efficient identification of healthy soybean seeds, this paper proposes a soybean seed pest detection method based on spatial frequency domain imaging combined with RL-SVM. Firstly, soybean optical data is obtained using single integration sphere technique, and the vigor index of soybean seeds is obtained through germination experiments. Then, based on the above two data items using feature extraction algorithms (the successive projections algorithm and the competitive adaptive reweighted sampling algorithm), the characteristic wavelengths of soybeans are identified. Subsequently, the spatial frequency domain imaging technique is used to obtain the sub-surface images of soybean seeds in a forward manner, and the optical coefficients such as the reduced scattering coefficient μ'_s and absorption coefficient μ_a of soybean seeds are inverted. Finally, RL-MLR, RL-GRNN, and RL-SVM prediction models are established based on the ratio of the area of insect-damaged sub-surface to the entire seed, soybean varieties, and μ_a at three wavelengths (502 nm, 813 nm, and 712 nm) for predicting and identifying soybean the stinging and sucking pest damage levels of soybean seeds. The experimental results show that the spatial frequency domain imaging technique yields small errors in the optical coefficients of soybean seeds, with errors of less than 15 μ_a and less than 10 μ'_s . After parameter adjustment through reinforcement learning, the Macro-Recall metrics of each model have improved by 10

Soybean plants form symbiotic nitrogen-fixing nodules with specific rhizobia bacteria. The root hair is the initial infection site for the symbiotic process before the nodules. Since roots and nodules grow in soil and are hard to perceive, little knowledge is available on the process of soybean root hair deformation and nodule development over time. In this study, adaptive microrhizotrons were used to observe root hairs and to investigate detailed root hair deformation and nodule formation subjected to different rhizobia densities. The result showed that the root hair curling angle increased with the increase of rhizobia density. The largest curling angle reached 268° on the 8th day after inoculation. Root hairs were not always straight, even in the uninfected group with a relatively small angle (<45°). The nodule is an organ developed after root hair curling. It was inoculated from curling root hairs and swelled in the root axis on the 15th day after inoculation, with the color changing from light (15th day) to a little dark brown (35th day). There was an error between observing the diameter and the real diameter; thus, a diameter over 1 mm was converted to the real diameter according to the relationship between the perceived diameter and the real diameter. The diameter of the nodule reached 5 mm on the 45th day. Nodule number and curling number were strongly related to rhizobia density with a correlation coefficient of determination of 0.92 and 0.93, respectively. Thus, root hair curling development could be quantified, and nodule number could be estimated through derived formulation.

As brown mushrooms are both delicious and beneficial to health, the global production and consumption of brown mushrooms have increased significantly in recent years. Currently, to ensure the quality of brown mushrooms, selective manual picking is required, and the delicate surface of the mushrooms must not be damaged during the picking process. The labor cost of picking accounts for 50–80% of the total labor cost in the entire production process, and the high-humidity, low-temperature plant environment poses a risk of rheumatism for the laborers. In this paper, we propose a novel underactuated gripper based on a lead screw and linear bearings, capable of operating with flexible force control while simultaneously measuring the diameter of the mushrooms. The gripper features three degrees of freedom: lifting, grasping, and rotation, and enabling it to approach, grasp, and detach the mushroom. A thin-film force sensor is installed on the inner side of the fingers to achieve accurate grip force measurement. The use of a PID algorithm ensures precise grip force control, thereby protecting the brown mushrooms from damage. Experimental results demonstrate that the proposed gripper has a static grasping force error of 0.195 N and an average detachment force overshoot of 1.31 N during the entire picking process. The in situ measurement of the mushroom diameter achieves 97.3% accuracy, with a success rate of 98.3%. These results indicate that the gripper achieves a high success rate in harvesting, a low damage rate, and accurate diameter measurement.

Download This Paper Open PDF in Browser Add Paper to My Library Share: Permalink Using these links will ensure access to this page indefinitely Copy URL Copy DOI

Soybean oil produces harmful substances after long durations of frying. A rapid and nondestructive identification approach for soybean oil was proposed based on photoacoustic spectroscopy and stacking integrated learning. Firstly, a self-designed photoacoustic spectrometer was built for spectral data collection of soybean oil with various frying times. At the same time, the actual free fatty acid content and acid value in soybean oil were measured by the traditional titration experiment, which were the basis for soybean oil quality detection. Next, to eliminate the influence of noise, the spectrum from 1150 cm−1 to 3450 cm−1 was selected to remove noise by ensemble empirical mode decomposition. Then three dimensionality reduction methods of principal component analysis, successive projection algorithm, and competitive adaptive reweighting algorithm were used to reduce the dimension of spectral information to extract the characteristic wavelength. Finally, an integrated model with three weak classifications was used for soybean oil detection by stacking integrated learning. The results showed that three obvious absorption peaks existed at 1747 cm−1, 2858 cm−1, and 2927 cm−1 for soluble sugars and unsaturated oils, and the model based on stacking integrated learning could improve the classification accuracy from 0.9499 to 0.9846. The results prove that photoacoustic spectroscopy has a good detection ability for edible oil quality detection.

[目的]针对现有根系表型检测方法存在价格昂贵、需要专人操作以及无法对根系表型进行原位无损检测等问题,提出一种基于电阻层析成像技术(Electrical resistance tomography,ERT)和深度残差神经网络(Deep residual network,ResNet)的萝卜根系表型无损检测方法.[方法]首先,利用COMSOL软件对萝卜?琼脂场域不同情况的ERT正问题进行仿真分析,并获得大量边界电压数据;然后,基于ResNet对萝卜?琼脂场域的内部电导率分布与边界电压之间的非线性映射关系建立模型,对萝卜?琼脂场域进行图像重建;最后,基于ERT研制一套萝卜根系表型检测装置,并进行试验验证.[结果]基于ERT和ResNet的萝卜根系表型检测方法能够实现萝卜根系表型的可持续无损检测,试验装置操作简单、成本低,图像重建相对误差小于5％.[结论]基于ERT的萝卜根系表型检测方法可以实现对萝卜根系表型的无损检测;结合ResNet算法,成像精度较高.该方法可有效应用于萝卜根系表型的检测.

To satisfy the increasing demand for soybeans, identifying and sorting high-vigor seeds before sowing is an effective way to improve the yield. Polarized hyperspectral imaging (PHI) technology is here proposed as a rapid, non-destructive method for detecting the vigor of naturally aged soybean seeds. First, the spectrum of 396.1–1044.1 nm was collected to automatically extract the region of interest (ROI). Then, first derivative (FD), Savitzky–Golay (SG), multiplicative scatter correction (MSC), and standard normal variate (SNV) preprocessed hyperspectral and polarized hyperspectral data (0°, 45°, 90°, and 135°) for the soybean seeds was obtained. Finally, the seed vigor prediction model based on polarized hyperspectral components such as I, Q, and U was constructed, and partial least squares regression (PLSR), back-propagation neural network (BPNN), generalized regression neural network (GRNN), support vector regression (SVR), random forest (RF), and blending ensemble learning were applied for modeling analysis. The results showed that the prediction accuracy when using PHI was improved to 93.36%, higher than that for the hyperspectral technique, with a prediction accuracy up to 97.17%, 98.25%, and 97.55% when using the polarization component of I, Q, and U, respectively.

Soybean is an important grain and oil crop worldwide and is rich in nutritional value. Phenotypic morphology plays an important role in the selection and breeding of excellent soybean varieties to achieve high yield. Nowadays, the mainstream manual phenotypic measurement has some problems such as strong subjectivity, high labor intensity and slow speed. To address the problems, a three-dimensional (3D) reconstruction method for soybean plants based on structure from motion (SFM) was proposed. First, the 3D point cloud of a soybean plant was reconstructed from multi-view images obtained by a smartphone based on the SFM algorithm. Second, low-pass filtering, Gaussian filtering, Ordinary Least Square (OLS) plane fitting, and Laplacian smoothing were used in fusion to automatically segment point cloud data, such as individual plants, stems, and leaves. Finally, Eleven morphological traits, such as plant height, minimum bounding box volume per plant, leaf projection area, leaf projection length and width, and leaf tilt information, were accurately and nondestructively measured by the proposed an algorithm for leaf phenotype measurement (LPM). Moreover, Support Vector Machine (SVM), Back Propagation Neural Network (BP), and Back Propagation Neural Network (GRNN) prediction models were established to predict and identify soybean plant varieties. The results indicated that, compared with the manual measurement, the root mean square error (RMSE) of plant height, leaf length, and leaf width were 0.9997, 0.2357, and 0.2666 cm, and the mean absolute percentage error (MAPE) were 2.7013%, 1.4706%, and 1.8669%, and the coefficients of determination (R2) were 0.9775, 0.9785, and 0.9487, respectively. The accuracy of predicting plant species according to the six leaf parameters was highest when using GRNN, reaching 0.9211, and the RMSE was 18.3263. Based on the phenotypic traits of plants, the differences between C3, 47-6 and W82 soybeans were analyzed genetically, and because C3 was an insect-resistant line, the trait parametes (minimum box volume per plant, number of leaves, minimum size of single leaf box, leaf projection area).The results show that the proposed method can effectively extract the 3D phenotypic structure information of soybean plants and leaves without loss which has the potential using ability in other plants with dense leaves.

Owing to the ill-conditioned nature of electrical resistivity tomography and the measurement error of the hardware equipment, the reconstructed resistivity distribution image often contains artifacts of varying degrees. Other soft-field imaging technologies, such as electrical impedance tomography and electrical capacitance tomography, also encounter artifacts. Artifacts interfere with the assessment of damaged areas. To eliminate the influence of artifacts on the reconstructed image, a novel artifact elimination algorithm called the fast artifact filtering (FAF) algorithm is proposed. Based on the calculation results of existing algorithms, such as the Newton’s one-step error reconstructor (NOSER) algorithm, the FAF algorithm can remove the damaged areas with low confidence from the potentially damaged areas and only retain the damaged areas with high confidence for final imaging. Several simulation models were used to test the effectiveness of the artifact elimination algorithm proposed in this study. The test results show that the number of artifacts in the final reconstructed image is significantly reduced after the NOSER algorithm is combined with the FAF algorithm. In addition, when the number of finite element model division elements was 4802, the refresh time of a single image increased by approximately 1 ms. A structural health monitoring test for hollow structure is provided. The results show that the FAF also performs well on the measured voltage data.

In this paper, we proposed a nondestructive detection method for egg freshness based on infrared thermal imaging technology. We studied the relationship between egg thermal infrared images (different shell colors and cleanliness levels) and egg freshness under heating conditions. Firstly, we established a finite element model of egg heat conduction to study the optimal heat excitation temperature and time. The relationship between the thermal infrared images of eggs after thermal excitation and egg freshness was further studied. Eight values of the center coordinates and radius of the egg circular edge as well as the long axis, short axis, and eccentric angle of the egg air cell were used as the characteristic parameters for egg freshness detection. After that, four egg freshness detection models, including decision tree, naive Bayes, k-nearest neighbors, and random forest, were constructed, with detection accuracies of 81.82%, 86.03%, 87.16%, and 92.32%, respectively. Finally, we introduced SegNet neural network image segmentation technology to segment the egg thermal infrared images. The SVM egg freshness detection model was established based on the eigenvalues extracted after segmentation. The test results showed that the accuracy of SegNet image segmentation was 98.87%, and the accuracy of egg freshness detection was 94.52%. The results also showed that infrared thermography combined with deep learning algorithms could detect egg freshness with an accuracy of over 94%, providing a new method and technical basis for online detection of egg freshness on industrial assembly lines.

Given the shortcomings of subjectivity and destructiveness associated with traditional methods of measuring food texture parameters (such as fruit firmness and meat tenderness) when using a texture meter, a noncontact measurement method utilizing an air puff combined with structured light imaging (ASLI) was proposed in this study. First, the finite-element simulation method was used to simulate the airflow impact model and optimize the ranges of the angle and distance parameters. The detection device was then developed, and binocular 3-D phenotype acquisition and 3-D feature extraction algorithms were investigated. Denoising, point cloud segmentation, greedy projection triangulation, Delaunay triangulation, and surface fitting algorithms were used to process the point cloud, and concave parameters, such as depth, mapping area, surface area, and volume of the concave area on the beef surface, were obtained. Finally, four typical samples with different meat (beef and chicken) and fruit (kiwifruit and peach) parameters were tested. Five machine-learning methods were used to model and analyze the correlation among texture parameters. The results revealed that the model established by the ensemble learning modeling method with extracted concave parameters possessed the highest accuracy. The prediction accuracy of the shear force values of the beef and chicken samples was 0.92 and 0.91, respectively. The accuracy of meat and fruit grading was 0.942 and 1.0, respectively. Therefore, the noncontact detection method proposed in this study to determine food texture characteristics based on airflow can be used to replace the traditional texture analyzer in food engineering and exhibits good application potential.

This study developed a flexible and wearable paper-based chemoresistive sensor (FWPCS) by modifying a SWCNT-PdNP-polystyrene microsphere (SPPM) composite (SPPM/FWPCS) for the low-cost and online deter-mination of fruit ripeness and corruption. A new method for the batch and low-cost fabrication of SPPM/FWPCSs based on laser direct writing was proposed. The sensing mechanism of FWPCS relies on the electron depletion layer in the sensing composite created by the Schottky barriers among SWCNTs, PdNPs, and the adsorbed ox-ygen, along with the construction of O-2(-). When the SPPM sensing film is exposed to ethylene, trapped electrons are released into the conduction band through oxidation and cleavage of ethylene, causing a decrease in resis-tance. The properties and morphology of the synthesized SPPM composite were investigated by X-ray diffraction (XRD), scanning electron microscopy (SEM), transmission electron microscopy (TEM), and Raman spectroscopy. Additionally, the key parameters for the fabrication of SPPMs/FWPCS related to the sensing performance were optimized. The concentration of C2H4 can be detected down to 100 ppb using the SPPMs/FWPCS at 25 degrees C. Finally, the real-time determination of banana ripeness and corruption verified the feasibility of the sensor, indicating that the SPPMs/FWPCS has prospects in monitoring fruit ripeness and corruption during storage and transportation.