Vegetable seeds are small and diverse in morphology. Under high-speed sowing conditions, missing-seeding detection is affected by seed image blur, uneven illumination, and impurities, leading to low accuracy, slow processing, and poor model generalization. This study proposes an image-registration-based method for detecting missing-seeding in plug trays for small vegetable seeds. The method captures before- and after-sowing images of the same tray, performs high-precision registration to correct angular and positional discrepancies, and applies uneven illumination compensation to reduce reflections and brightness gradients. This enables stable identification of seeds and empty cells. Experiments were conducted with pepper, tomato, and radish seeds at five sowing speeds: 600, 800, 1000, 1200, and 1500 trays/h. The results showed that the detection accuracy was above 96% under all sowing speeds, with an average F1 score of 97.97%. The average F1 scores for single-seed, multi-seed, and missing-seeding detection were 96.57%, 87.09%, and 92.92%, respectively. Single-seed and multi-seed F1 scores improved by 11.67% and 12.70%, compared to YOLO11n, with an average tray detection time of 0.78 s. Under high-speed conditions (1500 trays/h), the F1 score remained at 97.47%, with accuracy decreasing by about 1.3% and F1 score by 1.2%. The system showed high stability under both high and low-speed conditions. The detection performance followed pepper > radish > tomato, with F1 scores above 97.50%, demonstrating strong adaptability for small, irregular seeds. This method enables precise, stable, and fast missing-seeding detection in high-speed sowing, providing an efficient solution for smart seedling technologies.

Flower thinning plays a vital role in peach production, which significantly affects fruit yield and quality. Obtaining precise information about inflorescences is the key to scientific thinning and refined orchard management. However, the accurate detection of peach inflorescence still faces great challenges due to the complex and changeable light conditions, dense occlusion between flowers and significant scale differences in the actual orchard environment. In order to solve these problems, an enhanced YOLOv11s peach inflorescence detection model, termed MDI-YOLOv11, is proposed in this study to achieve accurate and stable recognition of flowers and buds. Considering the characteristics of small target and frequent occlusion in peach inflorescences, a collaborative design of the neck feature fusion structure and the backbone feature attention mechanism is adopted. Specifically, the RFCAConv module is added to the backbone network to increase sensitivity to salient regions, while a P2 layer for small target detection is embedded within the neck network and integrated with the RepGFPN structure to enhance multi-scale feature fusion, thereby improving detection accuracy and adaptability in complex orchard environments. The model’s performance was systematically assessed on a self-built dataset comprising 1,008 images. The dataset labeled 41,962 target instances after sample balancing, including 22,803 flower targets and 19,159 bud targets, covering typical orchard scenes with varying illumination, color characteristics, and high density occlusion. The five-fold cross-validation experiment demonstrated that MDI-YOLOv11 achieved an AP50 of 0.919 and an AR50 of 0.964 for peach tree inflorescences detection, along with a detection time of 13.46 ms per image. 10.97 million parameters, and a model size of 21.51MB, all of which meet practical application requirements. Compared with the YOLOv11s model, the MDI-YOLOv11 model achieved a 0.033 increase in both AP50 and AR50, and the detection performance and model complexity are better than YOLOv11m. Based on the detection results of MDI-YOLOv11, this study generated row-by-row inflorescence density distribution maps that intuitively displayed the spatial density distribution of peach inflorescences. The results indicate that the proposed method enables efficient and accurate detection of peach flowers and the generation of inflorescence density maps, which is expected to provide effective support for refined orchards management.

Hoof slipping in dairy cows is a subtle, transient hoof motion event distinct from lameness or falling, with short duration, limited displacement, and close resemblance to normal gait, making automated detection particularly challenging; relevant methods remain scarce. This study proposes a cascaded detection framework based on improved DeepLabCut and NeuFlow v2 for automated hoof-slipping detection and distance estimation in Holstein dairy cows. The four-stage framework covers hoof key point localization, pixel-level optical flow fusion, motion parameter curve feature extraction, and Random Forest classification. The framework was developed on Dataset 1, which contained 115 single-cow side-view videos. Of these, 31 contained slipping events and 84 were normal walking. It was further assessed on a smaller second-farm dataset of 17 single-cow videos (Dataset 2). ResNet-50 with a Coordinate Attention mechanism was adopted as the backbone, reducing mean four-hoof localization RMSE to 2.80 pixels across five independent training runs, showing a 15.2% improvement over the baseline, and outperforming YOLOv8s-Pose. NeuFlow v2 was applied to extract the localized optical flow from hoof regions, yielding velocity and directional curves from which slipping features were derived. The Random Forest classifier achieved an accuracy of 98.9%, precision of 93.3%, recall of 90.3%, F1 score of 91.8%, and AUC of 0.995, outperforming MViT, SlowFast, and STME. The slipping distance estimation RMSE was 1.22 pixels. With the localisation model retrained on new farm frames, the method reached comparable performance on the second farm, suggesting preliminary cross-farm generalisability that warrants larger-scale validation. The proposed framework provides a non-invasive basis for early hoof-health monitoring and welfare-oriented farm management.

Widespread infestation of pests and pathogens during winter wheat's heading stage poses significant risks to yield loss. In this study, an assessment model of health degree (HD) of winter wheat under field conditions was established by using unmanned aerial vehicle remote sensing (UAV RS) imagery. Firstly, non-photosynthetic features were identified from the UAV RS imagery based on different machine learning methods, including Minimum Distance (MD), Maximum Likelihood Estimation (MLE), and Support Vector Machine (SVM). Classification results indicated that MD demonstrates the best performance, according to the values of Overall Accuracy (0.898), Kappa Coefficient (0.863), and Precision (0.856). Therefore, the inversion model between the proportion of pixels classified as non-photosynthetic features and the corresponding ground truth of the incidence of non-photosynthetic features was established. Coefficient of determination (R2), RMSE (root mean square error), and RRMSE (Relative RMSE) of the inversion model are 0.73, 4.86%, and 19.81%, respectively, demonstrating strong correlation and high accuracy. Subsequently, an assessment model for HD of the wheat field was generated based on the predicted incidence of the non-photosynthetic features, and the conclusion was reached that HD1 (pre-symptoms of the infestation of pests and pathogens) dominated in the wheat field, with the proportion of area as 56.16%, while HD4 and HD5 (severe infestation of pests and pathogens) were negligible, with proportions of area of 2.29% and 17.75%. Finally, the assessment model of HD was used to simulate the precision OSMP (One-Spray-Multiple-Protection), and the agricultural chemical could be reduced to 69.11% of the conventional OSMP operation, which provides theoretical and methodological support for the reduction of agricultural chemicals in the domain of precision agriculture.

To address the issue of reduced comfort and operational accuracy in tractors caused by the clutch engagement process in combination with automatic transmissions, dual-clutch transmissions, and hybrid power tractors, a new method is proposed that considers both the slipping process and synchronous instantaneous control at the moment of engagement. In this method, the optimal engagement process model of the clutch considering the clutch sliding process and synchronous instantaneous control was established, and the optimal trajectory of the clutch engagement process was solved based on the pseudo-spectral method. Then, the optimization results were compared with those obtained without considering the synchronous instantaneous. The results show that the proposed method considering the sliding process and synchronous instantaneous constraint can reduce the frictional loss of the clutch by 9%, and suppress the impact to below 10 m/s³. Finally, this method was applied to the control of tractor starting up, gear shifting and hybrid power mode switching processes. Simulation results demonstrate that this method can be effectively applied to these three operating conditions.

IntroductionLeaf water content is a key physiological indicator of plant growth and health status. Constructing leaf water content estimation models based on spectroscopy is an effective method for monitoring plant physiological conditions.MethodsTo improve the accuracy of leaf water content estimation and develop models applicable to different plants, this study collected 1,680 groups of hyperspectral and water content data from peach tree leaves. Estimation models were established using two methods: “constructing vegetation indices” and “selecting characteristic wavelengths.” The accuracy and number of wavelengths used in each model were systematically evaluated. The optimal model was used to predict the water content of each pixel in the hyperspectral images, achieving visualization of leaf water distribution. Additionally, 244 groups of hyperspectral and water content data from apple tree and lettuce leaves were collected to validate the generalization ability of the optimal model.ResultsResults showed that the optimal models established using the two methods were the linear regression model based on the vegetation index NISDI (3 wavelengths, RP2 = 0.9636, RMSEP=0.0356), and the CARS-RF model (12 wavelengths, RP2 = 0.9861, RMSEP=0.0219). Although the accuracy of the two models was similar, the latter used four times more wavelengths than the former, so the former was chosen as the optimal model. Using the optimal model to estimate the water content of apple tree leaves, the RP2 and RMSEP were 0.9504 and 0.1226, respectively. For lettuce containing only leaf tissue, the RP2 and RMSEP were 0.8211 and 0.1771, respectively.DiscussionThese results indicate that the model has some generalization ability and can accurately estimate the water content of leaves of woody plants in the same family, with some performance degradation across different growth forms. The study results achieved accurate estimation of leaf water content for three types of plants and also provided a reference for establishing plant leaf water content estimation models with generalization ability.

Regenerative agriculture (RA) is nature-based solution which has its core intention to improve soil health through elevated soil organic content and nitrogen, as well as reduced chemicals input and mechanical operations, so as to realise the sustainable development of agriculture, namely enhanced soil carbon sequestration, maximised biodiversity and improved quality of water, vegetation and land-productivity. No-till seeding and ground cover are two critical practices of RA in terms of managing soil, reducing water and wind erosion, and enhancing water infiltration characteristics. However, no-till seeders are regularly blocked by crop residue in RA systems, causing more energy inputs and less crop yield outputs. Anti-blocking methods are essential for the no-till seeding technology (NST) and have been developed over the past years. A comprehensive summary about typical furrow openers and corresponding tool characteristics and current anti-blocking methods for no-till seeding are reported. The anti-blocking methods for NST are introduced in five aspects: residue management before seeding, straw cutting, straw removing, improving straw fluidity and other methods (i.e. automatic navigation and high pressure air shooting). This review also recommends five future research directions about further improving the anti-blocking performance of no-till seeders to promote the development of NST and RA. The study discusses implications for improving the anti-blocking ability of no-till seeders and satisfying the principles of the RA by reducing energy inputs (i.e. less fuel consumption) and raising the outputs of crop yields.

While straw mulching has been recognized for mitigating compaction, the multifactorial effects of straw parameters (content, length, laying modes) under static versus dynamic loads remain poorly quantified. Straw mulching may alter the stress transfer in the soil when applying static or dynamic loads. This study systematically evaluated stress and energy dissipation mechanisms using laboratory simulations: a plate sinkage test and an adapted Proctor test. The results demonstrated that the straw content (0-20 Mg/hm(2)) dominantly governs dissipation efficiency, with maximum stress dissipation ratios of 45.6% (static load >200 kPa) and energy dissipation ratios of 38.64% (dynamic high-energy). Longer straw (0.20 m) and ordered laying modes enhanced stress dispersion only under low static loads, while dynamic loads exhibited weaker dissipation. The study reveals that the damping effect of straw is strongest under low stress static load, so it is necessary to reduce the compaction of agricultural machinery and optimize the allocation of straw, such as 15-20 Mg/hm(2), to alleviate compaction in clay loam soils. These findings can provide actionable insights for designing straw-based soil conservation strategies and improving compaction prediction models in mechanized agriculture.

To address several challenges, including low efficiency, significant damage, and high costs, associated with the manual harvesting of Agaricus bisporus, in this study, a machine vision-based intelligent harvesting device was designed according to its agronomic characteristics and morphological features. This device mainly comprised a frame, camera, truss-type robotic arm, flexible manipulator, and control system. The FES-YOLOv5s deep learning target detection model was used to accurately identify and locate Agaricus bisporus. The harvesting control system, using a Jetson Orin Nano as the main controller, adopted an S-curve acceleration and deceleration motor control algorithm. This algorithm controlled the robotic arm and the flexible manipulator to harvest Agaricus bisporus based on the identification and positioning results. To confirm the impact of vibration on the harvesting process, a stepper motor drive test was conducted using both trapezoidal and S-curve acceleration and deceleration motor control algorithms. The test results showed that the S-curve acceleration and deceleration motor control algorithm exhibited excellent performance in vibration reduction and repeat positioning accuracy. The recognition efficiency and harvesting effectiveness of the intelligent harvesting device were tested using recognition accuracy, harvesting success rate, and damage rate as evaluation metrics. The results showed that the Agaricus bisporus recognition algorithm achieved an average recognition accuracy of 96.72%, with an average missed detection rate of 2.13% and a false detection rate of 1.72%. The harvesting success rate of the intelligent harvesting device was 94.95%, with an average damage rate of 2.67% and an average harvesting yield rate of 87.38%. These results meet the requirements for the intelligent harvesting of Agaricus bisporus and provide insight into the development of intelligent harvesting robots in the industrial production of Agaricus bisporus.

For the past decades, fully automatic transplanting machines have been rapidly flooding into the field of modern agricultural transplanting production, replacing humans to quickly and orderly remove seedlings from plug trays and plant them in the fields. However, transplanting machines cannot identify the biological characteristics of seedlings, and mechanized operations result in inferior seedlings in plug trays being planted in the fields. The agricultural planting wisdom of " planting robust seedlings and eliminating inferior seedlings" is difficult for the transplanter to achieve during mechanized transplanting, resulting in machine operation quality problems such as missed planting and reduced production. This paper proposes a selective transplanting robot solution and designs the Selective Intelligent Seedling Picking Framework (SISPF) based on deep learning. The selective transplanting robot is integrated with the mechanisms of seedling picking, seedling dividing, and planting mechanisms, which realizes the selective operation of the transplanting equipment and is applied in operational production. The experiment showed that the mAP of SISPF is 86.4 %, the weight is 12.4 M, and the Inference time is 28.9 ms. When the detection results are linearly analyzed with the real data, the R2 of SISPF is 0.9975, and the RMSE is 8.27, which are better than the current mainstream target detection algorithms. The field test showed that the average robustness score of seedlings after selective transplanting was 87.70 %, and the missing planting was 2.13 %. Compared with automatic transplanting, selective transplanting improved the robustness score by 18.92 % and reduced the missing planting by 9.91 %. This system meets the requirements of selective transplanting and is of great significance in promoting the intelligent upgrading of transplanting equipment.

IntroductionMonitoring nitrogen nutrition indices is crucial for assessing current wheat growth conditions and guiding nitrogen fertilizer application.MethodsTo estimate the wheat nitrogen nutrition index (NNI) and explore the effects of planting density and nitrogen application rates on NNI, this study employed UAVs to capture multispectral canopy imagery of wheat at key growth stages (tillering, jointing, booting, and filling) under varying planting densities and nitrogen application rates. Vegetation indices were selected using Pearson correlation and feature importance analysis. A Bayesian optimized random forest model was constructed to estimate the NNI.ResultsExperimental results indicate that vegetation indices DVI, MDD, NGI, MEVI, NDVI, EVI, and ENDVI exhibit strong resistance to interference, enabling the construction of highly robust models. The NNI estimation model developed under nitrogen application level N2 (210 kg/hm2) demonstrated optimal performance, with R2 and RMSE values of 0.785 and 0.137, respectively. The NNI estimation model constructed at planting density P1 (1 million plants/hm2) was optimal, with R2 and RMSE of 0.716 and 0.158, respectively. It was also found that NNI generally exhibited an initial increase followed by a decrease as planting density increased.DiscussionThe research findings systematically reveal the patterns of planting density and nitrogen application levels affecting wheat NNI. The constructed NNI estimation model plays a crucial role in assessing wheat growth status and also provides reference for rationally determining planting density and nitrogen application levels for spring wheat.

Highlights: Seven grain storage modes were defined as the empty warehouse, half-full warehouse, new grain addition, self-heated, aeration, summer, and winter, based on changes in inventory status. Seven vector features (RGB, LAB, COR, CCV, LBP, JU, GLCM) were extracted from temperature field cloud maps, and fused as 13 elements for inputs to a GA-BP neural network. Results indicate that the GA- BP neural network classification was more stable and accurate than a BP neural network alone. ABSTRACT. In order to reduce the loss of stored grain, managers regularly inspect product quantity and quality changes to determine grain inventory. The inspection of grain temperature changes is the primary method to detect abnormalities. Previous research proposed a grain storage mode detection method based on temperature field cloud maps. However, this method was inconsistent with the actual storage modes of grain warehouses and is difficult to apply to grain warehouse management. To address this issue, this study redefined the grain storage inventory modes as seven conditions, namely the empty warehouse, half-full warehouse, new grain addition, self-heating, aeration, summer, and winter modes. Meanwhile, a new classification method for grain storage modes was constructed based on a GA-BP neural network by fusing seven new features of temperature field cloud maps. The results showed that the optimal vector feature combination of the GA-BP neural network was COR+LBP+JU+GLCM+RGB+LAB, and F-values of the seven modes mentioned were 98.6%, 95%, 96.3%, 96%, 94.6%, 94.1%, and 93.3%, respectively. Results indicated that the GA-BP neural network yielded higher inventory accuracy and more stable performance for each grain storage mode. Therefore, the grain warehouse managers can use the method proposed in this study to grain inventory more accurately. Keywords: GA-BP, Grain, Inventory, Storage modes, Temperature field cloud map.

Rapid economic growth since the turn of the century has often been accompanied by significant challenges, including fossil fuel depletion, environmental degradation, and energy security concerns. Urgent measures are essential to promote environmentally friendly advancements and adopt sustainable energy solutions. Biomass energy, an important component of renewable energy, stands out as the sole renewable energy source containing carbon and has attracted significant attention from governments and the scientific community worldwide. Attention to biomass conversion technologies and their practical applications has gradually increased. This paper provides an in-depth analysis of the utilization of biomass and its wastes, and systematically introduces the progress of the application of biomass conversion technologies, including biochemical and thermochemical conversion, to provide readers with a clear picture of the technological development. By meticulously summarizing the current status of the application of different products produced by these technologies, it provides a valuable reference for researchers and practitioners in the field of biomass energy, aiming to meet the challenges of clean energy production and biomass waste management, and to mitigate the adverse impacts of human activities on the environment. In addition, this paper explores the application of machine learning in the field of biomass conversion, especially its potential in optimizing the biomass conversion process, improving the accuracy of energy yield prediction, and enhancing process control. Despite challenges such as data quality and model interpretability, developments in machine learning, particularly advances in feature engineering and interpretable AI, promise to address these issues. This study contributes positively to advancing biomass energy technologies.

With the aging of the rural population,standardized and intelligent orchard production has be-come a key development direction for modem orchards.Orchard production currently faces several challenges,including labor shortages,low mechanization levels,and inefficient resource utilization.In-telligent orchard technologies offer potential solutions to these problems by improving productivity,re-ducing costs,and minimizing resource waste.This study systematically reviews the four key technologi-cal areas of intelligent orchards:information perception,intelligent decision-making,precision opera-tions,and intelligent management,and analyzes their current status,future directions,and applications in modem orchards based on recent research and developments at home and abroad.First,information perception technology forms the foundation of intelligent orchard production.By integrating various sensors,drones,and Internet of Things(IoT),orchards can achieve real-time,multidimensional monitor-ing of their environments,crop growth,and operational equipment.Environmental perception technolo-gies cover factors such as climate,soil moisture,and temperature.These data are collected using tools like LiDAR,remote sensing,and soil sensors,helping orchard managers better understand the microen-vironment of the orchard.Crop perception technology monitors the health,growth,and pest infestation status of trees using hyperspectral imaging,infrared technology,and other advanced sensors.This en-ables early interventions to prevent losses in yield or quality.Additionally,operational equipment per-ception technology provides real-time monitoring of the status and performance of agricultural machin-ery,supporting autonomous navigation and precision operations by providing crucial data for optimiz-ing equipment use and ensuring efficient orchard management.By fusing multiple sources of informa-tion,intelligent orchards can monitor and manage their operations across the full lifecycle of the or-chard,from planting to harvest.Secondly,intelligent decision-making systems are essential for achiev-ing smart orchard production.By analyzing and processing the collected data,these systems can opti-mize various orchard production processes such as irrigation,fertilization,flower thinning,pesticide ap-plication,and harvesting.For example,intelligent irrigation systems analyze soil moisture levels and meteorological data to determine the best times and quantities for irrigation,ensuring efficient use of water resources.Fertilization and pesticide application systems adjust the timing and dosage based on the specific growth needs of the trees,promoting healthy growth while reducing the use of fertilizers and pesticides,thus minimizing environmental pollution.Additionally,smart harvesting systems use fruit maturity detection to schedule harvests efficiently,improving productivity while reducing fruit damage.Precision operations are a vital component of smart orchard production.Autonomous naviga-tion technologies allow agricultural machinery to operate autonomously in the complex environments of orchards.Using LiDAR,vision-based navigation,and obstacle avoidance algorithms,machinery can complete tasks safely and efficiently.Precision operations are enhanced by real-time sensor data,which enables machinery to adjust its parameters automatically to ensure accuracy and quality.For example,precision fertilization and pesticide application systems adjust the amount applied based on the actual needs of each tree,improving resource utilization efficiency and production outcomes.In terms of sys-tem integration,intelligent orchards rely on cloud platforms to achieve unmanned and automated man-agement.Orchard inspection robots collect real-time data on tree growth and pest status,which is up-loaded to the cloud for analysis by intelligent decision-making systems.Orchard management robots carry out tasks such as fertilization,flower thinning,and pruning based on the instructions from the in-telligent decision systems,executing complex operations automatically.Harvesting robots,equipped with visual recognition technology and deep learning algorithms,can assess fruit maturity and perform harvesting tasks efficiently.In multi-machine collaborative operation systems,several robots in the or-chard are coordinated through cloud platforms to work together,improving overall efficiency.For exam-ple,during harvest seasons,harvesting robots and transport robots collaborate to ensure that picked fruits are swiftly transported to designated locations,reducing spoilage and enhancing workflow effi-ciency.Finally,this study looks forward to the future development direction of smart orchard technolo-gy and provides specific research ideas.Future smart orchard technology will place greater emphasis on multi-source information fusion,autonomous operation of agricultural machinery,and intelligent man-agement throughout the entire process.By deeply integrating perception data from different sources,or-chard managers can more accurately grasp the production dynamics of the orchard,further improving the scientificity of decision-making.The autonomous operation technology of agricultural machinery will continue to improve,achieving autonomous navigation and operation in more complex environ-ments.The fully intelligent management system will optimize the production process,reduce operating costs,improve the overall production efficiency and fruit quality of the orchard through technologies such as big data analysis and cloud computing,in order to provide reference and guidance for the devel-opment of key technologies in standardized orchards.In conclusion,the future of intelligent orchards lies in the continuous improvement and integration of these technologies.Through the development of more advanced sensing technologies,intelligent decision-making systems,and autonomous machinery,orchards will become more efficient,sustainable,and productive,helping farmers to manage their re-sources better while meeting the demands of modern agriculture.This will not only enhance orchard productivity and fruit quality but also contribute to the overall sustainability and competitiveness of the agricultural sector.

In this study, waste catering oil was processed into a gel (NJ), sesame straw was processed into biochar (ZM), and the two were modified with zinc oxide nanoparticles and phosphoric acid to form NJZ and ZMP, respectively. Both materials were then coupled to obtain a polymer gel-biochar composite (N-Z). The materials were characterised by infrared spectroscopy and scanning electron microscopy, and the adsorption kinetics and isothermal adsorption characteristics of copper(II) ions (Cu2+) in water were analysed. The actual adsorption effect was tested by a soil column leaching experiment. The results showed that the modification and coupling processes strengthened the adsorption properties of the adsorbent materials. The Cu2+ adsorption processes of the five materials (NJ, NJZ, ZM, ZMP and N-Z) were consistent with the quasi-second-order adsorption kinetic model. The maximum Cu2+ adsorption capacities of NJ, NJZ, ZM, ZMP and N-Z were 6.38, 9.04, 25.41, 119.42 and 47.82 mg g-1, respectively. Furthermore, a comparison of the parameters of Langmuir and Freundlich model equations revealed that the adsorption process was homogeneous and involved both single and double layers, which is more consistent with the Langmuir equation. The adsorption efficiency (% removal) of NJZ towards cadmium(II) ions leached from soil reached 90.67%, while that of N-Z reached approximately 50%. The lead(II) ion adsorption efficiencies of the five adsorbents were low. The adsorption efficiencies of ZMP and N-Z towards Cu2+ leached from soil were over 90%, which demonstrates their good application prospects.

Aqueous Zn-ion batteries (AZIBs) have been regarded as promising alternatives to Li-ion batteries due to their advantages, such as low cost, high safety, and environmental friendliness. However, AZIBs face significant challenges in limited stability and lifetime owing to zinc dendrite growth and serious side reactions caused by water molecules in the aqueous electrolyte during cycling. To address these issues, a new eutectic electrolyte based on Zn(ClO4)2·6H2O-N-methylacetamide (ZN) is proposed in this work. Compared with aqueous electrolyte, the ZN eutectic electrolyte containing organic N-methylacetamide could regulate the solvated structure of Zn2+, effectively suppressing zinc dendrite growth and side reactions. As a result, the Zn//NH4V4O10 full cell with the eutectic ZN-1-3 electrolyte demonstrates significantly enhanced cycling stability after 1000 cycles at 1 A g−1. Therefore, this study not only presents a new eutectic electrolyte for zinc-ion batteries but also provides a deep understanding of the influence of Zn2+ solvation structure on the cycle stability, contributing to the exploration of novel electrolytes for high-performance AZIBs.

To improve multi-scale disease detection in orchard environments with complex backgrounds while reducing model complexity, this study proposes FHLE-RTDETR, a lightweight detection model based on RT-DETR. Firstly, a dataset of common peach tree diseases was self-constructed, comprising 6692 images of diseases, including gummosis, wood rot, shot hole disease, anthracnose, and red leaf disease. Through data augmentation, the dataset was expanded to 16,060 images to ensure the robustness of the model. Secondly, the FPCBlock module was designed to replace BasicBlock, utilizing partial convolution to reduce computational redundancy. Additionally, the high-frequency and low-frequency (HiLo) attention mechanism was integrated into the encoder to enhance detail and global feature extraction, improving the model's ability to capture key disease characteristics. Finally, the Esilm-neck feature fusion structure was introduced, combining global heterogeneous kernel selection with multi-scale convolution modules. Efficient upsampling convolution block (EUCB) and fast normalised fusion (FastNF) modules were employed to achieve efficient feature enhancement, reducing computational overhead while further improving detection accuracy. Experimental results demonstrate that FHLE-RTDETR achieved a mean average precision at 50 % IoU (mAP50) of 92.1 %, with precision, recall, and mAP50 increasing by 1.9 %, 2.8 %, and 1.7 %, respectively, outperforming the original RT-DETR. Meanwhile, the model's parameters, giga floating point operations per second (GFLOPs), and size were reduced by 26 %, 25.9 %, and 25.4 %, respectively, with an inference time of only 3.2 milliseconds per image. The proposed model surpasses baseline RT-DETR and other mainstream methods in precision, speed, and computational efficiency, and is capable of detecting diseases under different scales and complex backgrounds.

Biochar can be prepared from biomass through pyrolysis in a low-oxygen or anaerobic environment. The characteristics of biochar are influenced by raw materials and the specific preparation process employed. To enhance the efficacy of biochar, various modification methods can be applied. These primarily include chemical modification, physical modification, and biological modification. This paper conducts a literature review covering the period from 2018 to 2023, focusing on biochar preparation, modification methods, and environmental applications. It is a comprehensive overview of the processes involved in preparing biochar, analyzes its characteristics, and explores various modification techniques. Additionally, the paper discusses the diverse and highly efficient application of biochar in environmental remediation, with a particular emphasis on its use in soil, water, and the atmosphere. While biochar holds significant potential for environmental remediation, a thorough understanding of its mechanism in environmental applications is still lacking. Further in-depth studies are required to elucidate the complex roles biochar plays in various environmental contexts.