Evaluation of Three Portable Optical Sensors for Non-Destructive Diagnosis of Nitrogen Status in Winter Wheat.

The accurate estimation and timely diagnosis of crop nitrogen (N) status can facilitate in-season fertilizer management. In order to evaluate the performance of three leaf and canopy optical sensors in non-destructively diagnosing winter wheat N status, three experiments using seven wheat cultivars and multi-N-treatments (0-360 kg N ha-1) were conducted in the Jiangsu province of China from 2015 to 2018. Two leaf sensors (SPAD 502, Dualex 4 Scientific+) and one canopy sensor (RapidSCAN CS-45) were used to obtain leaf and canopy spectral data, respectively, during the main growth period. Five N indicators (leaf N concentration (LNC), leaf N accumulation (LNA), plant N concentration (PNC), plant N accumulation (PNA), and N nutrition index (NNI)) were measured synchronously. The relationships between the six sensor-based indices (leaf level: SPAD, Chl, Flav, NBI, canopy level: NDRE, NDVI) and five N parameters were established at each growth stages. The results showed that the Dualex-based NBI performed relatively well among four leaf-sensor indices, while NDRE of RS sensor achieved a best performance due to larger sampling area of canopy sensor for five N indicators estimation across different growth stages. The areal agreement of the NNI diagnosis models ranged from 0.54 to 0.71 for SPAD, 0.66 to 0.84 for NBI, and 0.72 to 0.86 for NDRE, and the kappa coefficient ranged from 0.30 to 0.52 for SPAD, 0.42 to 0.72 for NBI, and 0.53 to 0.75 for NDRE across all growth stages. Overall, these results reveal the potential of sensor-based diagnosis models for the rapid and non-destructive diagnosis of N status.

Using an Active Sensor to Develop New Critical Nitrogen Dilution Curve for Winter Wheat.

Critical nitrogen (N) dilution curves (CNDCs) have been developed to describe the dilution dynamic of N and to diagnose N status in plants. In this study, to develop a convenient alternative CNDC determination method, four field experiments involving different N rates (0-360 kg N ha-1) and six wheat varieties were performed at different eco-sites from 2014 to 2019. The normalised difference red-edge (NDRE) index extracted from the RapidSCAN CS-45 (Holland Scientific Inc., Lincoln, NE, USA) sensor was used as a driving factor instead of plant dry matter (PDM) to establish a new alternative winter wheat CNDC. The newly developed CNDC was described by the equation Nc = 0.90NDRE-0.88, when NDRE values were ≤ 0.19 and constant Nc = 3.81%, which was independent of the NDRE values. Compared to PDM-derived CNDC (R2 = 0.73) developed with the same dataset, a comparable precision was obtained using NDRE-derived CNDC (R2 = 0.76) and both CNDCs could accurately discriminate wheat N status. Moreover, the NDRE could be inexpensively and rapidly measured using the active sensor. The relationship between NDRE-derived CNDC and grain yield was also analysed to facilitate in-season N management, and the R2 value reached 0.79 and 0.87 at jointing and booting stages, respectively. The NDRE-based CNDC can be used to effectively diagnose wheat N status and as an alternative approach for non-destructive determination of crop N levels.

Using a Portable Active Sensor to Monitor Growth Parameters and Predict Grain Yield of Winter Wheat.

Rapid and effective acquisition of crop growth information is a crucial step of precision agriculture for making in-season management decisions. Active canopy sensor GreenSeeker (Trimble Navigation Limited, Sunnyvale, CA, USA) is a portable device commonly used for non-destructively obtaining crop growth information. This study intended to expand the applicability of GreenSeeker in monitoring growth status and predicting grain yield of winter wheat (Triticum aestivum L.). Four field experiments with multiple wheat cultivars and N treatments were conducted during 2013⁻2015 for obtaining canopy normalized difference vegetation index (NDVI) and ratio vegetation index (RVI) synchronized with four agronomic parameters: leaf area index (LAI), leaf dry matter (LDM), leaf nitrogen concentration (LNC), and leaf nitrogen accumulation (LNA). Duration models based on NDVI and RVI were developed to monitor these parameters, which indicated that NDVI and RVI explained 80%, 68⁻70%, 10⁻12%, and 67⁻73% of the variability in LAI, LDM, LNC and LNA, respectively. According to the validation results, the relative root mean square error (RRMSE) were all <0.24 and the relative error (RE) were all <23%. Considering the variation among different wheat cultivars, the newly normalized vegetation indices rNDVI (NDVI vs. the NDVI for the highest N rate) and rRVI (RVI vs. the RVI for the highest N rate) were calculated to predict the relative grain yield (RY, the yield vs. the yield for the highest N rate). rNDVI and rRVI explained 77⁻85% of the variability in RY, the RRMSEs were both <0.13 and the REs were both <6.3%. The result demonstrates the feasibility of monitoring growth parameters and predicting grain yield of winter wheat with portable GreenSeeker sensor.

Analysis and Evaluation of the Image Preprocessing Process of a Six-Band Multispectral Camera Mounted on an Unmanned Aerial Vehicle for Winter Wheat Monitoring.

Unmanned aerial vehicle (UAV)-based multispectral sensors have great potential in crop monitoring due to their high flexibility, high spatial resolution, and ease of operation. Image preprocessing, however, is a prerequisite to make full use of the acquired high-quality data in practical applications. Most crop monitoring studies have focused on specific procedures or applications, and there has been little attempt to examine the accuracy of the data preprocessing steps. This study focuses on the preprocessing process of a six-band multispectral camera (Mini-MCA6) mounted on UAVs. First, we have quantified and analyzed the components of sensor error, including noise, vignetting, and lens distortion. Next, different methods of spectral band registration and radiometric correction were evaluated. Then, an appropriate image preprocessing process was proposed. Finally, the applicability and potential for crop monitoring were assessed in terms of accuracy by measurement of the leaf area index (LAI) and the leaf biomass inversion under variable growth conditions during five critical growth stages of winter wheat. The results show that noise and vignetting could be effectively removed via use of correction coefficients in image processing. The widely used Brown model was suitable for lens distortion correction of a Mini-MCA6. Band registration based on ground control points (GCPs) (Root-Mean-Square Error, RMSE = 1.02 pixels) was superior to that using PixelWrench2 (PW2) software (RMSE = 1.82 pixels). For radiometric correction, the accuracy of the empirical linear correction (ELC) method was significantly higher than that of light intensity sensor correction (ILSC) method. The multispectral images that were processed using optimal correction methods were demonstrated to be reliable for estimating LAI and leaf biomass. This study provides a feasible and semi-automatic image preprocessing process for a UAV-based Mini-MCA6, which also serves as a reference for other array-type multispectral sensors. Moreover, the high-quality data generated in this study may stimulate increased interest in remote high-efficiency monitoring of crop growth status.

A Comparative Assessment of Measures of Leaf Nitrogen in Rice Using Two Leaf-Clip Meters.

Accurate estimation and monitoring of crop nitrogen can assist in timely diagnosis and facilitate necessary technical support for fertilizer management. Four experiments, involving three cultivars and six nitrogen (N) treatments, were conducted in southeast China to compare the two leaf-clip meters (Dualex 4 Scientific+, Force-A, Orasy, France; Soil and Plant Analyzer Development (SPAD) meter, Minolta Camera Co., Osaka, Japan) for their ability to measure nitrogen nutrient-related indicators. The results indicated that Chl had a better monitoring accuracy for chlorophyll in per unit leaf area as compared to SPAD value, and there was no saturation to appear under high leaf chlorophyll concentration status. Flavonoids (Flav) presented the advantage of early diagnosis of rice N nutrition status (about one day as compared to SPAD value). As a reliable N nutrient diagnosis indicator, it also improved the estimation accuracy compared with the classical SPAD-based method. The other Dualex value also obtained good monitoring results. Flav was positively correlated with N deficiency, and with higher R2 in panicle initiation and booting stages with low RMSE, respectively; whereas SPAD value was negatively correlated with nitrogen deficiency. Therefore, the Flav-based nitrogen application model was found to provide an early rice nitrogen fertilizer application approach, especially in the panicle initiation and booting stages.

Wireless Channel Propagation Characteristics and Modeling Research in Rice Field Sensor Networks.

Wireless channel propagation characteristics and models are important to ensure the communication quality of wireless sensor networks in agriculture. Wireless channel attenuation experiments were carried out at different node antenna heights (0.8 m, 1.2 m, 1.6 m, and 2.0 m) in the tillering, jointing, and grain filling stages of rice fields. We studied the path loss variation trends at different transmission distances and analyzed the differences between estimated values and measured values of path loss in a free space model and a two-ray model. Regression analysis of measured path loss values was used to establish a one-slope log-distance model and propose a modified two-slope log-distance model. The attenuation speed in wireless channel propagation in rice fields intensified with rice developmental stage and the transmission range had monotone increases with changes in antenna height. The relative error (RE) of estimation in the free space model and the two-ray model under four heights ranged from 6.48⁻15.49% and 2.09⁻13.51%, respectively, and these two models were inadequate for estimating wireless channel path loss in rice fields. The ranges of estimated RE for the one-slope and modified two-slope log-distance models during the three rice developmental stages were 2.40⁻2.25% and 1.89⁻1.31%, respectively. The one-slope and modified two-slope log-distance model had better applicability for modeling of wireless channels in rice fields. The estimated RE values for the modified two-slope log-distance model were all less than 2%, which improved the performance of the one-slope log-distance model. This validates that the modified two-slope log-distance model had better applicability in a rice field environment than the other models. These data provide a basis for modeling of sensor network channels and construction of wireless sensor networks in rice fields. Our results will aid in the design of effective rice field WSNs and increase the transmission quality in rice field sensor networks.

Assessing the Spectral Properties of Sunlit and Shaded Components in Rice Canopies with Near-Ground Imaging Spectroscopy Data.

Monitoring the components of crop canopies with remote sensing can help us understand the within-canopy variation in spectral properties and resolve the sources of uncertainties in the spectroscopic estimation of crop foliar chemistry. To date, the spectral properties of leaves and panicles in crop canopies and the shadow effects on their spectral variation remain poorly understood due to the insufficient spatial resolution of traditional spectroscopy data. To address this issue, we used a near-ground imaging spectroscopy system with high spatial and spectral resolutions to examine the spectral properties of rice leaves and panicles in sunlit and shaded portions of canopies and evaluate the effect of shadows on the relationships between spectral indices of leaves and foliar chlorophyll content. The results demonstrated that the shaded components exhibited lower reflectance amplitude but stronger absorption features than their sunlit counterparts. Specifically, the reflectance spectra of panicles had unique double-peak absorption features in the blue region. Among the examined vegetation indices (VIs), significant differences were found in the photochemical reflectance index (PRI) between leaves and panicles and further differences in the transformed chlorophyll absorption reflectance index (TCARI) between sunlit and shaded components. After an image-level separation of canopy components with these two indices, statistical analyses revealed much higher correlations between canopy chlorophyll content and both PRI and TCARI of shaded leaves than for those of sunlit leaves. In contrast, the red edge chlorophyll index (CIRed-edge) exhibited the strongest correlations with canopy chlorophyll content among all vegetation indices examined regardless of shadows on leaves. These findings represent significant advances in the understanding of rice leaf and panicle spectral properties under natural light conditions and demonstrate the significance of commonly overlooked shaded leaves in the canopy when correlated to canopy chlorophyll content.

Using an Active-Optical Sensor to Develop an Optimal NDVI Dynamic Model for High-Yield Rice Production (Yangtze, China).

The successful development of an optimal canopy vegetation index dynamic model for obtaining higher yield can offer a technical approach for real-time and nondestructive diagnosis of rice (Oryza sativa L) growth and nitrogen (N) nutrition status. In this study, multiple rice cultivars and N treatments of experimental plots were carried out to obtain: normalized difference vegetation index (NDVI), leaf area index (LAI), above-ground dry matter (DM), and grain yield (GY) data. The quantitative relationships between NDVI and these growth indices (e.g., LAI, DM and GY) were analyzed, showing positive correlations. Using the normalized modeling method, an appropriate NDVI simulation model of rice was established based on the normalized NDVI (RNDVI) and relative accumulative growing degree days (RAGDD). The NDVI dynamic model for high-yield production in rice can be expressed by a double logistic model: RNDVI = ( 1 + e - 15.2829 × ( R A G D D i - 0.1944 ) ) - 1 - ( 1 + e - 11.6517 × ( R A G D D i - 1.0267 ) ) - 1 (R2 = 0.8577**), which can be used to accurately predict canopy NDVI dynamic changes during the entire growth period. Considering variation among rice cultivars, we constructed two relative NDVI (RNDVI) dynamic models for Japonica and Indica rice types, with R2 reaching 0.8764** and 0.8874**, respectively. Furthermore, independent experimental data were used to validate the RNDVI dynamic models. The results showed that during the entire growth period, the accuracy (k), precision (R2), and standard deviation of RNDVI dynamic models for the Japonica and Indica cultivars were 0.9991, 1.0170; 0.9084**, 0.8030**; and 0.0232, 0.0170, respectively. These results indicated that RNDVI dynamic models could accurately reflect crop growth and predict dynamic changes in high-yield crop populations, providing a rapid approach for monitoring rice growth status.

Spatio-temporal change of winter wheat yield and its quantitative responses to compound frost-dry events - An example of the Huang-Huai-Hai Plain of China from 2001 to 2020.

Extreme climate events such as frost and drought have great influence on wheat growth and yield. Understanding the effects of frost, drought and compound frost-dry events on wheat growth and yield is of great significance for ensuring national food security. In this study, wheat yield prediction model (SCYMvp) was developed by combining crop growth model (CGM), satellite images and meteorological variables. Wheat yield maps in the Huang-Huai-Hai Plain (HHHP) during 2001-2020 were generated using SCYMvp model. Meanwhile, accumulative frost days (AFD), accumulative dry days (ADD) and accumulative frost-dry days (AFDD) in different growth periods of wheat were calculated, and the effects of frost and drought on wheat yield were quantified by the first difference method and linear mixed model. The results showed that wheat yield increased significantly, while the rising trend was obvious at more than half of the regions. Extreme climate events (ECEs) showed a relatively stable change trend, although the change trend was significant only in a few areas. Compared with frost and drought in the early growth period, ECEs in the middle growth period (spring ECEs) had more negative effects on wheat growth and yield. Wheat yield was negatively correlated with spring ECEs, and yield loss was between 4.6 and 49.8 kg/ha for each 1 d increase of spring ECEs. The effects of spring ECEs on wheat yield were ranked as AFDD > AFD > ADD. The negative effect of ADD on wheat yield in the late growth period was higher than that in the other periods. The negative effects of spring ECEs on yield in southern regions were higher than those in northern regions. Overall, due to the adverse effects of frost and drought on wheat yield in the middle and late growth periods, the mean annual yield loss was 6.4 %, among which spring AFD caused the greatest loss to wheat yield (3.1 %). The results have important guiding significance for formulating climate adaptation management strategies.

A method for small-sized wheat seedlings detection: from annotation mode to model construction.

The number of seedlings is an important indicator that reflects the size of the wheat population during the seedling stage. Researchers increasingly use deep learning to detect and count wheat seedlings from unmanned aerial vehicle (UAV) images. However, due to the small size and diverse postures of wheat seedlings, it can be challenging to estimate their numbers accurately during the seedling stage. In most related works in wheat seedling detection, they label the whole plant, often resulting in a higher proportion of soil background within the annotated bounding boxes. This imbalance between wheat seedlings and soil background in the annotated bounding boxes decreases the detection performance. This study proposes a wheat seedling detection method based on a local annotation instead of a global annotation. Moreover, the detection model is also improved by replacing convolutional and pooling layers with the Space-to-depth Conv module and adding a micro-scale detection layer in the YOLOv5 head network to better extract small-scale features in these small annotation boxes. The optimization of the detection model can reduce the number of error detections caused by leaf occlusion between wheat seedlings and the small size of wheat seedlings. The results show that the proposed method achieves a detection accuracy of 90.1%, outperforming other state-of-the-art detection methods. The proposed method provides a reference for future wheat seedling detection and yield prediction.

A deep learning approach for deriving wheat phenology from near-surface RGB image series using spatiotemporal fusion.

Accurate monitoring of wheat phenological stages is essential for effective crop management and informed agricultural decision-making. Traditional methods often rely on labour-intensive field surveys, which are prone to subjective bias and limited temporal resolution. To address these challenges, this study explores the potential of near-surface cameras combined with an advanced deep-learning approach to derive wheat phenological stages from high-quality, real-time RGB image series. Three deep learning models based on three different spatiotemporal feature fusion methods, namely sequential fusion, synchronous fusion, and parallel fusion, were constructed and evaluated for deriving wheat phenological stages with these near-surface RGB image series. Moreover, the impact of different image resolutions, capture perspectives, and model training strategies on the performance of deep learning models was also investigated. The results indicate that the model using the sequential fusion method is optimal, with an overall accuracy (OA) of 0.935, a mean absolute error (MAE) of 0.069, F1-score (F1) of 0.936, and kappa coefficients (Kappa) of 0.924 in wheat phenological stages. Besides, the enhanced image resolution of 512 × 512 pixels and a suitable image capture perspective, specifically a sensor viewing angle of 40° to 60° vertically, introduce more effective features for phenological stage detection, thereby enhancing the model's accuracy. Furthermore, concerning the model training, applying a two-step fine-tuning strategy will also enhance the model's robustness to random variations in perspective. This research introduces an innovative approach for real-time phenological stage detection and provides a solid foundation for precision agriculture. By accurately deriving critical phenological stages, the methodology developed in this study supports the optimization of crop management practices, which may result in improved resource efficiency and sustainability across diverse agricultural settings. The implications of this work extend beyond wheat, offering a scalable solution that can be adapted to monitor other crops, thereby contributing to more efficient and sustainable agricultural systems.

Design and implementation of a portable snapshot multispectral imaging crop-growth sensor.

The timely and accurate acquisition of crop-growth information is a prerequisite for implementing intelligent crop-growth management, and portable multispectral imaging devices offer reliable tools for monitoring field-scale crop growth. To meet the demand for obtaining crop spectra information over a wide band range and to achieve the real-time interpretation of multiple growth characteristics, we developed a novel portable snapshot multispectral imaging crop-growth sensor (PSMICGS) based on the spectral sensing of crop growth. A wide-band co-optical path imaging system utilizing mosaic filter spectroscopy combined with dichroic mirror beam separation is designed to acquire crop spectra information over a wide band range and enhance the device's portability and integration. Additionally, a sensor information and crop growth monitoring model, coupled with a processor system based on an embedded control module, is developed to enable the real-time interpretation of the aboveground biomass (AGB) and leaf area index (LAI) of rice and wheat. Field experiments showed that the prediction models for rice AGB and LAI, constructed using the PSMICGS, had determination coefficients (R²) of 0.7 and root mean square error (RMSE) values of 1.611 t/ha and 1.051, respectively. For wheat, the AGB and LAI prediction models had R² values of 0.72 and 0.76, respectively, and RMSE values of 1.711 t/ha and 0.773, respectively. In summary, this research provides a foundational tool for monitoring field-scale crop growth, which is important for promoting high-quality and high-yield crops.

Comparison and intercalibration of vegetation indices from different sensors for monitoring above-ground plant nitrogen uptake in winter wheat.

Various sensors have been used to obtain the canopy spectral reflectance for monitoring above-ground plant nitrogen (N) uptake in winter wheat. Comparison and intercalibration of spectral reflectance and vegetation indices derived from different sensors are important for multi-sensor data fusion and utilization. In this study, the spectral reflectance and its derived vegetation indices from three ground-based sensors (ASD Field Spec Pro spectrometer, CropScan MSR 16 and GreenSeeker RT 100) in six winter wheat field experiments were compared. Then, the best sensor (ASD) and its normalized difference vegetation index (NDVI (807, 736)) for estimating above-ground plant N uptake were determined (R2 of 0.885 and RMSE of 1.440 g·N·m(-2) for model calibration). In order to better utilize the spectral reflectance from the three sensors, intercalibration models for vegetation indices based on different sensors were developed. The results indicated that the vegetation indices from different sensors could be intercalibrated, which should promote application of data fusion and make monitoring of above-ground plant N uptake more precise and accurate.

TRiP: a transfer learning based rice disease phenotype recognition platform using SENet and microservices.

Classification of rice disease is one significant research topics in rice phenotyping. Recognition of rice diseases such as Bacterialblight, Blast, Brownspot, Leaf smut, and Tungro are a critical research field in rice phenotyping. However, accurately identifying these diseases is a challenging issue due to their high phenotypic similarity. To address this challenge, we propose a rice disease phenotype identification framework which utilizing the transfer learning and SENet with attention mechanism on the cloud platform. The pre-trained parameters are transferred to the SENet network for parameters optimization. To capture distinctive features of rice diseases, the attention mechanism is applied for feature extracting. Experiment test and comparative analysis are conducted on the real rice disease datasets. The experimental results show that the accuracy of our method reaches 0.9573. Furthermore, we implemented a rice disease phenotype recognition platform based microservices architecture and deployed it on the cloud, which can provide rice disease phenotype recognition task as a service for easy usage.

Improving the spatial and temporal estimation of ecosystem respiration using multi-source data and machine learning methods in a rainfed winter wheat cropland.

The investigation of ecosystem respiration (RE) and its vital influential factors along with the timely and accurate detection of spatiotemporal variations in RE are essential for guiding agricultural production planning. RE observation in the plot region is primarily based on the laborious chamber method. However, upscaling the spatial-temporal estimates of RE at the canopy scale is still challenging. The present study conducted a field experiment to determine RE using the chamber method. A multi-rotor unmanned aerial vehicle (UAV) equipped with a multispectral camera was employed to acquire the canopy spectral data of wheat during each RE test experiment. Moreover, the agronomic indicators of aboveground plant biomass, leaf area index, leaf dry mass as well as agrometeorological and soil data were measured simultaneously. The study analyzed the potential of multi-information for estimating RE at the field scale and proposed two strategies for RE estimation. In addition, a semiempirical, yet Lloyd and Taylor-based, remote sensing model (LT1-NIRV) was developed for estimating RE observed across different growth stages with a small margin of error (coefficient of determination [R2] = 0.60-0.64, root-mean-square error [RMSE] = 285.98-316.19 mg m-2 h-1). Further, five machine learning (ML) algorithms were utilized to independently estimate RE using two different datasets. The rigorous analyses, which included statistical comparison and cross-validation for estimating RE, confirmed that the XGBoost model, with the highest R2 and lowest RMSE (R2 = 0.88 and RMSE = 172.70 mg m-2 h-1), performed the best among the evaluated ML models. The LT1-NIRV model was less effective in estimating RE compared with the other ML models. Based on this comprehensive comparison analysis, the ML model can successfully estimate variations in wheat field RE using high-resolution UAV multispectral images and environmental factors from the wheat cropland system, thereby providing a valuable reference for monitoring and upscaling RE observations.

Potential of Establishing the Universal Critical Nitrogen Dilution Curve for Japonica Rice.

Establishing the universal critical nitrogen (NC) dilution curve can assist in crop N diagnosis at the regional scale. This study conducted 10-year N fertilizer experiments in Yangtze River Reaches to establish universal NC dilution curves for Japonica rice based on simple data-mixing (SDM), random forest algorithm (RFA), and Bayesian hierarchical model (BHM), respectively. Results showed that parameters a and b were affected by the genetic and environmental conditions. Based on RFA, highly related factors of a (plant height, specific leaf area at tillering end, and maximum dry matter weight during vegetative growth period) and b (accumulated growing degree days at tillering end, stem-leaf ratio at tillering end, and maximum leaf area index during vegetative growth period) were successfully applied to establish the universal curve. In addition, representative values (most probable number [MPN]) were selected from posterior distributions obtained by the BHM approach to explore universal parameters a and b. The universal curves established by SDM, RFA, and BHM-MPN were verified to have a strong N diagnostic capacity (N nutrition index validation R 2 ≥ 0.81). In summary, compared with the SDM approach, RFA and BHM-MPN can greatly simplify the modeling process (e.g., defining N-limiting or non-N-limiting groups) while maintaining a good accuracy, which are more conducive to the application and promotion at the regional scale.

Small and Oriented Wheat Spike Detection at the Filling and Maturity Stages Based on WheatNet.

Accurate wheat spike detection is crucial in wheat field phenotyping for precision farming. Advances in artificial intelligence have enabled deep learning models to improve the accuracy of detecting wheat spikes. However, wheat growth is a dynamic process characterized by important changes in the color feature of wheat spikes and the background. Existing models for wheat spike detection are typically designed for a specific growth stage. Their adaptability to other growth stages or field scenes is limited. Such models cannot detect wheat spikes accurately caused by the difference in color, size, and morphological features between growth stages. This paper proposes WheatNet to detect small and oriented wheat spikes from the filling to the maturity stage. WheatNet constructs a Transform Network to reduce the effect of differences in the color features of spikes at the filling and maturity stages on detection accuracy. Moreover, a Detection Network is designed to improve wheat spike detection capability. A Circle Smooth Label is proposed to classify wheat spike angles in drone imagery. A new micro-scale detection layer is added to the network to extract the features of small spikes. Localization loss is improved by Complete Intersection over Union to reduce the impact of the background. The results show that WheatNet can achieve greater accuracy than classical detection methods. The detection accuracy with average precision of spike detection at the filling stage is 90.1%, while it is 88.6% at the maturity stage. It suggests that WheatNet is a promising tool for detection of wheat spikes.

Improving multi-scale detection layers in the deep learning network for wheat spike detection based on interpretive analysis.

BACKGROUND: Detecting and counting wheat spikes is essential for predicting and measuring wheat yield. However, current wheat spike detection researches often directly apply the new network structure. There are few studies that can combine the prior knowledge of wheat spike size characteristics to design a suitable wheat spike detection model. It remains unclear whether the complex detection layers of the network play their intended role. RESULTS: This study proposes an interpretive analysis method for quantitatively evaluating the role of three-scale detection layers in a deep learning-based wheat spike detection model. The attention scores in each detection layer of the YOLOv5 network are calculated using the Gradient-weighted Class Activation Mapping (Grad-CAM) algorithm, which compares the prior labeled wheat spike bounding boxes with the attention areas of the network. By refining the multi-scale detection layers using the attention scores, a better wheat spike detection network is obtained. The experiments on the Global Wheat Head Detection (GWHD) dataset show that the large-scale detection layer performs poorly, while the medium-scale detection layer performs best among the three-scale detection layers. Consequently, the large-scale detection layer is removed, a micro-scale detection layer is added, and the feature extraction ability in the medium-scale detection layer is enhanced. The refined model increases the detection accuracy and reduces the network complexity by decreasing the network parameters. CONCLUSION: The proposed interpretive analysis method to evaluate the contribution of different detection layers in the wheat spike detection network and provide a correct network improvement scheme. The findings of this study will offer a useful reference for future applications of deep network refinement in this field.

Comparison of six risk scores for stroke-associated pneumonia in patients with acute ischemic stroke: A systematic review and Bayesian network meta-analysis.

Background: Stroke-associated pneumonia (SAP) is one of the major causes of death after suffering a stroke. Several scoring systems have been developed for the early prediction of SAP. However, it is unclear which scoring system is more suitable as a risk prediction tool. We performed this Bayesian network meta-analysis to compare the prediction accuracy of these scoring systems. Methods: Seven databases were searched from their inception up to April 8, 2022. The risk of bias assessment of included study was evaluated by the QUADAS-C tool. Then, a Bayesian network meta-analysis (NMA) was performed by R 4.1.3 and STATA 17.0 software. The surface under the cumulative ranking curve (SUCRA) probability values were applied to rank the examined scoring systems. Results: A total of 20 cohort studies involving 42,236 participants were included in this analysis. The results of the NMA showed that AIS-APS had excellent performance in prediction accuracy for SAP than Chumbler (MD = 0.030, 95%CI: 0.004, 0.054), A2DS2 (MD = 0.041, 95% CI: 0.023, 0.059), ISAN (MD = 0.045, 95% CI: 0.022, 0.069), Kwon (MD = 0.077, 95% CI: 0.055, 0.099) and PANTHERIS (MD = 0.082, 95% CI: 0.049, 0.114). Based on SUCRA values, AIS-APS (SUCRA: 99.8%) ranked the highest. Conclusion: In conclusion, the study found that the AIS-APS is a validated clinical tool for predicting SAP after the onset of acute ischemic stroke. Systematic review registration: https://www.crd.york.ac.uk/PROSPERO/display_record.php?RecordID=292375, identifier: CRD42021292375.

Study about photoinduction-based technology for trapping small brown planthopper.

To provide a theory to guide the selection of the illumination parameters of light emitting diode (LED)-based light sources used for trapping Laodelphax striatellus, we used LED light sources and devices built in-house to detect L. striatellus phototactic behavior. Through phototaxis screening experiments of different light sources and the comparative experimental method, we analyzed the response patterns of L. striatellus to wavelength, light intensity, layout, flash frequency of monochromatic light sources, as well as combined color light sources, and discussed the mechanisms of the phototactic behavior of L. striatellus under different light sources. The results of the monochromatic light experiment showed that the trapping rate of the L. striatellus to the linear blue light source of 460 nm was the highest and was also significantly affected by the light intensity. The results of the experiments with the combined color light sources showed that compared with the linear 460 nm blue light source, the trapping rate of the L. striatellus was significantly improved by the polychromatic light, and the blue-green light led to the best improvement, reaching 1.5 times that of the trapping rate in the case of monochromatic light sources. The wavelength composition, light intensity, shape, and flash frequency of the light source used in this study can provide a theoretical basis for the development of LED-based light traps specifically for L. striatellus.