查看更多>>摘要：? 2022 The AuthorsOne way to estimate ammonia emission rates from naturally ventilated housing systems is to scale-up mechanistic modeling results. However, obtaining the relevant data to set initial and boundary conditions adequately is usually very challenging and for a whole barn barely possible. This study has investigated the potential of coupling different mechanistic modeling approaches towards an overarching barn scale ammonia emission model, which might permit ammonia emission projections for naturally ventilated housing systems with minimal measurement efforts. To this end, we combined an ammonia volatilization model for shallow urine or slurry puddles with a dynamic mechanistic model of digestion and excretion of nitrogen, an empirical model to estimate urination volumes, semi-empirical models for pH and temperature dynamics of the puddles and a mechanistic air flow model. The ammonia volatilization model was integrated with a time step of one second over a period of twenty-four hours, while the relevant boundary conditions were updated on an hourly base (determined by the other mentioned submodels). Projections and uncertainties of the approach were investigated for a farm case with about ten months of on-farm measurements in a naturally ventilated dairy cattle building with scraped solid floor in Northern Germany. The results showed that the nested model was in general capable to reproduce the long-term emission trend and variability, while the short-term variability was damped compared with the emission measurements. A sensitivity study indicated that particularly a refinement of the submodules for urine puddle alkalizing, urination volume and urea concentration distributions as well as for local near-surface wind speeds have a great potential to further improve the overall model accuracy. The cleaning efficiency of the scraper has turned out to be a crucial and sensitive parameter in the modeling, which so far has been described insufficiently by measurements or modeling approaches.
查看更多>>摘要：? 2022 Elsevier B.V.Aquaculture plays a critical role in food security and nutrition strategies. The application of intelligent aquaculture technology has shown promising performance in improving aquaculture productivity and increasing economic benefits with its rapid advancement and good prospects. However, degraded underwater images have hampered the existing computer vision applications in intelligent aquaculture. To this end, a novel Tied Bilateral learning network is proposed for Aquaculture Image Enhancement (TBAIE), which improves the degraded aquaculture images to meet the requirements of various computer vision applications in aquaculture. Concretely, a novel multiple tied guidance module is designed to generate a multi-channel feature map and capture long-range features based on input. Then, a feature fusion module is introduced with a novel tied attention block to blend the feature and suppress noise with a low computational resource. Experimental results demonstrate that the proposed TBAIE can improve the quality of aquaculture images and remove color distortion. Moreover, TBAIE can achieve state-of-the-art in quantitative and qualitative metrics and meet the practical requirements of different aquaculture vision tasks.
查看更多>>摘要：? 2022 The Author(s)Plant-based measurements are recognized as key methods to obtain insightful data in the field. In general, they are labor-intensive and expensive. In this context, Non-Contact Resonant Ultrasonic Spectroscopy technique (NC-RUS) emerged as a powerful alternative that enabled plant water status determination in a non-destructive, non-invasive and rapid way. However, NC-RUS is not applicable to all plant species as it depends on the possibility to excite and sense thickness resonances in the leaves. In this work, we propose and test an ultrasonic technique that can be used in all leaves, regardless of the appearance of thickness resonances. This technique is based on the contactless measurement of through transmitted airborne ultrasonic pulses in the leaves at high-frequencies and in the absence of thickness resonances, to obtain the leaf ultrasonic velocity (vair). It benefits from the facts that: i) at sufficiently high frequencies (typically around 1 MHz) all leaves are non-resonant (so the technique can be applied to both resonant and non-resonant leaves), ii) the use of high-frequencies allows a greater time resolution and a further miniaturization, making possible to apply the technique to small and irregular leaves. Three different signal processing techniques were used to determine the time it takes to the ultrasonic pulse to cross the leaves (time-of-flight) from the measured signals. Two of them operate in time domain: cross-correlation, and edge detection, while the third one makes use of the Fast Fourier Transform (FFT) and operates in the frequency domain: phase-slope. If leaf thickness is also measured, ultrasound velocity can then be worked out. As ultrasound velocity is determined by density and elastic modulus, it is then closely related to water content and turgor pressure. Obtained ultrasound velocities were first validated by comparing them with those obtained by well-established and standard ultrasonic methods: water immersion transmission (vwater) and NC-RUS (vres). The conclusions of this comparison permitted us to propose a novel methodology that combines the three signal processing techniques used to improve robustness and accuracy for the measurement of ultrasound velocity in plant leaves. It is of interest to note that a bias towards higher values of vair compared to vres was observed. This behavior is considered the consequence of the different influence of the leaf layered structure in these two measurements, so this feature can be further used for leaf structure analysis.
查看更多>>摘要：? 2022 Elsevier B.V.Understanding the propagation process of acoustic waves in the soil is very important for designing high-performance acoustic wave detection devices. In this study, the discrete element method was used to simulate the propagation process of acoustic waves in the soil. A single factor experiment was carried out with soil compression ratio, excitation frequency, and excitation amplitude as factors. The influence of various factors on the time and frequency domain were analyzed from the perspective of the received signal and the force signal on the particle. The results showed that: with the increase of soil compression ratio, the speed of the acoustic wave, the amplitude of the received signal's first wave and the dominant frequency increased; with the increase of excitation frequency, the amplitude of the received signal's first wave decreased, and the dominant frequency and speed of acoustic wave remained unchanged; with the increase of excitation amplitude, the received signal amplitude increased, and the dominant frequency and speed of acoustic wave remained unchanged. According to the force signal of the particles on the sound wave propagation path, as the soil compression ratio increased, the dominant frequency of the acoustic wave increased, and the amplitude attenuation coefficient increased; as the excitation frequency increased, the dominant frequency and amplitude attenuation coefficient of the acoustic wave first increased and then decreased; as the excitation amplitude increased, the dominant frequency of the acoustic wave remained unchanged, and the amplitude attenuation coefficient increased. The occurrences of these phenomena were related to the natural frequency of the soil and the sound wave attenuation mechanism. During the propagation process, the dominant frequency of the sound wave would continue to attenuate and eventually reached the natural frequency, resulting in the same dominant frequency of the received signal under different excitation frequencies. Under different parameters, the decay speed of the sound wave amplitude was different, and the initial signal amplitude was different, which led to the various first wave amplitude of the received signal. The conclusion could be drawn through the simulation experiment: the DEM method could well characterize the acoustic wave propagation process in the soil and analyze the acoustic wave propagation law in the soil. Thus, it provides theoretical support and reference for the research and instrument manufacture of the acoustic wave detection of soil characteristics.
查看更多>>摘要：? 2022 Elsevier B.V.Recent deep learning methods have allowed important steps forward in the automatic detection of wheat ears in the field. Nevertheless, it was still lacking a method able to both count and segment the ears, validated at all the development stages from heading to maturity. Moreover, the critical step of converting the ear count in an image to an ear density, i.e. a number of ears per square metre in the field, has been widely ignored by most of the previous studies. For this research, wheat RGB images have been acquired from heading to maturity in two field trials displaying contrasted fertilisation scenarios. An unsupervised learning approach on the YOLOv5 model, as well as the cutting-edge DeepMAC segmentation method were exploited to develop a wheat ear counting and segmentation pipeline that necessitated only a limited amount of labelling work for the training. An additional label set including all the development stages was built for validation. The average F1 score of ear bounding box detection was 0.93 and the average F1 score of segmentation was 0.86. To convert the ear counts to ear densities, a second RGB camera was used so that the distance between the cameras and the ears could be measured by stereovision. That distance was exploited to compute the image footprint at ear level, and thus divide the number of ears by this footprint to get the ear density. The obtained ear densities were coherent regarding the fertilisation scenarios but, for a same fertilisation, differences were observed between acquisition dates. This highlights that the measurement was not able to retrieve absolute ear densities for all the development stages and conditions. The deep learning measurement considered the most reliable outperformed observations from three human operators.
de Castro Pereira R.da Costa R.M.Hirose E.Ferreira de Carvalho O.L....
12页
查看更多>>摘要：? 2022 Elsevier B.V.This paper presents a novel strategy to detect and classify adult whiteflies and five important related stages on images of detached soybean leaves. The whitefly Bemisia tabaci is a major pest in soybean crops, and by detecting, counting and differentiating its related life stages in field collected leaves control management decisions can be made. The proposed solution is based on a deep learning object detection algorithm (YOLOv4), modified into an specific new learning strategy with innovations on data augmentation, image mosaicking, and fusion of hypothesized object categories. A real and annotated dataset of images is provided from a controlled experiment infected with whitefly eggs having 121 images and 973 annotated objects. The experimental results showed a promising performance of the proposed system, reaching an f1-score of 0.87, in comparison with a single YOLOv4 algorithm that reached f1-score of 0.80. The overall strategy could be extended to work in other similar tasks for image based pest management.
查看更多>>摘要：? 2022 Elsevier B.V.The internet of things has revolutionized Chinese agricultural production and marketing sectors, creating “New Farmers.” WeChat is one of the most popular social media platforms in China, and WeChat marketing of agricultural products has become a widespread phenomenon. To respond to the quality and safety concerns of consumers on the social media marketing platform, New Farmers should pay more attention and maintain the quality and safety of the agricultural products they sell on WeChat. Our objective is to understand the determinants influencing the New Farmers' quality and safety knowledge of the fresh fruits they sell on WeChat. For this purpose, we collected data in 2018 from a sample of 504 purely sales-oriented New Farmers in China. We analyzed the influence of WeChat marketing, food safety awareness, personal characteristics, and external factors on purely sales-oriented New Farmers' quality and safety knowledge related to the fresh fruits they sell on WeChat based on the Stimulus-Organism-Response framework. Results indicated that purely sales oriented New Farmers' ranking of their understanding of the quality and safety status (pesticides, fertilizers, and quality and safety certifications) of the fruits depend on factors such as consumers' consultation frequency about quality and safety information to New Farmers, consumers' word of mouth, sales status of fruits, WeChat regulation, quality and safety awareness of New Farmers, gender, training received, government regulations, and consumers attention to food safety. The findings should be helpful to strengthen the quality and safety management, training, and supervision of New Farmers.
查看更多>>摘要：? 2022 The Author(s)Research is being extensively conducted on using deep learning in the field of crop and weed segmentation based on images captured with a camera. However, the segmentation performance for various crops and weeds varies significantly, implying that certain classes of crops or weeds are not being detected properly. This problem may also occur in the loss calculations used in crop and weed segmentation. In previous studies, the cross-entropy loss (corresponding to a distribution loss) and dice loss (using spatial information) have been widely used. However, such losses lead to large discrepancies in crop and weed segmentation performance, as the correlations between crop and weed classes are not considered. In order to solve these problems, this study proposes multi-task semantic segmentation-convolutional neural network for detecting crops and weeds (MTS-CNN) using one-stage training. This approach adds the crop, weed, and both (crop and weed) losses to heighten the correlations between the crop and weed classes, and designs the model so that the object (crop and weed) region is trained intensively. In experiments conducted using three types of open databases - the BoniRob dataset, a crop/weed field image dataset (CWFID), and rice seedling and weed dataset - the mean intersection of union (MIOU) values of the segmentation for the crops and weeds in the MTS-CNN are 0.9164, 0.8372, and 0.8260, respectively. Thus, the results indicate higher accuracy from the proposed approach than from the state-of-the-art methods.
查看更多>>摘要：? 2022 Elsevier B.V.Drought prediction of regional crops during the growth stages can get drought information in advance and prepare for the response, so as to effectively guide the water-saving irrigation and lessen yield losses of spring maize. This study used the daily meteorological data in the research area during 1965–2019 to calculate the Crop Water Deficit Index (CWDI) and used the Pearson correlation coefficient (PCC) method to select the relevant factors for the CWDI. Then, CWDI was predicted using Support Vector Machine (SVM), Back Propagation Neural Network (BPNN), Wavelet Neural Network (WNN), Support Vector Machine-Particle Swarm Optimize (SVM-PSO), Back Propagation Neural Network-Particle Swarm Optimize (BPNN-PSO), and Back Wavelet Neural Network-Particle Swarm Optimize (WNN-PSO) models. By comparing d, R2, Root Mean Square Error (RMSE), and Mean Relative Error (MRE), the best model was selected and used to predict drought in the next five years. The compared results showed that WNN-PSO models performed better than all the other models. When the input variables were CWDI and 3 main relevant meteorological factors relative humidity, maximum temperature and precipitation at sowing-seedling stage, the MRE (1.58%～6.65%), the MAE (1.2922 ～ 3.5866) as well as RMSE (0.0174 ～ 0.0481) were the smallest and the d(0.8608 ～ 0.915) and R2 (0.8402 ～ 0.9853) was the largest. The model R2 increased by 10.6%, 32.8% and 125.9% compared with WNN, BPNN-PSO and SVM-PSO. It is proved that WNN-PSO is suitable for predicting CWDI of spring maize in drought-affected areas.
查看更多>>摘要：? 2022 Elsevier B.V.Grape downy mildew is a major biotic constraint to grapevine production worldwide, and its impact is influenced by environmental conditions and varietal susceptibility. Here, we proposed a grape downy mildew model to optimize plant disease management and its application on northern Chinese grapevine areas. A primary and a secondary infection model of P. viticola were integrated to estimate the date of symptoms on set and disease incidence, to give the firSt brEakout and incidEnce of grape Downy Mildew model (SEE_DM). The experimental data for model calibration were collected on two grapevine varietals with high and moderate susceptibility to downy mildew grown in a multiyear (2009, 2011–2019) and multisite (Beijing, Shenyang and Yantai) trial. A model sensitivity analysis (Sobol analysis) drove the selection of the subset of relevant parameters to be adjusted in calibration. The model adequately reproduced the observed downy mildew incidence, obtaining high R2 (0.89), Nash-Sutcliffe modelling efficiency (0.72), and low RMSE (9–16%). The model correctly estimated the disease onset date in all conditions but two (late by 5 and 13 days), demonstrating to be a valid tool to take timely decision to limit the time course of the disease. The model was then projected in three sites under varying temperature and rainfall to analyses their effects on the trends of grape downy mildew incidence, using a variety with moderate susceptibility. This study highlights that Northern China is gradually becoming more suitable for grape downy mildew infections, along with warmer temperatures and more frequent rainfall.
NSTL
Elsevier