Search

ObjectiveCamellia oleifera is one of the four major woody oil plants in the world. Diseases is a significant factor leading to the decline in quality of Camellia oleifera and the financial loss of farmers. Among these diseases, anthracnose is a common and severe disease in Camellia oleifera forests, directly impacting yields and production rates. Accurate disease assessment can improve the prevention and control efficiency and safeguarding the farmers' profit. In this study, an improved You Only Look at CoefficienTs (YOLACT) based method was proposed to realize automatic and efficient grading of the severity of Camellia oleifera leaf anthracnose.MethodsHigh-resolution images of Camellia oleifera anthracnose leaves were collected using a smartphone at the National Camellia oleifera Seed Base of Jiangxi Academy of Forestry, and finally 975 valid images were retained after a rigorous screening process. Five data enhancement means were applied, and a data set of 5 850 images was constructed finally, which was divided into training, validation, and test sets in a ratio of 7:2:1. For model selection, the Camellia-YOLACT model was proposed based on the YOLACT instance segmentation model, and by introducing improvements such as Swin-Transformer, weighted bi-directional feature pyramid network, and HardSwish activation function. The Swin Transformer was utilized for feature extraction in the backbone network part of YOLACT, leveraging the global receptive field and shift window properties of the self-attention mechanism in the Transformer architecture to enhance feature extraction capabilities. Additionally, a weighted bidirectional feature pyramid network was introduced to fuse feature information from different scales to improve the detection ability of the model for objects at different scales, thereby improving the detection accuracy. Furthermore, to increase the the model's robustness against the noise in the input data, the HardSwish activation function with stronger nonlinear capability was adopted to replace the ReLu activation function of the original model. Since images in natural environments usually have complex background and foreground information, the robustness of HardSwish helped the model better handling these situations and further improving the detection accuracy. With the above improvements, the Camellia-YOLACT model was constructed and experimentally validated by testing the Camellia oleifera anthracnose leaf image dataset.Results and DiscussionsA transfer learning approach was used for experimental validation on the Camellia oleifera anthracnose severity grading dataset, and the results of the ablation experiments showed that the mAP75 of Camellia-YOLACT proposed in this study was 86.8%, mAPall was 78.3%, mAR was 91.6% which were 5.7%, 2.5% and 7.9% higher than YOLACT model. In the comparison experiments, Camellia-YOLACT performed better than Segmenting Objects by Locations (SOLO) in terms of both accuracy and speed, and its detection speed was doubled compared to Mask R-CNN algorithm. Therefore, the Camellia-YOLACT algorithm was suitable in Camellia oleifera gardens for anthracnose real-time segmentation. In order to verify the outdoors detection performance of Camellia-YOLACT model, 36 groups of Camellia oleifera anthracnose grading experiments were conducted. Experimental results showed that the grading correctness of Camellia oleifera anthracnose injection severity reached 94.4%, and the average absolute error of K-value was 1.09%. Therefore, the Camellia-YOLACT model proposed in this study has a better performance on the grading of the severity of Camellia oleifera anthracnose.ConclusionsThe Camellia-YOLACT model proposed got high accuracy in leaf and anthracnose segmentation of Camellia oleifera, on the basis of which it can realize automatic grading of the severity of Camellia oleifera anthracnose. This research could provide technical support for the precise control of Camellia oleifera diseases.

Considering the serious risks posed by the heavy metals present in pig diets, monitoring and controlling the amount of these metals is crucial. The laser-induced breakdown spectroscopy (LIBS) is a promising technique for performing elemental analysis due to its unique advantages, such as rapid, in situ, nondestructive, and online detection of various minerals. In order to improve the detection accuracy and detection limit (LOD) of elements, the spatial constraint combined with laser-induced breakdown spectroscopy (CC-LIBS) is used to detect the quantity of zinc in pig feed and achieve the detection of samples with lower concentrations. Zn I 480.86 nm is selected as the characteristic spectral line, and the effects of different two-dimensional (cylindrical) and three-dimensional (hemispherical, truncated cone) cavities on the enhancement factor are compared under different time delays. The results show that the optimal conditions include a truncated cone cavity D5H3, a delay time of 2.08, and an enhancement factor of 3.01 associated with the analytical spectral line. The detection limit of zinc in samples under CC-LIBS is 62.67 mg/kg (the recommended quantity of zinc in pig diet mentioned in the safe use of feed additives is 43–80 mg/kg), which is 35.65% lower than that under the LIBS (97.39 mg/kg). Therefore, this work provides a new test basis and idea for the detection of zinc in pig feed.

With the development of machine vision and spectral detection technology, online sorting of fruit internal and external quality has been developed rapidly. However, for spherical fruits, it is difficult to obtain full surface images during sorting, so it is difficult to accurately calculate the size of the surface defects and the ratio of defects to the full surface. In this paper, a full surface line scanning image acquisition device for spherical fruit is proposed. Based on this device, the line scanning hyperspectral image of spherical fruit is collected, and the original image is extracted by feature extraction and background removal. Next, the isometric projection image and the equivalent projection image of the feature image is obtained through cartography projection transformation; The number of feature pixels in the original feature image, the isometric projection image, the equivalent projection image, and the width of the original feature image are used as input parameters to predict the actual defect area with the help of the shallow neural network. In this paper, the equipment and method are verified using three test balls with different diameters and pasting different sizes of identification blocks at different positions on their surfaces. The experimental results show that the prediction accuracy R of the test set of the model is 0.9937, and the RMSE is 0.3391 cm2. It can be seen that the method has good prediction accuracy, which can provide a reference for the hyperspectral on‐line sorting method of external quality of spherical fruit. Practical application: This method provides an effective solution for the quality sorting production line of spherical fruits. In addition to agricultural product quality testing and food quality testing, similar to the detection of industrial products such as ball balls, the scheme provided in this manuscript can also be used as one of the options. The method proposed in this manuscript is suitable for all kinds of line scanning equipment, including hyperspectral imager and laser profilometer. [ABSTRACT FROM AUTHOR]

Imazalil, a preservative for navel orange in the process of postharvest processing, is easy to seep into the flesh through the peel and produce residues in the flesh, which is vulnerable to cause endanger to human body if it was eaten accidentally. Base on this, a fast detection method of imazalil residues in navel orange peel ,namely surface-enhanced Raman spectroscopy (SERS) was proposed in this study. Firstly, the SERS detection conditions of imazalil residues in navel orange peel were optimized, and the optimal detection conditions were determined as follows: Reaction time of 2 min, gold colloid of 400 µL, NaBr as electrolyte solution, NaBr dosage of 25 µL. Based on the above optimal conditions, 6 groups of spectral data processed by adaptive iterative penalized least squares (air PLS), air PLS combination with normalization, air PLS combination with baseline correction, air PLS combination with first derivative, air PLS combination with standard normal distribution (SNV), air PLS combination with multiplicative scatter correction (MSC) were used to establish support vector regression (SVR) models and compare the models prediction performance. And air PLS method was selected as the spectral pretreatment method, because the value of correlation coefficient computed value of prediction set (RP) is the largest, and the value of root mean square error calculated value of the prediction set (RMSEP) is the smallest. Then, principal component analysis (PCA) was used to extract the features from spectral data, and the first seven principal component scores were selected as the input values of SVR prediction model. SVR, multiple linear regression (MLR) and partial least squares regression (PLSR) were used to analyze and compare the prediction performances. The RP value of prediction set of SVR prediction model could reach 0.9156, the RMSEP value of their prediction set was 4.8407 mg/kg, and the relative standard deviation computation value (RPD) was 2.3103, which indicated that the closer the predicted value of imazalil residue on navel orange surface based on SVR algorithm was to the measured value, the more effective the prediction accuracy of the model could be. The above data indicated that the speedy detection of imazalil residues in navel orange peel could be emploied by SERS coupled with PCA and SVR modeling method.

Because antibiotics are regularly used for chicken, food safety is of utmost importance, and health experts pay attention to the effect antibiotics could have on human health. This study examines [...]

The final particle moisture content of ceramic slurry spray drying affects the processing, product quality, and process energy consumption. While the real-time moisture content determination is very difficult for spray drying process due to nonlinear, severe lag, interference factors, and complexity of the systems, the high temperature, high humidity drying environment and short drying process. In current work, an improved moth optimization (IMFO) algorithm combined with backpropagation neural network (BPNN) was proposed to predict the moisture content of ceramic slurry particles during spray drying. The performance of the model is trained and tested. Results demonstrate that IMFO has better convergence ability and speed compared to other algorithms such as particle swarm optimization (PSO) and gravitational search algorithm (GSA). The IMFO-BPNN model achieves a MAE of 0.0293, RMSE of 0.0383, and R2 value of 0.9113, outperforming the prediction performance of BPNN and MFO-BPNN models. The absolute error rate of the IMFO-BPNN model (0–5%) is lower compared to BPNN (5–10%) and MFO-BPNN model (>10%), showcasing superior accuracy in predicting moisture content. This mathematical model established in the study provides an efficient, accurate, and nondestructive method for predicting the drying endpoint of ceramic slurry spray drying. [ABSTRACT FROM AUTHOR]

The shell explosion by hot air drying is a critical step in the processing of Camellia oleifera fruit (COF), which directly affects the degree of the shell explosion, and the separation effect of Camellia oleifera seed and Camellia oleifera shell after the shell explosion of COF. To reveal the characteristics of the COF shell explosion, a hot air drying device was designed based on mass conservation and drying principles. The physical characteristics of COF and the evolution of drying parameters were thoroughly analyzed with a combination method of drying analysis and experimental. Moreover, under the conditions of air temperature 50–70 °C, relative humidity 20–50%, and air velocity 1.3–1.9 m/s, the internal relationship between COF shell explosion formation through hot air drying and the hot air drying medium was systematically investigated by response surface methodology, and a prediction model for the shell explosion rate of COF by hot air drying was constructed using statistical methods. Results demonstrated that decreasing the relative humidity and increasing the temperature and air velocity of the drying medium could reduce the dehydration time of COF. The moisture content of Camellia oleifera shell was found to be 177.45% d.b. (dry basis) at the initial cracking stage of COF. Furthermore, at temperatures ranging from 50 to 70 °C Deff values of COF were estimated to be within the range of 0.915 × 10−9 to 1.782 × 10−9 m2/s. Similarly, at relative humidity levels of 20 to 50%, Deff values ranged from 1.226 × 10−9 to 1.501 × 10−9 m2/s. At an air velocity of 1.3 to 1.9 m/s, Deff values ranged from 0.956 × 10−9 to 1.501 × 10−9 m2/s. The measured values of the shell explosion rate were in close agreement with that calculated using the fitted model, with a correlation coefficient of 0.997 and a root mean square error of 0.9743. This study will provide a theoretical basis for optimizing the shell explosion process and improving shell explosion rate of COF by hot air drying. [ABSTRACT FROM AUTHOR]

Federated learning, as a distributed machine learning framework, aims to protect data privacy while addressing the issue of data silos by collaboratively training models across multiple clients. However, a significant challenge to federated learning arises from the non-independent and identically distributed (non-iid) nature of data across different clients. non-iid data can lead to inconsistencies between the minimal loss experienced by individual clients and the global loss observed after the central server aggregates the local models, affecting the model's convergence speed and generalization capability. To address this challenge, we propose a novel federated learning algorithm based on update bias (FedUB). Unlike traditional federated learning approaches such as FedAvg and FedProx, which independently update model parameters on each client before direct aggregation to form a global model, the FedUB algorithm incorporates an update bias in the loss function of local models—specifically, the difference between each round's local model updates and the global model updates. This design aims to reduce discrepancies between local and global updates, thus aligning the parameters of locally updated models more closely with those of the globally aggregated model, thereby mitigating the fundamental conflict between local and global optima. Additionally, during the aggregation phase at the server side, we introduce a metric called the bias metric, which assesses the similarity between each client's local model and the global model. This metric adaptively sets the weight of each client during aggregation after each training round to achieve a better global model. Extensive experiments conducted on multiple datasets have confirmed the effectiveness of the FedUB algorithm. The results indicate that FedUB generally outperforms methods such as FedDC, FedDyn, and Scaffold, especially in scenarios involving partial client participation and non-iid data distributions. It demonstrates superior performance and faster convergence in tasks such as image classification. [ABSTRACT FROM AUTHOR]

In order to solve the problem of image boundary segmentation caused by the irregularity of paddy fields in southern China, a high-precision segmentation method based on the improved MultiResUNet model for paddy field mapping is proposed, combining the characteristics of paddy field scenes. We introduce the attention gate (AG) mechanism at the end of the encoder–decoder skip connections in the MultiResUNet model to generate the weights and highlight the response of the field ridge area, add an atrous spatial pyramid pooling (ASPP) module after the end of the encoder down-sampling, use an appropriate combination of expansion rates to improve the identification of small-scale edge details, use 1 × 1 convolution to improve the range of the sensory field after bilinear interpolation to increase the segmentation accuracy, and, thus, construct the AM-UNet paddy field ridge segmentation model. The experimental results show that the IoU, precision, and F1 value of the AM-UNet model are 88.74%, 93.45%, and 93.95%, respectively, and that inference time for a single image is 168ms, enabling accurate and real-time segmentation of field ridges in a complex paddy field environment. Thus, the AM-UNet model can provide technical support for the development of vision-based automatic navigation systems for agricultural machines. [ABSTRACT FROM AUTHOR]

Direct-seeded rice (Oryza sativa) (DSR) is gaining increasing acceptance worldwide because it saves both time and labor. A covering treatment is a simple method to improve early plant growth under direct-seeding conditions. Herein, field experiments were conducted to study the effects of different powdered organic covering materials, i.e., organic fertilizer, seedling substrate, nutrient soil A, and nutrient soil B. In addition, the effects of different covering thicknesses were studied, including 5, 10, 15, and 20 mm, on the germination and early growth of DSR varieties. The two rice varieties Wufengyou 286 and Zhongjiazao 17 were used for this study. The seedling substrate at 15 mm and nutrient soil A at 5 mm covering thickness significantly increased the rate of germination of Wufengyou 286 by 43.93% and 40.26%, respectively, compared to the control treatment. Organic fertilizer, seedling substrate, and nutrient soil A at covering thicknesses of 5 mm significantly increased the germination rate by 42.57%, 38.62%, and 40.10%, respectively, compared to the control. Notably, all the treatments increased the emergence rate, with Wufengyou 286 exhibiting the most pronounced effect. The seedling substrate at 10 mm covering thickness significantly increased the rate of emergence of Wufengyou286 by 15.42% compared to the control. Similarly, the seedling substrate at 10 mm covering thickness gave the best results by increasing the rate of emergence of Zhongjiazao 17 by 10.85% compared to the control. On average, the rate of emergence of Wufengyou 286 and Zhongjiazao 17 under the experimental treatment increased by 11.81% and 6.45%, respectively, compared to the control. The appropriate cover treatment enhanced the early growth of DSR by improving its morphological attributes, such as plant height and dry weight, and regulating the physio-biochemical responses, such as the production of antioxidants and photosynthetic pigments. The seedling growth rate of both varieties decreased as the thickness increased. This study elucidated the effects of different covering materials on the early growth of DSR, which can further guide their actual production. [ABSTRACT FROM AUTHOR]