Search

• A defencing algorithm based on deep learning was proposed. • The algorithm based on U-Net and Pix2pixHD. • Chicken contours occlude by the wire mesh can be recovered. • The accuracy of caged chicken detection is improved after defencing. Cage farming is the mainstream farming mode in China. Accurate individual identification and behavioral detection of caged chickens can provide managers with a better understanding of chicken status. However, for image detection of caged chickens, the cage may affect the accuracy of the detection algorithm. For this reason, CCD (caged chicken defencing), a defencing algorithm based on U-Net and pix2pixHD, was proposed to improve caged chickens' detection accuracy. The proposed defencing algorithm can accurately identify the cage wire mesh and recover the chicken contours completely. In the test set, the detection accuracy of the cage wire mesh was 94.71%, while a structural similarity (SSIM) of 90.04% and a peak signal-to-noise ratio (PSNR) of 25.24 dB were obtained in the image recovery. To verify the practicality of the method proposed in this paper, we analyzed the performance of the object detection algorithm before and after defencing from the perspective of the most basic individual detection in the caged chicken detection task. We validated the defencing algorithm with different YOLOv5 detection algorithms, including YOLOv5s, YOLOv5m, YOLOv5l, and YOLOv5x. The experimental results showed that the defencing algorithm improved the detection precision of caged chickens by 16.1%, 12.1%, 7.3%, and 5.4%, respectively, compared with before defencing. The recall improvement was 29.1%, 16.4%, 8.5%, and 6.8%. To our knowledge, this is the first time that a deep learning-based defencing algorithm has been applied to caged chickens, and the detection accuracy can be significantly improved. The method proposed in this paper can remove cage wire mesh greatly and provide a technical reference for subsequent poultry researchers. [ABSTRACT FROM AUTHOR]

• A multi-view field-road classification method was proposed for GNSS trajectories. • Trajectories were converted to trajectory images to encode rich motion information. • Pixel-level visual features were extracted by a specific image segmentation model. The understanding of agricultural machinery, whether it is used "in field" or "on road", plays an essential role in optimizing the efficiency of cross-regional agricultural mechanization services that have been available in China for decades. With the widespread availability of Global Navigation Satellite System (GNSS)-enabled devices, developing a method for automatically identifying the activity associated with each point in GNSS recorded trajectories can significantly enhance the optimization of these services. In this paper, we propose a novel field-road classification method that utilizes two feature vectors to represent each point in a GNSS trajectory. The first feature vector is a statistical feature vector extracted with existing methods, while the second feature vector is a visual feature vector obtained through an image segmentation model. To extract visual features, we first convert each GNSS trajectory into a trajectory image that encodes the motion information (e.g., speed, direction) of each point and the spatiotemporal relationship between points. Then, an image segmentation model specifically for field-road classification was developed, which can effectively extract a visual feature vector for each pixel in a trajectory image, corresponding to each point in the GNSS trajectory. The image segmentation model used in our approach has a robust feature extraction capability that enhances the statistical feature representation of each point by adding pixel-level visual features. We evaluated the effectiveness of our multi-view field-road classification method using trajectories of corn, wheat, and paddy harvesters. The results demonstrated that our method achieved accuracy scores of 94.17%, 90.93%, and 83.43% for respective datasets. Our method outperformed state-of-the-art field-road classification methods, especially for trajectories with high-frequency GNSS acquisition. [ABSTRACT FROM AUTHOR]

• A detection approach for late-autumn shoots of litchi tree based on UAV remote sensing is provided. • A remote sensing dataset of late-autumn shoots of litchi tree is constructed. • A deep learning model based on the combination of YOLOv5 and Swin-transformer is proposed. • The detection accuracy of late-autumn shoots in remote sensing images is improved. • It will provide technical support for other researchers on late-autumn shoots and orchards manager. Litchi is one of the most common economic fruits in southern China, however, the growth of late-autumn shoots of litchi hinders flower bud differentiation and reduces yield of fruit. The early identification of the late-autumn shoots is of great significance for orchard management to control shoots and then increase fruit yield. At present, the identification of late-autumn shoots still relies on manual methods, which is not suitable for smart orchard management in a large area due to low recognition efficiency and high subjectivity. Therefore, a convenient, fast and cost-effective method is urgently needed. In response to this problem, the paper proposes a method based on the combination of unmanned aerial vehicle (UAV) remote sensing and object detection algorithm to detect late-autumn shoots. For this purpose, a remote sensing dataset of late-autumn shoots of litchi is first constructed by UAV. An improved YOLOv5 algorithm called YOLOv5-SBiC is then developed for late-autumn shoots identification. In the YOLOv5-SBiC algorithm, the transformer module is introduced to speed up the convergence of the network and improve detection accuracy, the attention mechanism module is employed to help the model extracting details, and BiFPN is used to better solve the multi-scale problem in detecting and then improve the recognition effect of small-sized objects. In addition, CIOU is selected as the loss function of bounding boxes regression to achieve high-precision localization of the boxes. The test results demonstrate that the recognition accuracy of YOLOv5-SBiC reaches a relatively high value of 79.6%, which is 4.0% higher than that (75.6%) of the original YOLOv5 algorithm and 15.9% higher than that (63.7%) of the pure transformer algorithm. It's also demonstrated that YOLOv5-SBiC is more competitive than the mainstream target detection algorithms in the dataset of late-autumn shoots. [ABSTRACT FROM AUTHOR]

• Peach variety detection using VIS-NIR spectroscopy. • Detection and classification without handcrafted feature extraction. • Best detection accuracy was 100% in the validation and 94.4% in the test. More than 1000 peach varieties with significant differences in qualities are cultivated in China. Distinguishing peach varieties is not only needed by peach dealers, but also demanded by products processing enterprises and consumers. Detecting the varieties of fruit is one of the most important analyses. Over the past decades, there are many linear or nonlinear methods have been proposed, such as principal component analysis, partial least squares, support vector machine etc. However, these traditional methods commonly need some preprocessing steps including denoising, baseline correction, wavelength selection and so on. Hence, it requires the users to have enough skilled knowledge before they can establish a good performance detection model. To offer information on identifying peach varieties, visual near-infrared (VIS-NIR) diffuse reflectance spectra between 350 and 820 nm were collected for five peach varieties, 100 samples for each variety, the total samples consisted of 500 peaches. Spectral information has shown the rapid non-destructive measurement ability in many research fields. In recent years, the spectral information was researched for detecting the sugar and firmness of fruits, for examples, apple, peach, tomato and so on. Some traditional detection methods using spectra require application specific transforms, expertly designed constraints and model parameters, and have limited detection performance due to their maintenance costs. With the development of machine learning technology in recent years, deep learning which plays an important role in different research projects has won the eyes of fields from both academy and industry. In order to classify peach varieties by analyzing VIS-NIR spectra, a detection method based on deep learning principle is proposed in this paper. In this paper, the method can realize multi-identification peach varieties, by constructing a one dimensional convolution neural network and establishing a VIS-NIR spectra database containing five kinds of peaches. The network model is obtained through training, and then is employed to predict the testing set data. The accuracy of models based on deep learning reached 100% in the validation dataset and 94.4% in the test dataset. This study indicated that peach varieties could be distinguished successfully by using VIS-NIR spectroscopy and deep learning. [ABSTRACT FROM AUTHOR]

• The residual structure is introduced into the Inception structure as a convolutional block, and a plant disease classification network is established. • An improved Convolutional Block Attention Module combing channel attention and spatial attention is embedded in the classification network. • Corn, potato and tomato are identified as a whole and individually by using a classification network. • Designed and developed a web-based disease detection system, which can complete real-time identification of plant diseases on mobile phones and computers. In this paper, we proposed a convolutional neural network based on Inception and residual structure with an embedded modified convolutional block attention module (CBAM), aiming to improve the classification of plant leaf diseases. Corn, potatoes and tomatoes are the most cultivated grains in southern China. The leaves of the three crops are very fragile and sensitive and are susceptible to leaf diseases, such as leaf blight of corn, late blight of potato and mosaic virus of tomato. These diseases cannot be identified at early stages. Therefore, an efficient solution is proposed by deep learning techniques to detect the disease categories of crops, which can effectively prevent the spread of diseases and ensure the normal growth of plants. In this experiment, our model achieved an overall accuracy of 99.55% for the identification of the three diseases of corn, potato and tomato. In addition, we tested the three plants individually. The classification accuracy of our model on corn, potato and tomato was 98.44%, 99.43% and 95.20%, respectively. We have also developed a web-based real-time plant disease classification system and deployed our model. The system had good performance in time and accuracy evaluation metrics. The results of this study showed that our model had fewer parameters, shorter training time, and higher recognition accuracy compared to existing image classification models. [ABSTRACT FROM AUTHOR]

• We have established a multi-target data set of agricultural pests, Pest24, for precision agriculture. • We analyzed the data properties of Pest24 finding that three factors mainly affect pest detection. • Experiments suggest the practical values of several STOA detection methods for real-time monitoring pests. • Pest24 can be used as gold standard for DL-based detection algorithm evaluation in the community. Precision agriculture poses new challenges for real-time monitoring pest population in field based on new-generation AI technology. In order to provide a big data resource for training pest detection deep learning models, this paper establishes a large-scale multi-target standardized data set of agricultural pests, named Pest24. Specifically, the data set currently consists of 25,378 field pest annotated images collected from our automatic pest trap & imaging device. Totally, 24 categories of typical pests are involved in Pest24, which dominantly destroy field crops in China every year. We apply several state-of-the-art deep learning detection methods, Faster RCNN, SSD, YOLOv3, Cascade R-CNN to detect the pests in the data set, and obtain encouraging results for real-time monitoring field crop pests. To explore the factors that affect the detection accuracy of pests, we analyze the data set in a variety of aspects, finding that three factors, i.e. relative scale, number of instances and object adhesion, mainly influence the pest detection performance. Overall, Pest24 is featured typically with large scale multi-pest image data, very small object scales, high object similarity and dense pest distribution. We hope that Pest24 promotes accurate multi-pest monitoring for precision agriculture and also benefits the machine vision community by providing a new specialized object detection benchmark. [ABSTRACT FROM AUTHOR]

• An end-to-end model is proposed for rice yield prediction using deep learning fusion. • Deep spatial-temporal features significantly reduced prediction errors. • The model converges quickly to low errors and provides stable prediction performance. • The model achieves the optimal prediction performance when the number of layers is 6. Rice yield is essential for more than half of the world's population, and thus, accurate predictions of rice yield are of great importance for trade, development policies, humanitarian assistance, decision-makers, etc. However, traditional mechanistic models and statistical machine learning models need to identify features, making the research on and application of these models laborious and time-consuming. In this paper, a novel end-to-end prediction model that fuses two back-propagation neural networks (BPNNs) with an independently recurrent neural network (IndRNN), named BBI-model, is proposed to address these challenges. In stage one, BBI-model preprocesses the original area and meteorology data. In stage two, one BPNN and the IndRNN are used to learn deep spatial and temporal features in parallel. In stage three, another BPNN combines two kinds of deep features and learns the relationships between these deep features and rice yields to make predictions for summer and winter rice yields. The experimental results indicate that BBI-model achieved the lowest mean absolute error (MAE) and root mean square error (RMSE) for the summer rice prediction (0.0044 and 0.0057, respectively) and corresponding values of 0.0074 and 0.0192 for the winter rice prediction when the number of layers in the network was set to six. Moreover, the errors of the model using the combination of deep spatial-temporal features were significantly lower than when simply using deep temporal features. Furthermore, the model converged quickly with 100 iterations and then remained stable. These findings confirm that the model can make accurate predictions for summer and winter rice yields of 81 counties in the Guangxi Zhuang Autonomous Region, China. [ABSTRACT FROM AUTHOR]