Fusing attention mechanism with Mask R-CNN for instance segmentation of grape cluster in the field.

Accurately detecting and segmenting grape cluster in the field is fundamental for precision viticulture. In this paper, a new backbone network, ResNet50-FPN-ED, was proposed to improve Mask R-CNN instance segmentation so that the detection and segmentation performance can be improved under complex environments, cluster shape variations, leaf shading, trunk occlusion, and grapes overlapping. An Efficient Channel Attention (ECA) mechanism was first introduced in the backbone network to correct the extracted features for better grape cluster detection. To obtain detailed feature map information, Dense Upsampling Convolution (DUC) was used in feature pyramid fusion to improve model segmentation accuracy. Moreover, model generalization performance was also improved by training the model on two different datasets. The developed algorithm was validated on a large dataset with 682 annotated images, where the experimental results indicate that the model achieves an Average Precision (AP) of 60.1% on object detection and 59.5% on instance segmentation. Particularly, on object detection task, the AP improved by 1.4% and 1.8% over the original Mask R-CNN (ResNet50-FPN) and Faster R-CNN (ResNet50-FPN). For the instance segmentation, the AP improved by 1.6% and 2.2% over the original Mask R-CNN and SOLOv2. When tested on different datasets, the improved model had high detection and segmentation accuracy and inter-varietal generalization performance in complex growth environments, which is able to provide technical support for intelligent vineyard management.

UAV spraying on citrus crop: impact of tank-mix adjuvant on the contact angle and droplet distribution.

Adding tank-mix adjuvants into the spray mixture is a common practice to improve droplet distribution for field crops (e.g., rice, wheat, corn, etc.) when using Unmanned Aerial Vehicle (UAV) sprayers. However, the effectiveness of tank-mix adjuvant for UAV spraying in orchard crops is still an open problem, considering their special canopy structure and leaf features. This study aims to evaluate the effects of a typical tank-mix adjuvant concentrations (i.e., Nong Jian Fei (NJF)) on Contact Angle (CA) and droplet distribution in the citrus tree canopy. Three commonly used parameters, namely dynamic CA, droplet coverage, and Volume Median Diameter (VMD), are adopted for performance evaluation. The dynamic CAs on the adaxial surface of citrus leaves, for water-only and NJF-presence sprays, respectively, are measured with five concentration levels, where three replications are performed for each concentration. The sprays with 0.5‰ NJF are adopted in the field experiment for evaluating droplet distributions, where Water Sensitive Papers (WSPs) are used as collectors. Two multi-rotor UAVs (DJI T20 and T30) which consist of different sizes of pesticide tanks and rotor diameters are used as the spraying platforms. Both water-only and NJF-presence treatments are conducted for the two UAVs, respectively. The results of the CA experiment show that NJF addition can significantly reduce the CAs of the sprays. The sprays with 0.5‰ NJF obtain the lowest CA within the observing time, suggesting a better spread ability on solid surface (e.g., WSPs or/and leaves). With respect to the effects of NJF addition on individual UAVs, the field trial results indicate that NJF addition can remarkably increase both the droplet coverage and VMD at three canopy layers, except for T30 droplet coverage of the inside and bottom layers. Comparing the difference of droplet coverage between two UAVs, while significant difference is found in the same layer before NJF addition, there is no notable difference appearing in the outside and bottom layers after NJF addition. The difference of VMD in the same layer between two UAVs is not affected by NJF addition except for the bottom layer. These results imply that the differences of droplet coverage among different UAVs might be mitigated, thus the droplet distribution of some UAVs could be improved by adding a tank-mix adjuvant into the sprays. This hypothesis is verified by investigating the droplet penetration and the correlation coefficient (CC) of droplet coverage and VMD. After NJF addition, the total percentage of T20 droplet coverage in the bottom and inside layers is increased by 5%. For both UAVs, the CCs indicate that both droplet coverage and VMD increase at the same time in most cases after NJF addition. In conclusion, the addition of a tank-mix adjuvant with the ability to reduce CA of the sprays, can effectively improve droplet distribution using UAV spraying in the citrus canopy by increasing droplet coverage and VMD.

A multi-sensory stimulating attention model for cities' taxi service demand prediction.

Taxi demand forecasting is crucial to building an efficient transportation system in a smart city. Accurate taxi demand forecasting could help the taxi management platform to allocate taxi resources in advance, alleviate traffic congestion, and reduce passenger waiting time. Thus, more efforts in industrial and academic circles have been directed towards the cities' taxi service demand prediction (CTSDP). However, the complex nonlinear spatio-temporal relationship in demand data makes it challenging to construct an accurate forecasting model. There remain challenges in perceiving the micro spatial characteristics and the macro periodicity characteristics from cities' taxi service demand data. What's more, the existing methods are significantly insufficient for exploring the potential multi-time patterns from these demand data. To meet the above challenges, and also stimulated by the human perception mechanism, we propose a Multi-Sensory Stimulus Attention (MSSA) model for CTSDP. Specifically, the MSSA model integrates a detail perception attention and a stimulus variety attention for capturing the micro and macro characteristics from massive historical demand data, respectively. The multiple time resolution modules are employed to capture multiple potential spatio-temporal periodic features from massive historical demand data. Extensive experiments on the yellow taxi trip records data in Manhattan show that the MSSA model outperforms the state-of-the-art baselines.

The short-term associations of chronic obstructive pulmonary disease hospitalizations with meteorological factors and air pollutants in Southwest China: a time-series study.

Chronic obstructive pulmonary disease (COPD) is the fourth major cause of mortality and morbidity worldwide and is projected to be the third by 2030. However, there is little evidence available on the associations of COPD hospitalizations with meteorological factors and air pollutants in developing countries/regions of Asia. In particular, no study has been done in western areas of China considering the nonlinear and lagged effects simultaneously. This study aims to evaluate the nonlinear and lagged associations of COPD hospitalizations with meteorological factors and air pollutants using time-series analysis. The modified associations by sex and age were also investigated. The distributed lag nonlinear model was used to establish the association of daily COPD hospitalizations of all 441 public hospitals in Chengdu, China from Jan/2015-Dec/2017 with the ambient meteorological factors and air pollutants. Model parameters were optimized based on quasi Akaike Information Criterion and model diagnostics was conducted by inspecting the deviance residuals. Subgroup analysis by sex and age was also performed. Temperature, relative humidity, wind and Carbon Monoxide (CO) have statistically significant and consistent associations with COPD hospitalizations. The cumulative relative risk (RR) was lowest at a temperature of 19℃ (relative humidity of 67%). Both extremely high and low temperature (and relative humidity) increase the cumulative RR. An increase of wind speed above 4 mph (an increase of CO above 1.44 mg/m3) significantly decreases (increases) the cumulative RR. Female populations were more sensitive to low temperature and high CO level; elderly (74+) populations are more sensitive to high relative humidity; younger populations (< = 74) are more susceptible to CO higher than 1.44 mg/m3. Therefore, people with COPD should avoid exposure to adverse environmental conditions of extreme temperatures and relative humidity, low wind speed and high CO level, especially for female and elderly patients who were more sensitive to extreme temperatures and relative humidity.

Wheat Stripe Rust Grading by Deep Learning With Attention Mechanism and Images From Mobile Devices.

Wheat stripe rust is one of the main wheat diseases worldwide, which has significantly adverse effects on wheat yield and quality, posing serious threats on food security. Disease severity grading plays a paramount role in stripe rust disease management including breeding disease-resistant wheat varieties. Manual inspection is time-consuming, labor-intensive and prone to human errors, therefore, there is a clearly urgent need to develop more effective and efficient disease grading strategy by using automated approaches. However, the differences between wheat leaves of different levels of stripe rust infection are usually tiny and subtle, and, as a result, ordinary deep learning networks fail to achieve satisfying performance. By formulating this challenge as a fine-grained image classification problem, this study proposes a novel deep learning network C-DenseNet which embeds Convolutional Block Attention Module (CBAM) in the densely connected convolutional network (DenseNet). The performance of C-DenseNet and its variants is demonstrated via a newly collected wheat stripe rust grading dataset (WSRgrading dataset) at Northwest A&F University, Shaanxi Province, China, which contains a total of 5,242 wheat leaf images with 6 levels of stripe rust infection. The dataset was collected by using various mobile devices in the natural field condition. Comparative experiments show that C-DenseNet with a test accuracy of 97.99% outperforms the classical DenseNet (92.53%) and ResNet (73.43%). GradCAM++ network visualization also shows that C-DenseNet is able to pay more attention to the key areas in making the decision. It is concluded that C-DenseNet with an attention mechanism is suitable for wheat stripe rust disease grading in field conditions.

Dimension Reduction Aided Hyperspectral Image Classification with a Small-sized Training Dataset: Experimental Comparisons.

Hyperspectral images (HSI) provide rich information which may not be captured by other sensing technologies and therefore gradually find a wide range of applications. However, they also generate a large amount of irrelevant or redundant data for a specific task. This causes a number of issues including significantly increased computation time, complexity and scale of prediction models mapping the data to semantics (e.g., classification), and the need of a large amount of labelled data for training. Particularly, it is generally difficult and expensive for experts to acquire sufficient training samples in many applications. This paper addresses these issues by exploring a number of classical dimension reduction algorithms in machine learning communities for HSI classification. To reduce the size of training dataset, feature selection (e.g., mutual information, minimal redundancy maximal relevance) and feature extraction (e.g., Principal Component Analysis (PCA), Kernel PCA) are adopted to augment a baseline classification method, Support Vector Machine (SVM). The proposed algorithms are evaluated using a real HSI dataset. It is shown that PCA yields the most promising performance in reducing the number of features or spectral bands. It is observed that while significantly reducing the computational complexity, the proposed method can achieve better classification results over the classic SVM on a small training dataset, which makes it suitable for real-time applications or when only limited training data are available. Furthermore, it can also achieve performances similar to the classic SVM on large datasets but with much less computing time.