Unmanned aerial vehicles (UAVs), known as drones, have played a significant role in recent years in creating resilient smart cities. UAVs can be used for a wide range of applications, including emergency response, civil protection, search and rescue, and surveillance, thanks to their high mobility and reasonable price. Automatic recognition of human activity in aerial videos captured by drones is critical for various tasks for these applications. However, this is difficult due to many factors specific to aerial views, including camera motion, vibration, low resolution, background clutter, lighting conditions, and variations in view. Although deep learning approaches have demonstrated their effectiveness in a variety of challenging vision tasks, they require either a large number of labelled aerial videos for training or a dataset with balanced classes, both of which can be difficult to obtain. To address these challenges, a hybrid data augmentation method is proposed which combines data transformation with the Wasserstein Generative Adversarial Network (GAN)-based feature augmentation method. In particular, we apply the basic transformation methods to increase the amount of video in the database. A Convolutional Neural Network–Long Short-Term Memory (CNN-LSTM) model is used to learn the spatio-temporal dynamics of actions, then a GAN-based technique is applied to generate synthetic CNN-LSTM features conditioned on action classes which provide a high discriminative spatio-temporal features. We tested our model on the YouTube aerial database, demonstrating encouraging results that surpass those of previous state-of-the-art works, including an accuracy rate of 97.83%.

<p>The agriculture sector in Tunisia plays a vital role in the Tunisian economy with 6% of the country's exports earning, 12.6% of GDP and almost one<br>quarter of the country's labor force. However, Tunisian agriculture is still increasingly exposed to a variety of vulnerabilities and uncertainties including in particular the climate variability such as drought and floods. In fact, Rainfall quantity and its geographic distribution are the main drivers of water productivity and agriculture production and a predominant key factor in the overall agriculture hazard risk management processes. This paper uses the daily open rainfall data from the national observatory of Tunisian agriculture (ONAGRI) to develop an ETL (Extract,Transform and load) tool to automatically spatialize and load the historical data into a big data platform by continuously incrementing the new daily disseminated records. In addition, this paper applies the Voronoi spatial analysis model to estimate rainfall measures for the newly added spatial units using VGI data from OSM world mapping project. Then, based on these spatial estimations, the paper examines the feasibility of applying ARIMA (Auto Regressive Integrated Moving Average) for time series forecasting by comparing it with deep learning methods ANN (Arti cial Neural Network) and LSTM (Long Short Term Memory) in order to predict the rainfall values corresponding to particular agriculture area belonging to a Tunisian region. Our experimental results showed that prediction accuracy increased with LSTM model comparing to the other models for the rainfall time series forecasting embedded now with geographic location.</p>

No abstract