Various generative and discriminative methods have been transferred from the computer vision field to remote sensing applications using different low and high semantic level descriptors. However, as classical approaches have shown their limits in representation learning and are not intended to deal with the great variability of the data. With the emergence of large-scale annotated datasets in vision, the convolutional deep approaches represent the most winning solutions by supporting this variability with spatial context integration through different semantic abstraction levels. In the lack of annotated remote sensing data, in this paper, we are comparing the performances of deep features produced by six different CNNs that have been trained on well established computer vision datasets with respect to the detection of small objects (cars) in very high resolution Pleiades imagery. Our findings show good generalization performance and are very encouraging for future applications.

Extracting and identifying objects in very high resolution imagery has been a popular research topic in remote sensing. Since the beginning of this decade, deep learning techniques have revolutionized computer vision providing significant performance gains compared to traditional “shallow” techniques in various challenging vision problems. The training of deep neural networks usually requires very large training datasets. The advantage of using deep features is to exploit already trained Convolutional Neural Networks (CNN) in order to produce high level features without the burden of having to train a CNN from scratch. In this paper, we are investigating the use of deep features for the detection of small objects (cars and individual trees) in high resolution Pleiades imagery. Preliminary results show good detection performance and are very encouraging for future applications.

The last three decades have seen significant mining development in the northern regions of Canada, where the freeze and thaw cycle of permafrost and corresponding surface subsidence and heave represent a significant challenge at all mining stages, from the design of infrastructures to the monitoring of restored areas. Over the past ten years, SAR interferometry has been widely used to monitor ground surface deformation. With this technique, changes in phase between two SAR acquisitions are used to detect centimetre to millimetre surface displacements over a large area with high spatial resolution. This paper presents the results of a project that aims to develop a SAR solution to provide useful information for environmental monitoring and assessing the stability of mining sites. RADARSAT-2 and TerraSAR-X images acquired during the summer of 2014 were used to measure the displacements of ground surface, infrastructures and stockpiles caused by seasonal changes in permafrost extent. The study area is an open-pit mine located in Nunavut, northern Canada, in the continuous permafrost zone. Results shown that surface displacements calculated from RADARSAT-2 and TerraSAR-X are very similar and in agreement with scientific and terrain knowledge. Significant displacements were observed in loose soil areas while none was detected in bedrock and rock outcrop areas. The areas most affected by active layer changes showed surface subsidence during the thaw settlement period. Thus, InSAR can be used as a tool to guide the siting and design of new infrastructure as well as highlighting risks in areas of unstable terrain.

This paper presents three applications of SAR interferometry. In the first study, DInSAR technique applied to TerraSAR-X images acquired in Nunavik, northern Canada, showed that surface subsidence observed in the permafrost thaw period is more important in loose soil areas while no surface movement were detected in rock outcrop areas. In the second one, electricity transmission tower movement was simulated through controlled vertical displacements of corner reflectors. These displacements were clearly observed in six TerraSAR-X interferometric pairs, while no movement was detected in the surrounding stable surfaces. The third application relates to monitoring of a dam movement, before and after water impoundment, using SAR interferometry.

Spatial information related to vegetation status and soil properties is needed in precision farming, especially early in the growing season. At these stages, vegetation has already emerged while soil is also visible in multispectral EO images. In this paper, linear spectral unmixing is applied to an 8 bands WorldView-2 image to extract information on both vegetation and soil acquired at in-season nitrogen sidedress stage in 2010 and 2011 for four corn fields located in the Montérégie region of Quebec, Canada. EO derived soil properties were strongly correlated to ECa. Correlation between dark soil abundance and ECa reached R=0.9 and correlation between bright soil abundance and ECa was about R=-0.7. Vegetation abundance for combined data of several fields was better correlated to measured biomass than NDVI and SAVI. The possibility to get valuable soil and plant information from a single multispectral image offers an interesting cost reduction opportunity for precision farming applications.