• Rural Digital Europe
• Other research productsclose
• CemOA close

Estimating exact 3D distribution of canopy components using terrestrial lidar in forest is limited by signal occlusion. We propose a method to address this limitation: it uses voxels, beam returns and beam propagation through the scene. The proposed method was validated using simulated forest scenes and a lidar simulator.

This work presents a novel approach for mapping the spatial distribution of natural habitats in the 'Foothills of Larzac' Natura 2000 listed site located in a French Mediterranean Biogeographical Region. Sparse Partial Least Square Discriminant Analysis was used to analyze two RapidEye datasets (June 2009 and July 2010) with the purpose of choosing the most informative spectral, textural and thematic variables that allow discriminating the classes of habitats. The Sparse Partial Least Square Discriminant Analysis selected relevant and stable variables for the discrimination of habitat classes that could be linked to ecological or biophysical characteristics. It also gave insight into the similarities and the differences between habitats classes with comparable physiognomic characteristics. The highest user accuracy was obtained for dry improved grasslands (u=91.97%) followed by riparian ash woods (u= 88.38%). These results are very encouraging given that these two classes were identified in Annex 1 of the EC Habitats Directive as of community interest. Due to limited data input requirements and to its computational efficiency, the approach developed in this paper is a good alternative to other types of variable selection approaches in a supervised classification framework and can be easily transferred to other Natura 2000 sites.

In recent years, with the rise of wireless sensors, wireless sensor networks (WSN) are used in many applications such as in smart agriculture. Thus, the aim of this internship is to realize a wireless sensor network to monitor crops development from remote places. In this network, the monitored data include temperature, humidity and soil moisture. / Un capteur sans fil est élaboré autour d'un microcontrôleur auquel est associé un module de communication sans fil et différentes sondes (ou capteurs) (thermomètre, etc.). Son utilisation la plus courante consiste donc à acquérir les données provenant de son environnement de déploiement grâce à sa (ou ses) sonde(s) et de les transmettre ensuite par communication sans fil vers un système central (ou station) de collecte. Cette transmission des données peut être directe ou transiter par d'autres capteurs sans fil qui forment ainsi un «Réseau de Capteurs Sans Fil» (RCSF). De par ses caractéristiques, cette technologie des RCSF, bien que relativement récente, offre de nouvelles perspectives dans de multiples domaines d'application dont ceux de l'agriculture et de l'environnement. Le projet de recherche dans lequel s'intègre ce sujet de stage vise à l'utilisation de cette technologie pour venir alimenter un outil d'aide à la décision (OAD) notamment dédié au suivi de cultures agricoles. L'objectif global de ce stage est donc d'aider à la mise en place d'un tel RCSF ce qui implique, entre autres, le choix de l'organisation (topologie et mode de fonctionnement) du réseau avec le paramétrage des différents capteurs sans fil qui le constituent.

Remote sensing has been used to map river bathymetry for several decades. Non-contact methods are necessary in several cases: inaccessible rivers, large-scale depth mapping, very shallow rivers. The remote sensing techniques used for river bathymetry are reviewed. Frequently, these techniques have been developed for marine environment and have then been transposed to riverine environments. These techniques can be divided into two types: active remote sensing, such as ground penetrating radar and bathymetric lidar; or passive remote sensing, such as through-water photogrammetry and radiometric models. This last technique which consists of finding a logarithmic relationship between river depth and image values appears to be the most used. Fewer references exist for the other techniques, but lidar is an emerging technique. For each depth measurement method, we detail the physical principles and then a review of the results obtained in the field. This review shows a lack of data for very shallow rivers, where a very high spatial resolution is needed. Moreover, the cost related to aerial image acquisition is often huge. Hence we propose an application of two techniques, radiometric models and through-water photogrammetry, with very high-resolution passive optical imagery, light platforms, and off-the-shelf cameras. We show that, in the case of the radiometric models, measurement is possible with a spatial filtering of about 1 m and a homogeneous river bottom. In contrast, with through-water photogrammetry, fine ground resolution and bottom textures are necessary.

Remote sensing images allow the monitoring of field techniques among crops. CIRAD and SPOT imaging set up between 2003 and 2005 a powerful tool to help the management of sugar cane crops within the framework of GIS and optical images (SUCRETTE project). CIRAD sets up MARGOUILLA in 2009, its own project in La Réunion. The purpose of this training period is to study the potentiality of the RADAR images to characterize the cultural stage of sugar cane crops (harvested mapping), in order to increment and to fulfill information already acquired by optical images. That said, we focused on the influence on three different parameters affecting the RADAR signal: incident angle, polarization, thanks to RADAR images acquired by the sensor TERRASAR-X (X Band). First, we studied backscatter coefficient values (σ) of sugar cane reference field at different dates and secondly, analyze the temporal evolution of σ. We observed on several points the RADAR signal sensibility at different growing stage of sugar cane, validating a characterization potential. Nonetheless, this potential remains limited for soil and sugar cane under moisture (causing by rain) which could be confused with other growing stage. / Les images satellitaires permettent d'étudier les pratiques agricoles de cultures diverses. Le CIRAD et SPOT Image ont mis en place entre 2003 et 2005 un outil d'aide à la gestion de la culture de la canne à sucre avec un SIG et les images optiques (projet SUCRETTE). Le CIRAD installe en 2009 à La Réunion son propre projet, baptisé MARGOUILLA. L'objectif de ce stage est de prolonger ces travaux et d'évaluer le potentiel des images RADAR à caractériser les états culturaux des parcelles de cannes à sucre (cartographie des coupes), afin d'enrichir et de compléter les informations déjà obtenues avec les images optiques. Pour cela, nous avons analysé plus précisément l'influence sur le signal RADAR des paramètres suivants : angle d'incidence et polarisation grâce à l'analyse d'images RADAR acquises par le capteur hyperfréquence TERRASAR-X (bande X). Nous nous sommes intéressés aux valeurs de coefficient de rétrodiffusion (σ°) des parcelles de canne à sucre de référence à différentes dates et nous avons suivi dans un second temps l'évolution dans le temps de σ°. Nous avons observé sous plusieurs approches la sensibilité du signal RADAR aux différents stades de développement de la canne à sucre. Les résultats confirment bien le fort potentiel de caractérisation de ces stades.

The routing protocol for low power and lossy network (RPL) was designed in the ROLL working group at IETF since the year of 2008. Until the latest version of draft 19 released, this protocol algorithms and its four application scenario, such as home automation, industrial control, urban environment and building automation, have been nearly grounded. However, it is still very difficult to find effective approaches to simulate and evaluate RPL's behavior and other extensions of its application. In this paper, first we provide a brief presentation of the RPL protocol including two case studies ContikiRPL and TinyRPL, and an initial simulation experiment results obtained from the RPL capable COOJA simulator and its developed module. Second we then focus on the utilization of this protocol in the precision agriculture area and propose our dedicated instances hybrid network architecture to meet the specific requirement of this application. As a conclusion, we summarized our ongoing work and future solutions of the current technology issues.

Airborne lidar systems (ALS) provide 3D point clouds of the topography by direct time measurement of a short laser pulse after reflection on the Earth surface. For the last decade, this technique has proved to be the ideal remote sensing tool for delivering very accurate digital terrain model (DTM) of the Earth surface, and then for answering main environmental issues such as natural hazard prevention and natural ressource management. Moreover, such active systems, also called multiple echo lidar, allow to detect several return signals for a single laser shot. It is particularly relevant in case of vegetation areas since a single lidar survey allows to acquire not only the canopy top (the only visible layer from passive sensors), but also points inside the vegetation layer and on the ground underneath. Thus, among the different remote sensing techniques, airborne laser scanning has also proved to be the most efficient technique to characterize both forest structure and ground topography. For a few years, new airborne laser scanning systems called full-waveform lidar systems have emerged, providing not only 3D point clouds as classical ALS systems, but entire altimeter profiles of reflected energy from the Earth surface. These profiles represent the laser backscattered energy as a function of time. They give to the end-user more control and flexibility on the signal processing steps and enable to extract more information than classical multi-echo lidar data. A detailed state-of-the-art of such systems can be found in [1]. However, managing these data with spacial and time dependency is much more complex than images or 3D point clouds : raw full-waveform lidar data are sets of range profiles of various lengths that are stored in the sensor geometry following both the scan angle of the lidar system and the chronological order along the flight track. Moreover, the data volume is drastically larger than 3D point clouds: it takes about 140 GB for an acquisition time of 1.6 h with a pulse repetition frequency (PRF) of 50kHz. Furthermore, there is neither commercial nor opensource toolkit to handle full-waveform lidar data, but some constructor solutions, that are black boxes, can only extract 3D point clouds from raw data and are designed to their own sensors. Finally, there is not standard file format for full-waveform data (such as the LAS format for multi-echo data). Managing full-waveform lidar data is therefore a challenging task, and we adress this issue by developping a specific research tool: FullAnalyze.

/ Ce mémoire porte sur le développement d'algorithmes de perception par vision artificielle pour le guidage d'un bras robotisé évoluant en milieu agricole, et plus précisément dans des champs de betteraves sucrières. Pour un bras robotisé, la détection de son environnement et des objets qu'il doit analyser (dans ce cas, les feuilles de betterave) est importante afin de prendre de bonnes décisions au niveau de sa commande. De nombreuses contraintes sont à prendre en compte lors de la conception de ces algorithmes, parmis lesquelles la variation de luminosité en milieu extérieur, les différences de forme des feuilles et plants de betterave, ou encore la mobilité du bras robotique. Les méthodes de vision artificielle présentées dans ce mémoire tiennent compte de ces contraintes.

Accurate estimation of leaf chlorophyll content (Cab) from remote sensing is of tremendous significance to monitor the physiological status of vegetation or to estimate primary production. Many vegetation indices (VIs) have been developed to retrieve Cab at the canopy level from meter- to decameter-scale reflectance observations. However, most of these VIs may be affected by the possible confounding influence of canopy structure. The objective of this study is to develop methods for Cab estimation using millimeter to centimeter spatial resolution reflectance imagery acquired at the field level. Hyperspectral images were acquired over sugar beet canopies from a ground-based platform in the 400-1000 nm range, concurrently to Cab, green fraction (GF), green area index (GAI) ground measurements. The original image spatial resolution was successively degraded from 1 mm to 35 cm, resulting in eleven sets of hyperspectral images. Vegetation and soil pixels were discriminated, and for each spatial resolution, measured Cab values were related to various VIs computed over four sets of reflectance spectra extracted from the images (soil and vegetation pixels, only vegetation pixels, 50% darkest and brightest vegetation pixels). The selected VIs included some classical VIs from the literature as well as optimal combinations of spectral bands, including simple ratio (SR), modified normalized difference (mND) and structure insensitive pigment index (SIPI). In the case of mND and SIPI, the use of a blue reference band instead of the classical near-infrared one was also investigated. For the eleven spatial resolutions, the four pixel selections and the five VI formats, similar band combinations are obtained when optimizing VI performances: the main bands of interest are generally located in the blue, red, red edge and near-infrared domains. Overall,mNDblue[728,850] defined as (R440

The use of consumer digital cameras or webcams to characterize and monitor different features has become prevalent in various domains, especially in environmental applications. Despite some promising results, such digital camera systems generally suffer from signal aberrations due to the on-board image processing systems and thus offer limited quantitative data acquisition capability. The objective of this study was to test a series of radiometric corrections having the potential to reduce radiometric distortions linked to camera optics and environmental conditions, and to quantify the effects of these corrections on our ability to monitor crop variables. In 2007, we conducted a five-month experiment on sugarcane trial plots using original RGB and modified RGB (Red-Edge and NIR) cameras fitted onto a light aircraft. The camera settings were kept unchanged throughout the acquisition period and the images were recorded in JPEG and RAW formats. These images were corrected to eliminate the vignetting effect, and normalized between acquisition dates. Our results suggest that 1) the use of unprocessed image data did not improve the results of image analyses; 2) vignetting had a significant effect, especially for the modified camera, and 3) normalized vegetation indices calculated with vignetting-corrected images were sufficient to correct for scene illumination conditions. These results are discussed in the light of the experimental protocol and recommendations are made for the use of these versatile systems for quantitative remote sensing of terrestrial surfaces.