Supervised classification systems used for land cover mapping require accurate reference databases. These reference data come generally from different sources such as field measurements, thematic maps, or aerial photographs. Due to misregistration, update delay, or land cover complexity, they may contain class label noise, i.e., a wrong label assignment. This study aims at evaluating the impact of mislabeled training data on classification performances for land cover mapping. Particularly, it addresses the random and systematic label noise problem for the classification of high resolution satellite image time series. Experiments are carried out on synthetic and real datasets with two traditional classifiers: Support Vector Machines (SVM) and Random Forests (RF). A synthetic dataset has been designed for this study, simulating vegetation profiles over one year. The real dataset is composed of Landsat-8 and SPOT-4 images acquired during one year in the south of France. The results show that both classifiers are little influenced for low random noise levels up to 25%–30%, but their performances drop down for higher noise levels. Different classification configurations are tested by increasing the number of classes, using different input feature vectors, and changing the number of training instances. Algorithm complexities are also analyzed. The RF classifier achieves high robustness to random and systematic label noise for all the tested configurations; whereas the SVM classifier is more sensitive to the kernel choice and to the input feature vectors. Finally, this work reveals that the cross-validation procedure is impacted by the presence of class label noise.

Contrary to aerial images, satellite images are often affected by the presence of clouds. Identifying and removing these clouds is one of the primary steps to perform when processing satellite images, as they may alter subsequent procedures such as atmospheric corrections, DSM production or land cover classification. The main goal of this paper is to present the cloud detection approach, developed at the French Mapping agency. Our approach is based on the availability of multi-temporal satellite images (i.e. time series that generally contain between 5 and 10 images) and is based on a region-growing procedure. Seeds (corresponding to clouds) are firstly extracted through a pixel-to-pixel comparison between the images contained in time series (the presence of a cloud is here assumed to be related to a high variation of reflectance between two images). Clouds are then delineated finely using a dedicated region-growing algorithm. The method, originally designed for panchromatic SPOT5-HRS images, is tested in this paper using time series with 9 multi-temporal satellite images. Our preliminary experiments show the good performances of our method. In a near future, the method will be applied to Pléiades images, acquired during the in-flight commissioning phase of the satellite (launched at the end of 2011). In that context, this is a particular goal of this paper to show to which extent and in which way our method can be adapted to this kind of imagery.

Detecting clouds and their shadows is one of the primaries steps to perform when processing satellite images because they may alter the quality of some products such as large-area orthomosaics. The main goal of this paper is to present the automatic method developed at IGN-France for detecting clouds and shadows in a sequence of satellite images. In our work, surface reflectance orthoimages are used. They were processed from initial satellite images using a dedicated software. The cloud detection step consists of a region-growing algorithm. Seeds are firstly extracted. For that purpose and for each input ortho-image to process, we select the other ortho-images of the sequence that intersect it. The pixels of the input ortho-image are secondly labelled seeds if the difference of reflectance (in the blue channel) with overlapping ortho-images is bigger than a given threshold. Clouds are eventually delineated using a region-growing method based on a radiometric and homogeneity criterion. Regarding the shadow detection, our method is based on the idea that a shadow pixel is darker when comparing to the other images of the time series. The detection is basically composed of three steps. Firstly, we compute a synthetic ortho-image covering the whole study area. Its pixels have a value corresponding to the median value of all input reflectance ortho-images intersecting at that pixel location. Secondly, for each input ortho-image, a pixel is labelled shadows if the difference of reflectance (in the NIR channel) with the synthetic ortho-image is below a given threshold. Eventually, an optional region-growing step may be used to refine the results. Note that pixels labelled clouds during the cloud detection are not used for computing the median value in the first step; additionally, the NIR input data channel is used to perform the shadow detection, because it appeared to better discriminate shadow pixels. The method was tested on times series of Landsat 8 and Pléiades-HR images and our first experiments show the feasibility to automate the detection of shadows and clouds in satellite image sequences.