Version 3.0 of ORFEO Toolbox
Release Notes: Cloud detection was implemented for QB/Pléiades images. Alignment and right angle detection (Burns, Grompone) were addded. Radiometric indices were added for vegetation, water, and soil. Texture computations were optimized, including Haralick, SFS, Pantex, and Edge density. A SIFT density image function and filter were added. Object-based segmentation and filtering were added. A LSD line segment detector and right angle detector were added. PCA computation was added.
Other releases
Release Notes: Large JPEG2000 file (Pleiades-like) support and Pleiades metadata handling. Efficient JPEG2000 visualization and ROI decompression tools in Monteverdi. A revamp of otb applications in a generic and scalable framework: launch applications from the command line, from an auto-generated QT GUI, from Python, from within QGis, etc. There are many new algorithms: Dimensionality Reduction (ICA, PCA, MNF, MAF, etc.), change detection (MAD), Hyperspectral Unmixing, elevation map from stereo data, compare segmentation with a ground truth (Hoover), etc. There are various bugfixes.
Release Notes: Lots of new features were added, including a multi-image supervised classification suite in OTB-Applications, an application to perform segmentation based on connected components and object based image analysis, a framework for cartographic database creation, and validation based on the fusion of features in the framework of Dempster-Shafer evidence theory.
Release Notes: A new object detection framework was added. Local descriptors tools based on ImageFunction were improved. In SVMClassifier, hyperplanes distances as output were added. GDALImageIO now has support for writing non-streamable formats (such as JPEG or PNG). Reading vector images of std::complex is supported. In BandMathFilter, physical and image coordinates variables were added. A class to generate a kmz file from an image was added.
Release Notes: A filter was added to perform arbitrary mathematical operations on image bands. A class to parse mathematical expressions and an interpolation function were added. Filters were added to compute extrema of big images. A filter was added to perform fine registration between images. Filters for higher order texture computation were added.
Release Notes: This release produces hill shade images from DEM, compute point set density functions (Gaussian, Epanechnikov), extract a subset from a point set, applys a transform to a point set, generates a random point set, computes confusion matrices, generates a list of samples from vector data, computes ground spacing, reads 6S spectral sensitivity files, and includes improvements in the computation of Haralick Texture coefficients.
