CloudCompare

*** CloudCompare versions history *** v2.7.0 - 04/22/2016 - Enhancements: * Cross section tool - Now based on OpenGL for a much faster display (+ proper integration in the new LOD mechanism) * Rasterize tool improved - Now handles RGB colors - Raster export dialog updated - The 'Cancel' button of the grid update progress bar is now handled properly - Now exports images with the active (default) color scale * Support for the Oculus Rift device - new option of the 'Stereo' mode (you have to use the 'Stereo' version of CloudCompare) - development is still in 'alpha' state - units must be expressed in meters - works best with a 3D mouse - works best in 'bubble view' mode (i.e. with FARO or PTX files) * The 'Global Shift & Scale' mechanism now allows for shifting values above 1.0e9 (for georeferenced clouds expressed in millimeters!) * New 'File > Global Shift settings' dialog - can be used to set the limits that trigger the 'Global Shift & Scale' mechanism - settings are persistent - settings are used for both the standard application and the command line mode - default triggering value (for the coordinates) have been lowered to 10^4 * New 'Tools > Sand-box > Distance Map' tool * allows to compute distance maps for clouds (= distances from regularly sampled positions around the points) * the output points can be filtered inside a specified distance range * The 'Edit > Multiply / Scale', 'Tools > Registration > Match bounding-box centers' and 'Tools > Registration > Match scales' methods now update the 'transformation history matrix' of the entities * Bubble View mode - now allows for horizontal AND vertical rotation - the 3D mouse is now properly handled * Command line mode - I/O plugins are now loaded in command line mode (i.e. the FARO, DP and PCD formats can now be used) - new option 'MAX_TCOUNT' to specify the maximum number of threads to be used for computing distances (C2C or C2M) as well as for ICP registration - the SPLIT_XYZ and MAX_DIST options can now be used together for C2C distance * Default light position - the default (sun) light position is now in the middle of the screen so as to get a brighter visualization of meshes * Distance computation - C2C distance can now be split along the 3 dimensions (X, Y, Z) even when using local models or a maximum distance (in which case some values may be NaN) * ICP registration - the user can now set the maximum number of threads/core to use * LOD mechanism enhanced (faster and smarter) * The 'Link camera' checkbox has been moved to the 'Display menu' * Mac OS X adds support for OpenGL Frame Buffer Objects (FBOs) * Mac OS X adds support for the CSV Matrix I/O file format * Mac OS X adds support for the following plugins: - qFacet - qAnimation - Bug fix: * The HSV to RGB method was broken * The 'Convert normals to HSV colors' mehod doesn't rely on the Dip / Dip direction anymore as the way these values are computed have been changed recently (with a symmetry about the plane Z = 0) * When playing with the 'skip lines' parameter of the ASCII file loading dialog, the roles assignments could be cleared (when a line was reappearing while it had less elements than the other lines) * Mac OS X properly saves and restores the main window's state & geometry * Mac OS X uses the standard menu names and keyboard shortcuts for entering and exiting full screen v2.6.3.1 - 03/XX/2016 - Enhancements: * Camera sensors can now be created 'freely' (i.e. not necessarily attached to a given entity) and moved freely in the DB tree * The cross section box orientation can now be setup via the 'advanced' button - Bug fixes: * The meshes over the user-specified limit (display options) were ALWAYS decimated (and not only when the mouse is moved) * 2D area label picking was broken (first corner was misplaced) * 3D point picking on a cloud with a temporary GL transformation (e.g. in the Align tool) was not working properly * The camera sensor dialog was presenting the camera orientation vectors in the wrong order (horizontal instead of vertical) * OpenMP support disabled for Ransac Shape Detection on Windows (the process loops infinitely if enabled) * When custom labels were defined for a custom color scale, the values that were not present in the active SF range were ignored v2.6.3 (= pre 2.7.0 for systems that don't support Qt5) - 03/13/2016 - New features: * Polyline tracing/picking tool: - accessible via a dedicated icon in the main tool bar ('Trace a polyline by point picking') - allows to pick points on all the visible clouds in the active 3D view to define one or more 3D polylines - allows to generate a tight contour polyline by automatically picking several points between each vertex ('oversampling') * 2.5D volume calculation tool: - Tools > Volume > Compute 2.5D volume - can compute the volume of a single cloud (relatively to a constant height) or between two clouds - the tool projects the points in a 2.5D grid (therefore the interface is similar to the 'Rasterize' tool) - outputs several other statistics (added/removed volume, surface, etc.) * Exclusive full screen mode for 3D views * Support for the NVidia 3D Vision glasses (thanks to Amfax (UK) - www.amfax.co.uk) - new option of the 'Stereo' mode - the graphic card must support OpenGL quad buffering (i.e. latest GeForce or Quadro cards) - the 3D stereo mode must be enabled in the NVidia Control Pannel - the screen frequency must be manually set to the right frequency (i.e. 100 or 120Hz) if not already - (the 3D view is forced to exclusive full-screen mode) - shaders (EDL, etc.) are supported * New method: Edit > Colors > Convert to grey scale - Enhancements: * Point-pair based alignment tool: - the tool 3D view now has the same viewport/camera parameters as the source 3D view * Section Extraction tool: - new 'Unfold cloud' option: lets the user 'unfold' the cloud along the active polyline (very useful to unwrap a cloud that has not a cylinder shape) - the tool now keeps the active GL filter active (if any) * qPoissonRecon now uses PoissonRecon V8.0 (https://github.com/mkazhdan/PoissonRecon) * qAnimation: - the plugin has been fixed and can now output mp4 videos - the user now sees all the viewports in the list and can enable/disable each one separately - new 'loop' option (to close the loop) - the preview can be started from the selected step (see dedicated checkbox) - new 'super resolution' option (to smooth the output video) * Point sampling (on a mesh): maximum density is now 100 M. per square units * Sensors: - Sensors (camera, TLS) are now properly updated when their associated cloud is scaled - principal point added to Camera sensor parameters - differentiation of the vertical and horizontal focal lengths for Camera sensors * Command line mode: - new option '-ORIENT_NORMS_MST' (+ number of neighbors): to (re)orient the normals of the loaded clouds with a Minimum Spanning Tree - new option '-SF_OP' (+ SF index + operation + scalar value): to add / sub / multiply or divide a scalar field by a constant value - new option '-DROP_GLOBAL_SHIFT': to drop the Global Shift information of all loaded entities * SF Arithmetic tool: - a constant value can now be used instead of a second scalar field - the first scalar field (SF1) can now be updated directly (instead of creating a new SF) - division by zero is now properly handled * Structured point clouds: - the sensor position of associated scans is now properly updated when the cloud is rotated * M3C2 plugin: - new option to output the scalar fields (distance, confidence, etc.) on the original core point entity (i.e. either the original cloud or