LIDAR Processing SDK for Fully Automated Aerial Point Cloud Analysis
SRI’s LIDAR Processing SDK is specifically designed for industry and academia developers of large-scale mapping solutions with a requirement to process LIDAR point clouds from aerial platforms.
The SRI LIDAR SDK is composed of powerful algorithms and utilities that handle millions of points generated by LIDAR while remaining scalable, development-oriented, and robust. The SDK can process linear or Geiger-mode data and has high accuracy in all terrain types, including mountainous, barren, or urban environments.
LIDAR mapping has many uses in state and local government, emergency response, disaster mitigation and management, forestry, natural resources management, and oil and gas exploration applications. The SDK leverages more than 10 years of SRI leading-edge research in LIDAR analysis to provide high accuracy bare earth and feature extraction across a broad range of platforms.
Accelerate Product Development
With SRI’s LIDAR Processing SDK, developers can focus on the requirements of their mapping software project or emerging product. The additional risk and expenditure of researching, building, and testing automated terrain analysis algorithms are eliminated. End users of LIDAR data who need to map terrain features, vegetation, buildings, and other structures will be interested in the capabilities this SDK adds to processing applications.
Contact SRI for more information about how to develop LIDAR processing applications with robust, automated feature extraction.
LIDAR Processing SDK Feature
Benefit to Developer and End User
|Robust Bare Earth Extraction. Produces an independently derived bare earth layer that is robust to varying data density and terrain type.||An accurate, independently derived bare earth layer is a fundamental step in any feature extraction process, providing necessary context for interpreting scene content.|
|Automatically Adaptive to Sensor Type and Terrain. Works with both linear and Geiger-mode LIDAR sensors at different resolutions without being tuned or customized to a particular sensor or terrain type (mountain, rural, urban, etc.).||
Automated feature extraction saves significant time in data exploitation—there is no need to customize or reconfigure for new products or environments.
|3D Feature Extraction Capabilities. Fully automated detection of terrain features.||Provides detection of the most general classes of features and objects to meet the broadest needs of the end user.|
Robust Bare Earth Extraction
Robust bare earth extraction is critical for accurate 3D analysis and feature detection in urban, rural, high-relief, and forested areas. Precise identification of the ground level—and of the points that lie on the ground—is not only necessary for terrain feature detection such as roads, but also for segmentation and classification of structures that are above the ground.
Automatically Adaptive to Sensor Type and Terrain
Since a bare earth derived from LIDAR is a fundamental layer, it must be accurate in all environments without requiring reconfiguration or hand adjustments. This is a challenge given the diversity of potential environments: barren desert, forested, rugged or mountainous, terraced valley, and heavily urbanized areas. A distinguishing aspect of the SRI bare earth algorithm is that it automatically adapts to the ruggedness of the terrain, making it highly accurate at labeling ground with steep slopes and discontinuities.
3D Feature Extraction Capabilities
Based on the bare earth segmentation, SRI adds fully automated feature extraction capabilities for aerial LIDAR exploitation. Building and foliage segmentation can be performed on source data from many challenging environments, including dense urban landscapes with several thousand buildings per square kilometer and heavily foliated areas with extensive tree canopy. In particular, the toolkit allows for automatic extraction of the following features:
- Bare earth
- Low foliage
- Tree canopy
- Small man-made structures near ground
Windows XP or 7
32bit or 64bit (64bit recommended)
2GHz or faster
4GB minimum (8GB or more recommended)