Pioneered a decade ago for remote sensing, eCognition enables geospatial data such as images, point clouds and GIS/CAD vectors to be integrated and analyzed to quantify features or detect changes over time. eCognition 9 simplifies and reduces the time taken to classify objects in data sets using deep learning, object based image analysis, and template matching functions. In addition, the lasted version adds also 3D point cloud analytics to the software so customers can perform a broader range of geospatial analysis with a greater level of control.
By leveraging deep learning technology from the Google TensorFlow™ library, eCognition empowers customers with highly sophisticated pattern recognition and correlation tools that automate the classification of objects of interest for faster and more accurate results. eCognition now provides new algorithms to directly leverage this state of the art machine learning technology. The new tools include a trainable convolutional neural network model and algorithms for the automatic generation of sample patches, train and apply model as well as the ability to save and load models into eCognition.
The new 3D point cloud capabilities in eCognition enable users to integrate aerial and terrestrial point cloud data to perform complex 3D data classification, extract information and analyze change over time. New algorithms allows automated point cloud classification and transferring thematic information between point clouds. The latest eCognition version provides new point cloud viewing features to get the full information potential of your input data and to effectively combine raster, vector and point cloud data within your project.
eCognition extends the existing knowledge-based and supervised classification methods with computer vision based object detection. The Template Editor Window allows users to easily collect samples to define the search template that is applied over the imagery data. Additionally, a template matching algorithm is available to create a correlation coefficient layer used to detect objects.
With the introduction of multi-core processing in eCognition, customers have been enabled to complete projects faster than before. Further enhancements in eCognition version 9 allow users to utilize up to 8 cores with selected algorithms that typically require longer processing times (multiresolution segmentation, template matching, and layer arithmetics). Additionally, the calculation of feature values on-the-fly in the feature view window is also multi-core enabled. The multi-core capabilities maximizes the power of computing hardware, reduces processing bottlenecks and allows users to complete projects in up to a third of the time taken with prior versions.