Christian Weise Product Manager APRIL 2016 What's New in ecognition 9
Introduction
Background ecognition Suite Advanced analysis software and development environment available for geospatial applications Enables users to create feature extraction or change detections solutions to transform geospatial data into geoinformation Tree Map Impervious Map Land Cover Map Elevation Data Image Data Real World
Overview ecognition Suite offers three different components which can be used stand-alone or in combination to cover even the most challenging fully automatic and semi-automated production workflows ecognition Developer: development & analysis environment for applications ecognition Architect: front-end for applications created in Developer ecognition Architect ecognition Developer ecognition Architect ecognition Developer ecognition Server: processing environment ecognition Server ecognition Architect ecognition Server
Use Cases EXTRACT INFORMATION UPDATE INFORMATION
Unique Features
Geospatial Data Fusion Fuse and use a variety of geospatial data, such as images, point clouds and GIS vectors Use the full information potential of your data
Dynamic Analysis Input Raster Controlled with RuleSet Segmentation Classification GIS Analytics Abstraction Result Raster Vector Combines analysis steps Segmentation, Classification, Object Manipulation, and GIS Analytics dynamically Vector Point cloud Point cloud Report Unique approach to translate your mind model (why you see something in the data) into computer understandable code (ecognition Solution [Rule Set])
Highlights in Version 9.2
ecognition 9.2 Highlights Extended Multi-Core processing capabilities for faster project completion Advanced Data Support to reduce time taken in converting and loading data Extended Multi-Core processing capabilities
Extended Multi-Core processing capabilities Further multi-core enhancements in ecognition 9.2 allow users to utilize up to 8 CPU cores with selected functions that typically require longer processing times: Multi-Core powered Multiresolution Segmentation Multi-Core powered Template Matching Multi-Core powered Feature Calculations seconds The new multi-core capabilities maximize the power of computing hardware, reduce processing bottlenecks and allow users to complete projects in up to a third 1000 500 8 cores of the time taken with prior versions 50 100 150 200 megapixel 3000 2500 2000 1500 1 core 4 cores
Advanced Data Support Reduced time taken in converting and loading data - ecognition 9.2 includes a new driver for a faster import of JPEG2000 files and now supports the usage of MrSID image data. Furthermore a new point cloud driver allows the usages of compressed point clouds in the LASzip format. Direct access to latest image and point cloud compression technology Offers faster turnaround times for projects using the newest set of earth observation sensors JP2 SID LAZ
Highlights in Version 9.1
ecognition 9.1 Highlights Performance improvements, significantly reducing processing time Better utilization of GIS-Data in image analysis workflows Improved mapping capabilities for robust and accurate classification results Packaging and installation of customized ecognition applications Enhanced mobile mapping support for Trimble MX2 systems Application Packaging
Performance Improvement Feature: Multi-core support in Multiresolution Segmentation Background: Multiresolution Segmentation is the first and most successful object creation technique in the remote sensing market & the most timeconsuming process of an typical ecognition analysis (~45% of the overall processing time) General trend in CPU development has moved from dual-core to quador hex-core chips - parallel programming techniques can benefit from multiple cores directly Benefits: Performance improvements of up to 50% during object creation freeing time to focus on the interpretation of results and decision making Reduced overall processing time allowing users to faster produce results Utilize the power of current desktop systems by using up to 4 CPU cores
Performance Improvement Performance Increase [%] 50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% 47% 48% 44% 47% 35% 34% 34% 34% 225.000.000 (15k x 15k) 400.000.000 (20k x 20k) 625.000.000 (25k x 25k) 900.000.000 (30k x 30k) Scene Size [pxl] Performance increase with 2 cpu cores Performance increase with 4 cpu cores
Better Utilization of GIS-Data Feature: New integrated GIS-techniques to inspect and analyze vector data, this includes: Buffering Conversions Distance calculations Benefits: Supports the dynamic combination of image and GIS analysis capabilities on a high level of automation which help reducing project turnaround times Further bridging the gap between GIS and Remote Sensing
Improved Mapping Capabilities Feature: Load, save and use sample statistics Background: Supervised learning algorithms analyzes spectral signatures obtained from sample areas and produces an classification model, which will be used for mapping ( convert pixel data into thematic information) Samples represent the class examples and can be stored in GIS layers to save time if users want to re-do or to modify the selection Problem is that new sample locations have to be selected for every new data set, because (beside atmospheric and spectral effects) they could not covered by the new scene(s) Benefits: Extend and transfer sample statistics for classifications make a reselection of sample areas superfluous and allowing users to faster produce results Combine sample statistics from different scenes offers the creation of transferable classification models which produce robust and accurate results
Other new features included Features: 1. Packaging and installation of customized ecognition applications 2. Enhanced mobile mapping support for Trimble MX2 systems Benefits: 1. Allows an easier roll out and commercialization of ecognition-based applications 2. Allow creation of mapping solution based on Trimble MX2 systems
Other new features included License Borrowing Customer Feedback Program
Highlights in Version 9.0
ecognition 9.0 Highlights Native Vector Handling Extend OBIA with GIS operations Template Matching Extend OBIA with computer vision based Object Detection Point Cloud Viewer Increase the productivity of creating Rule Sets JPEG2000 Support Integrate JPEG2000 images in OBIA Workflows Trusted Storage Flexible License Management Point Cloud Viewer JPEG 2000 Native Vector Handling and more Trusted Storage Template Matching
Native Vector Handling Efficient GIS data management inside ecognition, this includes Functions to visualize and inspect GIS vector data Common used GIS operations Algorithms to update and export GIS vectors This improved integration of GIS techniques into Remote Sensing workflows is further bridging the gap between the two domains
Native Vector Handling Visualize and inspect GIS vector layer (1) No Rasterization No Pixel Snapping
Native Vector Handling Visualize and inspect GIS vector layer (2) Edit Vector Layer Mixing dialog Vector related Features + + + +
Native Vector Handling Manipulate Vector Data (1) Combine GIS vector layers Input Union Subtract Intersect Difference Simplify/Smooth GIS vector layers Input Simplified Smoothed
Native Vector Handling Manipulate Vector Data (2) Orthogonalize GIS data layers i.e. Reshape Building Objects Integrate GIS data layers i.e. Combine Input/Result Layers Create GIS data layers i.e. Convert OBIA-Results into Vector Layers Update GIS data layers + i.e. Write Results back in Databases + + + + + + + ++ +++ + ++++++ + + + + + + +++ + + + + + + + + + + + + + + FileGDB +
Template Matching Extends the existing ecognition classification methods with computer vision based object detection, this includes GUI to collect Samples GUI to create and evaluate Templates Algorithm to execute Template Matching Provides additional value remote sensing experts by improving classification results and extending ecognition to new fields of use
Template Matching Template Matching Algorithm algorithm to execute matching procedure If standard deviation of the template image compared to the source image is small enough, template matching may be used Templates are most often used to identify printed characters, numbers, and other small objects (i.e. Trees)
Template Matching Template Editor (Template Generation) Functions to define and select Samples Functions to create Template(s) Functions to test Template(s) and generate additional Samples
Point Cloud Viewer Visualize Point Clouds for an easier and better data and result assessment Functions to render points by classification, intensity, RGB values Functions to define class colors This can significantly increase the productivity of creating rule sets as well as improving the classification quality
Trusted Storage License handling based on files is replaced with Trusted Storage Secure environment for license management Online license activation and re-hosting VM support Provides a Flexible ecognition License Management License Entitlement ID Activation Return
Summary View PCs CV based Object Detection Check PCs 3 rd Party Tools 3 rd Party Tools 3 rd Party Tools JPEG2000 Conversion Inspect Analyze Map Fusion Interpretation GIS Vectors Classified Points Statistics Generalize Results Combine Results Create Maps
Ordering, Availability, and Resources
Ordering, Availability, and Resources ecognition Suite 9.2 Release Date: 13 April 2016 Who s eligible? All users with a valid maintenance contract as of April 13 th, 2016 The software is available in the Customer Download Area: http://www.trimble.com/imaging/downloads Get the Free Trial and learn more about Trimble ecognition on http://www.ecognition.com Visit the ecognition Community (more than 12.000 users) to ask questions, follow discussions, download guided tours, and give feedback on http://www.ecognition.com/community
Contact Sales for Inquiry Find a Reseller Save up to 75% with EDU Discounts