1 ENVI Training ITT Visual Information Solutions Topographic Production Capability 2 ENVI • ENVI 4.8 with ENVI Zoom –3‐window “classic” interface. – One‐window "Zoom" interface with pan and zoom functionality. – Can launc Zoom or classic from the other interface in a linked session. – 4.8 introduces Viewshed Workflow, LiDAR viewer, and ENVI tools for ArcGIS® Topographic Production Capability 3 Notes on Organization • We’ll be taking one long break for lunch and a couple of short breaks throughout the day. • Training will be organized to go through guided workflows, followed up by challenge exercises. • A comprehensive set of ENVI capabilities will be organized into scenario‐based training. • A few standalone module sections at the end will cover Orthorectification, DEM Extraction, SPEAR, and THOR tools. Topographic Production Capability
32
Embed
ENVI Training - Goozmo, Boulder Colorado Company · 1 ENVI Training ITT Visual Information Solutions Topographic Production Capability 2 ENVI •ENVI 4.8 with ENVI Zoom –3‐window
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
ENVI Training
ITT Visual Information Solutions
Topographic Production Capability
2
ENVI
• ENVI 4.8 with ENVI Zoom– 3‐window “classic” interface.
– One‐window "Zoom" interface with pan and zoom functionality.
– Can launc Zoom or classic from the other interface in a linked session.
– 4.8 introduces Viewshed Workflow, LiDAR viewer, and ENVI tools for ArcGIS®
Topographic Production Capability
3
Notes on Organization
• We’ll be taking one long break for lunch and a couple of short breaks throughout the day.
• Training will be organized to go through guided workflows, followed up by challenge exercises.
• A comprehensive set of ENVI capabilities will be organized into scenario‐based training.
• A few standalone module sections at the end will cover Orthorectification, DEM Extraction, SPEAR, and THOR tools.
Topographic Production Capability
4
Disaster Mitigation ScenarioUsing change detection to guide
reconstruction efforts post‐Tsunami.
Topographic Production Capability
5
Indian Ocean Tsunami Analysis
In the immediate aftermath of the 2004 Indian Ocean Tsunami, image analysis played a key role in assessing damage caused by windstorm, flood and storm surge.
Topographic Production Capability
6
The extracted coastal damage assessment information was then available to GIS users and rescue personnel in their applications to quantify and classify the extent of the damage.
Indian Ocean Tsunami Analysis
Topographic Production Capability
7
Analysis Workflow
• Display and Subset Imagery
• Orthorectification
• Image to Image Registration
• Visualize and explore the imagery
• Perform Change Detection
Topographic Production Capability
8
Subsetting Imagery
• Using Resize Data tool in ENVI
• Spatial and/or Spectral subset
• Spatial subset:– by image
– by coordinates
• Other options– by size of another file
– by Region of interest or
vector file
Topographic Production Capability
9
Image Rectification
• Many tools in ENVI
• Image to Image Registration
• Image to Map Registration
• Convert and customize map projections
• Coordinate Converter
• Orthorectification– RPC
– Rigorous (separate module)
Topographic Production Capability
10
Image Rectification Exercises
• RPC Othorectification– Account for distortions are caused by topography, camera geometry, etc.
– Typically requires:• Image with approximate geolocation information
• RPC model in ancillary file
• Elevation information (DEM)
• Offset between mean sea level and the gravitational potential surface (known as the geoid), so the
Topographic Production Capability
11
Image Registration Exercises
• Image to Image Registration– Selection of Ground Control Points
– Warping and Resampling
Time 1 ImageTime 2 Image
Topographic Production Capability
12
• Image Difference or Thematic Change
• Use vegetation indices if desired
• Cleanup methods for filtering and aggregating results
• Preview changed areas in portal window
Change Detection Workflow
Topographic Production Capability
13
Change Detection• Two change detection algorithms in
ENVI Zoom– Thematic Change – takes two classification images and identifies differences between them. The resulting classification image shows class transitions, for example, from vegetation class to urban class.
– Image Difference ‐ compares two images taken at different times, and it identifies differences between them. The difference can be computed on a specified input band or on a feature index (e.g., veg), with optional thresholding
Topographic Production Capability
14
Change Detection
• Thresholding– Automatic methods
– Manual sliders
• Cleanup step – Smoothing
– Aggregation
Topographic Production Capability
15
Change Detection
• Export– Classification image
– Vectors (shapefiles) HLZ.pptx
Topographic Production Capability
16
Ship Landing ScenarioUsing ENVI to select suitable regionsfor shore approach.
Topographic Production Capability
17
Ship Landing Scenario
• Commercial data import– Combine individual bands into one file and reorder
• Image Mosaicking • SPEAR Relative Water Depth
– Masking• Density Slice• Export to ArcGIS
Topographic Production Capability
18
Relative Water Depth Product
SPEAR ‐ Relative Water Depth
• Generate a product depicting relative water depths (shallow to deep) for a region of interest
• Log ratio or principal components analysis methodologies
• Bottom albedo independent algorithm means sea floor brightness does not affect depth estimates
• Optional absolute depth calibration using ground truth
Quickbird MSI
Topographic Production Capability
19
Ship Landing Scenario• Data import
– Combine individual bands into one file and reorder
• Image mosaicking
Topographic Production Capability
20
Ship Landing Scenario ‐ results
• SPEAR Relative Water Depth
– Color tables
– Density slice
• Density Slice
Topographic Production Capability
21
Helicopter Landing Zone Scenario
• Using ENVI’s LiDAR and Topographic Modeling Capabilities to Determine Areas Where a Blackhawk Helicopter can be landed.
Topographic Production Capability
22
HLZ Requirements
• Exclude areas near Vertical Obstructions
• Exclude Water cover.
• Exclude areas of excessive slope or with otherwise unsuitable topography.
• Combine all exclusions into a single rule image.
• Select areas of appropriate morphology for the helicopter in question (area, shape, etc.)
Topographic Production Capability
23
ENVI Capabilities
• Use the LiDAR Viewer to gain actionable intelligence.
• Construct a vertical obstructions layer using Band Math.
• Mask out water by digitizing vectors aided with region growing.
• Create buffer zones around water and vertical obstructions.
Topographic Production Capability
24
ENVI Capabilities
• Perform topographic modeling.
• Threshold out unsuitable data ranges to an ROI.
• Use convolutions and morphology to locate appropriately sized areas for the landing zone.
Topographic Production Capability
25
What is LiDAR?• Light Detection And Ranging
• You may also see the terms Laser Altimetry or ALSM (Airborne Laser Swath Mapping)
• LiDAR sensors function by reflecting a dense array of laser beams over a surface and using them to record information about that surface.
Topographic Production Capability
26
LiDAR Datasets
• Image analysts work with LiDAR data after it has been preprocessed and turned into elevation raster images.
• DEM, DTM, DSM, and Intensity Image are all names you will see for datasets derived from LiDAR point clouds.
Topographic Production Capability
27
DSM
• The DSM (Digital Surface Model) is the most widely available LiDAR raster product. It is produced through a simple interpolation process, and contains elevation data for objects such as buildings and trees in addition to terrain.
Topographic Production Capability
28
DTM
• A DTM (Digital Terrain Model) is produced through LiDAR processing which filters out buildings, trees, and other above‐ground objects from a DSM to produce a layer which contains only ground elevation values.
Topographic Production Capability
29
LiDAR Viewer in ENVI
• ENVI 4.8 introduces a new high-performance LiDAR viewer• Efficiently use imagery for intelligence and actionable information
Topographic Production Capability
30
• Measure building height and volume
• Determine elevation• Visual Interpretation
Topographic Production Capability
31
Deriving New LiDAR Data• When working with
LiDAR data, you may need to use basic image processing tools to create other layers.
• One simple example is that of the Feature Height layer, which is created by subtracting the DTM from the DSM.
Topographic Production Capability
32
Threshold to ROI
• Specific data ranges can be used to construct ROIs when using the Threshold to ROI tool.
• For the HLZ, we can use it to identify areas which contain vertical obstructions by thresholding out feature height above a certain value.
Topographic Production Capability
33
ROIs and Region Growing
• If simple values cannot be used to mask out areas, then ROIs can be digitized.
• For objects like water bodies which are spectrally homogeneous, region growing can be used once an area inside the feature is digitized.
Topographic Production Capability
34
Topographic Modeling
• Slope, Aspect, and other topographic measures can be extraction from a DEM in ENVI.
• Use Threshold to ROI with these tools to mask out desirable or undesirable values.
Topographic Production Capability
35
Band Math in ENVI
• Band Math is a multi purpose tool which can be used to combine datasets, extract information from various datasets, or compare datasets.
• Don’t overestimate the difficulty of using the band math tool. Most operations use only basic arithmetic.
Topographic Production Capability
36
ENVI Morphology Filters
• Can use Erode to shrink or reduce islands of smaller pixels, or Dilate to fill, expand, or fill holes in an image.
• Morphology filters provide the capability to automatically filter out areas of potential helicopter landing zones that are too small to accept helicopters of a certain size.
Topographic Production Capability
37
Forest Clearance Scenario
• Using ENVI to locate regions where forest has been cleared.
Topographic Production Capability
38
Forest Clearance Scenario
• Can be used to locate and assess:
• Agricultural expansion.
• Illegal logging or production of illegal drugs.
• Construction or expansion of facilities.
Topographic Production Capability
39
Forest Clearance ‐ Logistics
• Forest clearance can be detected from a variety of sensors – presence of NIR and Red bands are critical.
• Forgiving of limits to spatial resolution of the sensor.
• Easily performed on common free, government, and commercial data formats.
Topographic Production Capability
40
Forest Clearance – ENVI components
– Data conversion to radiance or reflectance– Atmospheric correction – Masking cloud cover, water, etc.– Register images– Image Difference Workflow : the difference can be computed on a vegetation index (NDVI).
• Image analysis tools can now be accessed directly from ArcGIS
• ENVI tools work in both Server and desktop environments
• Support for ArcGIS 9.3 and 10
ENVI Tools
Can be accessed from the ArcGIS Desktop
Also from the ArcGIS Server®
Platform
Topographic Production Capability
49
D
Tools include:• Detecting Change• Extracting Features• Classifying land cover
• Create custom ENVI Tools
• Upon installation of ENVI, image processing tools will appear in ArcGIS
ENVI 4.8: Streamlined workflow and further integration with ArcGIS®
Topographic Production Capability
50
– Rosse Ice Shelf, Antarctica– Change detection from 2005-2008
51
• Blue areas = new, positive, change• Red areas = removed, negative
change
52
Orthorectification
• A process that accurately registers image pixels to ground coordinates and removes geometric distortions induced during image capture– Causes of geometric distortions
• Variable scene topography• Off-nadir sensor acquisition geometry• Platform instability or inaccurate knowledge of pointing
geometry
• Result: An image product with properties of a map– Projected to map coordinate system (e.g., UTM)– Constant scale
Topographic Production Capability
53
Applications of Orthoimagery
• Orthoimagery is essential for:– Accurate point coordinate determination– Object measurement (size, shape)– Precise collection of feature data for mapping applications
• Orthoimagery is often used as an image base in Geographic Information Systems (GIS)– Positional accuracy required for registration with map layers– Imagery often provides the most up-to-date information to map
users
Topographic Production Capability
54
ENVI Orthorectification Methods
• RPC-based Method (Rational Polynomial Coefficients)– Uses an approximate sensor model to construct the
geometry relating an image pixel (sample, line) to a map coordinate (X, Y, Z)
• Rigorous Sensor Model Method– Uses the exact sensor model defined by detailed
interior and exterior orientation parameters– Combines several analytical sub-models that:
• Construct a look vector for each image pixel • Intersect the vector with the earth ellipsoid• Project the intersection to a map coordinate system
– Multiple imagesTopographic Production Capability
55
ENVI RPC OrthorectificationRequired Inputs:
– Image– RPC data– Elevation– Geoid offset for
correct elevationOptional Inputs:
– Digital Elevation Model (DEM)
– Ground Control Points (GCPs)
Topographic Production Capability
56
Features and functionality:• Rigorous Sensor Models
• Five step wizard-based workflow– Select Input Imagery and DEM– Collect and Edit GCPs– Select Image-to-Image Tie Points– Order Images and Define Cutlines– Set Output Parameters
Topographic Production Capability
58
Project Management
• Save and Restore Project Files
• Layout Manager– View spatial relationship of
imagery, DEM, GCPs and tie points
– List input data sets• Error Visualization
– View residual errors in tabular and map form
Topographic Production Capability
59
ENVI - DEM Extraction Module
The ENVI DEM Extraction Module allows you to easily create accurate digital elevation models from stereo imagery• Wizard-based tool walks you
through all the necessary steps
• Flexibility allows you to edit, extract features, and more
Topographic Production Capability
60
DEM Extraction: Anaglyph Imagery
• Stereo pairs can be used to create a stereoscopic 3D image which can be viewed with anaglyph glasses.
• Each eye sees a slightly different picture based on chromatically opposite colors and the brain blends the image together.
Topographic Production Capability
61
How DEM Extraction Works
• Creation of Epipolar Images. Epipolar geometry describes the geometrical constraint between two frame images of a stereo pair.
• Image matching: locate points on both left and right images which correspond to the same ground feature.
• DEM Geocoding – The DEM must be oriented in a geographic or projected coordinate system.
• Relative or absolute DEM – whether or not DEM is calibrated to ground truth elevation.
Topographic Production Capability
62
LAS LiDAR to DEM
• LiDAR point clouds contain a lot of information, but 2D gridded formats allow a lot of quick topographic processing.
• ENVI supports import of LiDAR LAS files into a gridded format.
Topographic Production Capability
63
DEM Creation
• From point cloud, to TIN, to Raster.
• First vs. Last return biases.
Topographic Production Capability
64
• LiDAR results contain above surface features.
• End product can be used for topographic modeling, Orthorectification, etc.
Topographic Production Capability
65
Image to Map Registration
ITT Visual Information Solutions
Topographic Production Capability
66
Image to Map Registration
• Picking Ground Control Points– Fractional pixel possible with zoom window
• Warping
• Resampling
• RMS error – based on GCPs so not a direct measure of accuracy
Topographic Production Capability
67
Image to Map Registration• Ground Control Points
Image
Map
Topographic Production Capability
68
Warping Methods• Rotation, Scaling, Translation
– Affine transformation (straight and parallel lines maintained)– O‐order polynomial– Need at least 3 GCPs
• Polynomial Function– Allows shearing– Up to order n where GCPs>(n+1)2
– Need at least 4 GCPs– Shouldn’t go higher than 4th order – “over fit”
• Delaunay Triangulation– Warp model varies with triangle– Output limited to outermost GCPs ‐ However, can use Predict