Real-time Processing and Visualization of Massive Air-Traffic Data - - PowerPoint PPT Presentation

Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes

Digital Landscape Architecture 2015, Dessau

Stefan Buschmann, Matthias Trapp, and Jürgen Döllner

Hasso-Plattner-Institut, Universität Potsdam

05.06.2015

Movement Data

◮ Movement data

◮ Traffic data (e.g., road, naval, or air-traffic) ◮ Pedestrian movements ◮ Animal movements

◮ Features of movement data

◮ Spatio-temporal geodata ◮ Often represented by spatial trajectories ◮ Large data sets (in both, the spatial and temporal dimension) ◮ Advancing technology for real-time acquisition, transfer, and storage

◮ Visualization of massive movement data in digital landscapes

◮ Visualization of dynamic phenomena ◮ Embedded into 3D virtual environments such as digital landscape models,

city models, or virtual globes

◮ Interactive visualization, exploration, and analysis of 3D movement data ◮ Visual Analytics

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 2

Movement Data

◮ Movement data

◮ Traffic data (e.g., road, naval, or air-traffic) ◮ Pedestrian movements ◮ Animal movements

◮ Features of movement data

◮ Spatio-temporal geodata ◮ Often represented by spatial trajectories ◮ Large data sets (in both, the spatial and temporal dimension) ◮ Advancing technology for real-time acquisition, transfer, and storage

◮ Visualization of massive movement data in digital landscapes

◮ Visualization of dynamic phenomena ◮ Embedded into 3D virtual environments such as digital landscape models,

city models, or virtual globes

◮ Interactive visualization, exploration, and analysis of 3D movement data ◮ Visual Analytics

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 3

Movement Data

◮ Movement data

◮ Traffic data (e.g., road, naval, or air-traffic) ◮ Pedestrian movements ◮ Animal movements

◮ Features of movement data

◮ Spatio-temporal geodata ◮ Often represented by spatial trajectories ◮ Large data sets (in both, the spatial and temporal dimension) ◮ Advancing technology for real-time acquisition, transfer, and storage

◮ Visualization of massive movement data in digital landscapes

◮ Visualization of dynamic phenomena ◮ Embedded into 3D virtual environments such as digital landscape models,

city models, or virtual globes

◮ Interactive visualization, exploration, and analysis of 3D movement data ◮ Visual Analytics

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 4

Visualization of Movement Data

◮ InfoVis

◮ Visualization of complex

spatio-temporal data

◮ Visualization of attribute

values

◮ GIS

◮ Analytical view ◮ Often embedded in a map

context

◮ Temporal aspects

◮ Color mapping ◮ Space-Time Cube ◮ Animation Traffic volumes in the city of Potsdam (Google Maps, https://maps.google.de). Tominski, C., Schumann, H., Andrienko, G. & Andrienko, N.: Stacking-Based Visualization of Trajectory Attribute Data, IEEE Transactions on Visualization and Computer Graphics(18, 12), 2012.

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 5

Visualization of Movement Data

◮ InfoVis

◮ Visualization of complex

spatio-temporal data

◮ Visualization of attribute

values

◮ GIS

◮ Analytical view ◮ Often embedded in a map

context

◮ Temporal aspects

◮ Color mapping ◮ Space-Time Cube ◮ Animation Traffic volumes in the city of Potsdam (Google Maps, https://maps.google.de). Tominski, C., Schumann, H., Andrienko, G. & Andrienko, N.: Stacking-Based Visualization of Trajectory Attribute Data, IEEE Transactions on Visualization and Computer Graphics(18, 12), 2012.

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 6

Visualization of Movement Data

◮ InfoVis

◮ Visualization of complex

spatio-temporal data

◮ Visualization of attribute

values

◮ GIS

◮ Analytical view ◮ Often embedded in a map

context

◮ Temporal aspects

◮ Color mapping ◮ Space-Time Cube ◮ Animation Traffic volumes in the city of Potsdam (Google Maps, https://maps.google.de). Tominski, C., Schumann, H., Andrienko, G. & Andrienko, N.: Stacking-Based Visualization of Trajectory Attribute Data, IEEE Transactions on Visualization and Computer Graphics(18, 12), 2012.

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 7

Digital Landscapes

◮ 3D virtual environments

◮ Digital landscape models ◮ Terrain models ◮ Vegetation models ◮ 3D virtual city models

◮ Features

◮ Complex geometry ◮ Costly rendering

◮ Scenery for InfoVis?

◮ Visualize dynamic

phenomena

◮ Support interactive

exploration and analysis

3D virtual city model of the city of Nuremberg (image created by 3D Content Logistics, 2015).

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 8

Digital Landscapes

◮ 3D virtual environments

◮ Digital landscape models ◮ Terrain models ◮ Vegetation models ◮ 3D virtual city models

◮ Features

◮ Complex geometry ◮ Costly rendering

◮ Scenery for InfoVis?

◮ Visualize dynamic

phenomena

◮ Support interactive

exploration and analysis

3D virtual city model of the city of Nuremberg (image created by 3D Content Logistics, 2015).

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 9

Digital Landscapes

◮ 3D virtual environments

◮ Digital landscape models ◮ Terrain models ◮ Vegetation models ◮ 3D virtual city models

◮ Features

◮ Complex geometry ◮ Costly rendering

◮ Scenery for InfoVis?

◮ Visualize dynamic

phenomena

◮ Support interactive

exploration and analysis

3D virtual city model of the city of Nuremberg (image created by 3D Content Logistics, 2015).

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 10

Visualization of Movement Data in Virtual Landscapes

◮ Challenges

◮ Handle massive amounts of

trajectories in high-resolution data sets

◮ Geometric complex, high

detailed 3D scenes for digital landscapes

◮ Maintain interactivity for

exploration and mapping

◮ Goals

◮ Avoid additional creation

and storage of large geometry

◮ Reduce integration costs

(e.g., costly updates of geometry)

Example of dynamic spatio-temporal data: frequency data based on aggregation of traffic volumes. Visualization of frequency data using a 3D city model as context and scenery.

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 11

Visualization of Movement Data in Virtual Landscapes

◮ Challenges

◮ Handle massive amounts of

trajectories in high-resolution data sets

◮ Geometric complex, high

detailed 3D scenes for digital landscapes

◮ Maintain interactivity for

exploration and mapping

◮ Goals

◮ Avoid additional creation

and storage of large geometry

◮ Reduce integration costs

(e.g., costly updates of geometry)

Example of dynamic spatio-temporal data: frequency data based on aggregation of traffic volumes. Visualization of frequency data using a 3D city model as context and scenery.

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 12

Our Approach (1/2)

◮ GPU-based rendering pipeline

◮ Interactive spatio-temporal filtering ◮ Generic mapping of trajectory attributes to

geometric representations and appearance

◮ Real-time rendering within 3D virtual environments

◮ Advantages

◮ Processing and rendering of massive data sets ◮ Maintaining small memory footprint ◮ Configurable on-the-fly geometry generation Comparison of a traditional forward-rendering visualization pipeline (top) with our GPU-based mapping approach (bottom).

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 13

Our Approach (2/2)

◮ On-the-fly geometry generation

◮ Input data is represented and managed entirely on the GPU ◮ Real-time mapping of data attributes to visual properties, such as type of

geometry, width/radius, color, texture mapping, and animation

◮ Interactive configuration of the mapping can be applied based on data

attributes, classification, or user interaction

◮ Applications

◮ Real-time adjustment of mapping options ◮ Interactive spatial and temporal exploration ◮ Interactive generation of density maps Supported basic geometry types for attribute mapping: (1) lines, (2) tubes, (3) ribbons, and (4) spheres.

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 14

Real-Time Trajectory Rendering

◮ Interactive trajectory rendering

◮ Real-time exploration of massive trajectory data sets ◮ Spatial, temporal, and attribute-based filtering ◮ Interactive mapping ◮ Visualization of attributes using mapping configurations

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 15

Real-Time Aggregation and Density Maps

◮ Real-time aggregation of trajectories

◮ Generate density maps at arbitrary spatial and temporal scales ◮ Real-time exploration ◮ Spatial, temporal, and attribute-based filtering ◮ Visualization of differences and changes over time Visualization of density maps of moving objects: aggregated view on air-traffic movements over the time period of a week (left), comparison of two time periods using distinct color channels red and blue (right).

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 16

Visualization of Massive Trajectory Data Sets (2/2)

◮ Visualize large numbers of

trajectories

◮ Interactive exploration and

filtering

◮ Use mapping configurations

to visually distinguish classes

• f trajectories (e.g.,

approaching and departing air planes, or aircraft types)

Visualization of approaching (red) and departing (blue) aircrafts, depicting direction (texture mapping and animation) and velocity (texture stretching, animation speed, and color).

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 17

Visualization of Massive Trajectory Data Sets (2/2)

◮ Visualize large numbers of

trajectories

◮ Interactive exploration and

filtering

◮ Use mapping configurations

to visually distinguish classes

• f trajectories (e.g.,

approaching and departing air planes, or aircraft types)

Visualization of different aircraft types: the weight class of aircrafts is depicted by diameter and color (from red for large aircrafts to green for light aircrafts).

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 18

Individual Trajectory Visualization

◮ Detailed visualization of

individual trajectories

◮ Visualization of trajectory

attributes by attribute mapping and classification

◮ Use various geometric

primitives to distinguish between different features

Classification based on the time-stamp of each sample points: Detailed visualization (speed and acceleration) of trajectories in the vicinity of an airport, discrete visualization of far-away sample points.

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 19

Exploration and Interaction

◮ Image-based selection of

trajectories by user input

◮ Highlighting of selected

trajectories using distinct visual styles

◮ Choose mapping styles to

display selected trajectories in more detail, or visualize different sets of attributes

Highlighting of a trajectory representing a missed-approach on an airport, visualizing the current speed using color, texture, and animation.

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 20

Detail-And-Overview

◮ Overview visualization by

means of a density map

◮ Detailed inspection of

individual trajectories within the context

Detail-and-overview Visualization.

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 21

Temporal Exploration

◮ Space-Time-Cube (STC): map

the time attribute to the visual z-axis

◮ Understand the temporal order of

events, but omit the 3D characteristics of movements

◮ Examine temporal features and

relationships for a number of trajectories

STC visualization of approaching and departing aircrafts. Detailed spatio-temporal examination of a single trajectory.

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 22

Demonstration

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 23

Conclusions

◮ Generic technique for visualizing large movement data for a number of

use cases

◮ Air traffic impact using landscape/city models ◮ Pedestrian movements ◮ Animal movements ◮ Car traffic

◮ Support interactive Visual Analytics / Big Data Analytics of large

spatio-temporal data in digital landscapes

◮ Use of digital landscapes as a computational model and scenery for

data analytics

◮ What role can Exploratory Visual Analytics play for GeoDesign?

◮ Predictive Analytics ◮ Prescriptive Analytics

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 24

Thank You!

Dipl.-Inform. Stefan Buschmann stefan.buschmann@hpi.uni-potsdam.de Computer Graphics Systems

• Prof. Dr. Jürgen Döllner

Hasso-Plattner-Institut für Softwaresystemtechnik GmbH www.hpi3d.de Thanks to DFS Deutsche Flugsicherung GmbH for the provided data.

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 25

Our Approach

◮ GPU data representation

Central attribute storage buffer that is streamed to the GPU.

◮ Visualization configurations

Define how attribute values are mapped to visual properties.

◮ Dynamic data pulling

Fetch attribute data based on selected configuration.

◮ Geometry creation and attribute mapping

The actual geometry is created on-the-fly and passed on for rendering.

◮ Real-time rendering

The generated geometry is rendered according to the configuration.

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 26

Our Approach

◮ GPU data representation

Central attribute storage buffer that is streamed to the GPU.

◮ Visualization configurations

Define how attribute values are mapped to visual properties.

◮ Dynamic data pulling

Fetch attribute data based on selected configuration.

◮ Geometry creation and attribute mapping

The actual geometry is created on-the-fly and passed on for rendering.

◮ Real-time rendering

The generated geometry is rendered according to the configuration.

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 27

Our Approach

◮ GPU data representation

Central attribute storage buffer that is streamed to the GPU.

◮ Visualization configurations

Define how attribute values are mapped to visual properties.

◮ Dynamic data pulling

Fetch attribute data based on selected configuration.

◮ Geometry creation and attribute mapping

The actual geometry is created on-the-fly and passed on for rendering.

◮ Real-time rendering

The generated geometry is rendered according to the configuration.

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 28

Our Approach

◮ GPU data representation

Central attribute storage buffer that is streamed to the GPU.

◮ Visualization configurations

Define how attribute values are mapped to visual properties.

◮ Dynamic data pulling

Fetch attribute data based on selected configuration.

◮ Geometry creation and attribute mapping

The actual geometry is created on-the-fly and passed on for rendering.

◮ Real-time rendering

The generated geometry is rendered according to the configuration.

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 29

Our Approach

◮ GPU data representation

Central attribute storage buffer that is streamed to the GPU.

◮ Visualization configurations

Define how attribute values are mapped to visual properties.

◮ Dynamic data pulling

Fetch attribute data based on selected configuration.

◮ Geometry creation and attribute mapping

The actual geometry is created on-the-fly and passed on for rendering.

◮ Real-time rendering

The generated geometry is rendered according to the configuration.

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 30

Our Approach

◮ GPU data representation

Central attribute storage buffer that is streamed to the GPU.

◮ Visualization configurations

Define how attribute values are mapped to visual properties.

◮ Dynamic data pulling

Fetch attribute data based on selected configuration.

◮ Geometry creation and attribute mapping

The actual geometry is created on-the-fly and passed on for rendering.

◮ Real-time rendering

The generated geometry is rendered according to the configuration.

Buschmann, S.

• Real-time Processing and Visualization of Massive Air-Traffic Data in Digital Landscapes
• 31