Contact Deutsch
open4aviation

DIBMETSAT - Digital Image Processing Supported Meteorology Services for Air Traffic Management

Air traffic controllers face the challenge of dealing with a flood of information. An essential source of information here are images and measurements of current weather conditions. Providing automated, higher-level messages from weather radar and satellite images, combining them and generating new measurements for an improved visibility estimation should provide substantial support.

Short Description

Motivation

All essential operations (tactical planning of capacity, sector configuration, required staffing, runway configuration, routing of incoming aircraft) in Air Traffic Management are based on the most accurate measurements and forecasts of current weather conditions. An essential source of information is images of various origins, in particular weather radar and camera observations. By means of weather radar, an assessment of precipitation, icing risk, thunderstorm extent and strength, and associated dangers can be predicted. Due to Austria being mountainous and obstacles in transmitters' vicinity, weather radar measurements are affected by disturbances and/or cannot cover all areas of Austria.

Objectives

  • To automate weather observation messages.
  • To correct coverage gaps in weather radar data.
  • To correct interfering signals (non-meteorological echoes).
  • To evaluate standard cameras for local forecasts.

Content

With the weather radar, the received signal strength of the pulsed signal in the microwave range is recalculated to a precipitation equivalent on the basis of the backscatter on cloud particles, and a spatially and temporally high-resolution actual state of the troposphere is provided. The quality of these systems is highly dependent on data quality and coverage of the entire observation area, so it is necessary to complement unrecorded zones by combining them with other sensors and eliminating noise without losing significant meteorological information.

Methodology

Based on our partners', JR and ARC, in-depth knowledge of digital image processing, sensor data fusion, classification and pattern recognition, strategies, methods and algorithms are being developed to optimise the meteorological use of existing and operating weather radar infrastructure. 

It is an investigation of how artefacts and disturbances in weather radar data can be detected automatically, and how data from the twelve meteorological satellite channels can interpolate the weather radar data as temporally and spatially accurately as possible.
The analysis and evaluation of visibility and cloud formations with commercially available cameras shall generate measured values for small-scale weather forecasts, and corroborate the reconstructed weather radar data.

Expected results

Improving weather radar images is a key factor for more accurate forecasts of weather phenomena. Image processing methods (including texture analysis and geometric operations) are used to identify areas disturbed by artifacts or areas that are incomplete due to shading in the weather radar image. The recognised areas are meteorologically expediently supplemented and filled in by means of classification and tracking methods from multispectral satellite data (Meteosat Second Generation). 

For plausibility checking, local weather forecasts (visibility, cloud coverage) are emulated by means of cameras available on the market as well as methods of illumination compensation, feature extraction and local modeling in accordance with the method of human weather observation.

Outcome

Proof of possibility and first prototype system for weather radar correction by means of image processing and correction via meteorological satellite data. Preparation of local weather forecast from camera images.

Downloads

Endbericht DIBMETSAT (pdf, 2.74 MB)

Project Partners

Funding program: TAKE OFF