Radar Scan Conversion: The Hows, Whys and Where Nexts?

Radar scan conversion is the process of converting a sequence of radar returns into an image for display. Each radar return is an array of amplitude values corresponding to different range values for a specific angle. As the radar rotates, a new array of values is created for the next angle. Each sample is therefore defined by a range and angle and that is converted into an x and y value for display as an image. In its simplest form, radar scan conversion is a conversion from samples indexed by range and angle to samples indexed by x and y with simple trigonometric functions (x = range * sine (angle) and y = range * cos (angle) converting the coordinates.

When software came to replace hardware

In the early days of radar display, notably before modern graphics cards and multi-core CPUs came along, radar scan conversion was handled with dedicated hardware and special-purpose display hardware. The processing hardware would handle the coordinate transformation of the radar and the display hardware would then blend the radar data with graphics data. This solution worked, and indeed many such installations are still deployed and working well, but it was expensive in proprietary hardware, for example based on racks of VME boards. The transition to software-based processing and display mixing began around 2007, which is when Cambridge Pixel started in business.  Largely driven by the demands of the computer gaming market, graphics capabilities advanced leaps and bounds, so that high-resolution true-colour displays could be created, with multiple layers of graphics being combined into a displayed image. The ability to blend several different layers was exactly what was needed for radar displays - see Figure 1 for an example. The radar image needed to be transparently blended with a background to give the impression of the radar fading to the colour of the background (rather than fading to black, for example). This was made possible by commodity graphics cards which were relatively inexpensive, but also used standardised programming libraries from Windows, Open GL or X Windows. When combined with multi-core processors that could handle the radar scan conversion processing without overloading the main processing loop, the result was a paradigm shift to a low cost, standardised hardware platform, with the additional benefits of flexibility that software solutions can offer. 

Since Cambridge Pixel launched its first software-based radar scan converter in 2007 much has changed. The software approach enabled a rapid evolution of new ideas, so what could be imagined one week could be prototyped the next and valuable ideas quickly became extensions to the product. 

Data sampling

The basic conversion of range and angle data into x and y coordinates is the essence of radar scan conversion but it is just the start of what is needed for a viable solution...