Manipulation of radar readings

A method comprising:

• using a radar sensor to emit a radar signal and receive back reflections thereof, and thereby produce a set of radar readings distributed amongst a plurality of distance bins each corresponding to a different respective distance of reflection, wherein in each distance bin that contains a radar reading, the radar reading comprises at least a respective azimuth angle and elevation of the radar reading;
• projecting at least some of the radar readings onto an image comprising a 2D Cartesian grid of pixels based on the respective azimuth angles and elevations, wherein for each pixel where a radar reading is present the pixel comprises a respective value of at least one non-binary channel comprising at least one of measured property of the radar reading and a respective value of a binary channel comprising a binary indicator,
• wherein the binary indicator is asserted if the pixel contains one of the radar readings but not asserted otherwise, or
• the binary indicator is asserted if the radar reading contains a radar reading above a noise floor but not asserted if the pixel contains a radar reading below the noise floor;
• and
• using a machine learning model designed for image recognition, to receive the image and to detect an object therein, wherein the machine learning model performs the detection based on the values of the binary and non-binary channels for each pixel, including pixels with both asserted and non-asserted values of the binary channel.

A computer program embodied on a non-transitory computer-readable medium or media, the computer program being configured so as when run on one or more processors to perform a method comprising:

• receiving a set of radar readings distributed amongst a plurality of distance bins each corresponding to a different respective distance of reflection, wherein in each distance bin that contains a radar reading, the radar reading comprises at least a respective azimuth angle and elevation of the radar reading;
• projecting at least some of the radar readings onto an image comprising a 2D Cartesian grid of pixels based on the respective azimuth angles and elevations, wherein for each pixel where a radar reading is present the pixel comprises a respective value of at least one non-binary channel comprising at least one of measured property of the radar reading;
• and
• using a machine learning model to receive the image and to detect an object therein, wherein the machine learning model performs the detection based on at the values of the at least one non-binary channel for each pixel.