Is it possible to make 3D point cloud with RADAR? Both LIDAR and RADAR are EM waves, but with different wavelength. So, if it is technically possible to make point clouds with LIDAR, why is it not possible to do the same with a RADAR?

Is it because, the RADAR cannot have narrow beams like LIDAR? Or the problem is more related to the antenna? Also, why isnt doppler effect experienced on LIDARs?

In short, the question is how are RADAR and LIDAR technically different with respect to sensors?


A point cloud is related to how data is acquired and stored. One can usually transform between point clouds and rasters and say, since they can be be used in similar ways, the answer is essentially yes. The most similar type of RADAR to a LIDAR point cloud is a Synthetic Aperture RADAR which is just called SAR.

enter image description here

and contrast it with a LIDAR Image , centered on the same place. This might not be the best comparison but adequate for a 3 minute Google search.
enter image description here

clearly LIDAR has higher resolution. The SAR image is formed from a coherent process of data acquired over a scan. LIDAR points are one time, travel time shots. Each measurement adds a point. I'm not aware of coherent LIDAR processing but I'm not State of the Art LIDAR researcher. Doppler has an effect on the phase of the signal, and an incoherent energy detector is insensitive to phase. You could take a SAR image and turn it into something like a point cloud but the way the data is collected and processed is different, so maybe your answer is no.

LIDAR, used like this are usually looking down over a narrow range of angles under the aircraft. SAR can more offset over a wider range of angles. The SAR image is probably from a single flight path. A LIDAR Image of the same extent would take a number flight paths. One can use SAR in orbit. I'm not sure about LIDAR through a lot of atmosphere.

SAR will probably have the advantage with respect to transmission power for some time.

The highest resolution global digital elevation maps that are publicly available are derived from SAR data.


LIDAR derived elevation data is growing.

requested edit:

A raster is a regular sampling, like uniform sampling. Each point in a SAR image is produced by a calculation involving data that was collected over a time window, that for a pulsed waveform, would span multiple pulses. One needs to select those solution points. Adjacent solution points overlap in the raw collected data input to the calculation. SAR raw data can include polarization.

A point cloud is a set of individual measurements and the points aren’t regularly spaced.

SAR is a interferometric or coherent processor. SAR has some history to it and was first processed using analog processing. Digital processing is the norm today.

Im not familiar with coherent Lidar but the term heterodyning, in my experience is analog processing. Cloud sampling isn’t a necessary theoretical requirement for Lidar. The article that is cited in the comment speaks of continuous wave coherent Lidar which would seem to me a raster sort of sampling.

Addressing what one can do that the other can’t, has categories like can’t now, and can’t never. The can’t now category shifts with technology. The can’t never ignores the possibility of an ultra Wideband, ultra agile, hybrid system.

The missing comparison to Lidar is stereo vision and even in that case, they aren’t mutually exclusive.

  • $\begingroup$ Thank you for the great explanation. What do you mean by a raster ? And its clear now, that its seldom spoken about Doppler in LiDAR because their receptors are incoherent. But I read something about coherent LiDAR detectors here - lidar-uk.com/how-lidar-works. Is there a reason, RADARs do not have narrow beams like LiDAR? Do you know, what is that a LiDAR can do what a RADAR cannot and vice versa? Could you please update your answer with my following queries? $\endgroup$ – infoclogged Feb 3 '18 at 8:54

A point cloud in theory could be made with just about any ranging sensor, but I think you are right that for practical purposes the main reason is the beamwidth. An 800(nm) lidar has a wavelength several orders of magnitude smaller than even an Ku-band radar, and since beamwidth is directly related to wavelength, the beams are much narrower. This lets lidar systems get both fine range and angle measurements with a single pulse. A returned radar pulse, on the other hand, will contain reflections from a much larger angle, which at typical standoff ranges could mean kilometers-worth of targets. If you want to convince yourself of this, calculate the arc length of, say, a 1(ft) dish at a distance of 10(km) with a 0.03(m) wavelength, which is a typical X-band wavelength used in radar, and a 800(nm) wavelength. (I don't know if they use dishes per se, but I bet you can find typical beamwidths online.)

However, there are ways to coherently process radar data to compensate for the wide beamwidths. Also, multiple passes of the sensor at different altitudes (technically any non-linear flight path gives 3-D resolution) could be used to get 3-D resolution. But as a practical matter, it's easier to just use a lidar system.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.