Development of a UAV-mounted Holographic Cloud Particle Imager

Many current climate models assume a homogeneous and uncorrelated spatial distribution of the particles within clouds.  In situ measurements point toward small-scale (mm to m) correlations between particles due to droplet inertia and turbulence and adjusting climate models to account for the inhomogeneity of clouds would increase the accuracy of climate predictions. The spatial distribution of droplets in a cloud influences radiative transfer, collision and coalescence, and droplet growth by condensation. This work presents the development of a UAV-mountable holographic cloud particle imager (HCPI) that measures both the 3D spatial distribution and size distribution of cloud particles in the 10 μm to several millimeter size range. The instrument is currently configured for a lab bench, but the UAV-mountable components are in the design phase.  The instrument is expected to be airborne by 2018.

It is difficult to measure the 3D spatial distribution of cloud particles with existing commercial instruments because they take 2D cross-sections of volumes of air, while a HCPI can record holograms that contain the 3D locations and size of each particle within the sample volume. An existing HCPI is too large to be flown on a UAV, but has made useful measurements on piloted aircraft. UAVs allow measurements to be taken during longer flights to increase data collection as well as in locations that are too dangerous to send a pilot and crew. 

This instrument uses in-line holography, a common technique due to its simplicity and the resolution constraints of currently available imagers, to generate cloud particle holograms. An in-line hologram is recorded by illuminating the sample volume with a collimated beam of laser (coherent) light. The interference between the unscattered light in the beam and the light that scatters off the particles in the sample volume creates a fringe-like pattern at the imager. Using short laser pulses, holograms can be recorded within nanoseconds, eliminating any blur that would occur due to motion of the UAV. Diffraction theory enables a numerical reconstruction of the particle positions and sizes within the sample volume.

Each hologram allows for the sample volume to be divided into an arbitrary number of reconstruction planes with every plane associated with a certain distance from the imager. Each reconstruction is essentially a photograph that contains images of particles, some in focus, some not. An image-analyzing algorithm picks the particles that are in focus and determines their size and location within the volume.  The reconstruction and image analysis are time intensive and are performed in a parallelized fashion on a graphics processing unit (GPU). From this information a pair correlation function can be calculated to determine the degree of homogeneity in the cloud particle spatial distribution.

The performance of the system has been characterized for resolution with standard resolution targets and monodisperse microspheres.  The reconstruction and image-analysis algorithms have been evaluated for speed and accuracy. Artificial cloud-like conditions have also been examined in the lab. Heating elements with temperature feed-back are used to enable operation over the range of temperatures expected to be experienced during flight.

The earliest form of the instrument will allow for the holograms to be downloaded by the user upon landing, but in-flight data transfer is a future possibility. Expanding the sample volume in future versions is possible by increasing the size of the imager (adding pixels). This can be done with existing CCD and CMOS cameras, but increases the computational burden of reconstruction and image analysis. Further work on algorithm development will make this feasible.