Date: 4/19/22

Location: Zoom Only

Speaker: Willow Coleman (Harvey Mudd)

Abstract: Remotely sensed satellite imagery can be an important data source for observing and measuring environmental changes in urban areas. For example, thermal satellite imagery can be used to identify urban heat islands and provide motivation for urban cooling strategies. However, the spatial resolution of thermal satellite imagery is insufficient for many urban planning applications (e.g., tree planting and urban heat island monitoring) because of the inherent heterogeneity and complexity of cities. In order to improve the usability of this imagery, artificial intelligence techniques  can increase the spatial resolution of the thermal imagery without losing valuable spectral information. One such technique is pansharpening that fuses high-spatial resolution panchromatic (single band gray-scale) images and lower-spatial resolution multispectral and thermal infrared images. This project uses a modified generative adversarial network (GAN) to pansharpen lower resolution, remotely sensed thermal satellite imagery using high spatial resolution thermal imagery. This will involve developing a novel training dataset with patch-pairs of co-located thermal (70 m) and RGB imagery (3-5 m). Using this novel dataset, this project will assess whether a pansharpening model (PanColorGAN) trained on RGB and panchromatic imagery can be successfully applied to the RGB-thermal patch-pairs to solve the thermal pansharpening problem. In addition, this project will determine if a pansharpening model (pix2pix) trained on the RGB-thermal patch-pairs produces higher quality results than the model with pre-trained weights. Both pansharpening approaches will be validated to assess how realistic the resulting sharpened thermal imagery is. This approach will be applied to five cities with variable climate-urban environments across the United States: Austin, TX, Boulder, CO, Chicago, IL, Los Angeles, CA, and Washington, D.C. This project will fill a critical gap by investigating the feasibility of GANs for generating higher resolution thermal imagery that is valuable for studies in heterogeneous urban environments.

Bio: Willow Coleman is a senior at Harvey Mudd College in Claremont, CA, where she is majoring in Mathematical and Computational Biology with an emphasis in Environmental Analysis. During her time at Harvey Mudd, she completed multiple internships in remote sensing at NASA JPL, NASA DEVELOP, and Planet Labs. Today, she will be presenting on her senior thesis research that has taken place over the past year at Harvey Mudd under the guidance of Natasha Stavros (CU Boulder) and Matina Donaldson-Matasci (HMC). After graduation, she will be joining the Shapiro Lab at Caltech as part of the Schmidt Academy for Software Engineering to research functional ultrasound imaging for brain-computer interfaces.