Small body exploration provides three main benefits to our society: (1) as they are remnants of the early solar system, they are vital to our understanding of planetary formation, (2) they aid our planetary defense efforts, (3) they may be the key to the sustained exploration of our solar system. In order to facilitate the aforementioned benefits, characteristics of the body such as the orbit, size, shape, spin pole axis are necessary. Such characterization is performed by two main methods: ground based and spacecraft based. Ground based optical and radar telescopes are used to conduct large scale surveys of the solar system. These surveys produce a large catalog of small boides with acccurate orbit and size characterization but the spin pole axis and shape are often very poor. In order to produce improved characterization of any specific body of interest (e.g due to its high scientific value, high risk of collision etc.) spacecraft based is used. In this regime, a spacecraft with sensors such as cameras are sent to the body of interest and improved size, shape and pole axis information is obtained. The current state of the art algorithm to perform shape and pole axis estimation from spacecraft images is known as stereo-photoclinometry (SPC). SPC requires the use of high resolution optical images taken from specific illumination geometries to produce accurate shape and pole axis. The viewing geometry requirement in addition to the computational burden of the high resolution images prohibits the on-board, autonomous usage of SPC. Instead, a human-in-the-loop process is required in which a ground team processes the spacecraft images to produce the accurate spin pole axis and shape. Moreover, as optical images require an illumination source, SPC is not illumination robust.

In this thesis, we present algorithms that leverage infrared images from a spacecraft to estimate the pole axis and shape of a small body. In contrast with optical images, infrared images are more illumination robust as they detect re-radiated light in the infrared spectrum instead of reflected light in the optical spectrum. The first algorithm obtains the pole estimate from a set of on-board infrared images. Due to the aforementioned advantage of infrared images, the proposed method is applicable in a vast majority observational geometries. In addition, the algorithm doesn’t require a prior estimate. Once the pole axis is known, the silhouette (i.e. the edge between the body and the background) of space can be used to extract the shape of the body. While such Shape from Silhouette (SfS) methods can be applied to optical, infrared images offer the aforementioned robustness to illumination as well as more accurate characterization of the body’s center of mass. In the proposed SfS method, an optimal sampling algorithm is outlined which extracts only the salient pixels from each image to ensure shape accuracy while reducing computational cost. While SfS algorithms are useful in generating initial shapes of small bodies, there are limitations. Some shape morphologies (e.g. craters) cannot be observed on the silhouettes and thus aren’t accurately captured. While SPC can be used to refine the shape, we propose a new algorithm called stereothermoclinometry (STC) which uses measured temperatures from infrared images in conjunction with predicted temperatures from a thermophysical model (TPM) to refine the shape of an object. As direct surface temperature measurements are used, STC is able to characterize craters and other such concave shape morphologies. All algorithms presented in this thesis are tested on simulated infrared images and sensitivity analysis of relevant parameters is conducted to comprehensively characterize the algorithms. In addition to simulated data, some of the algorithms were tested on real infrared imagery when such data was available and applicable to the algorithm.

These algorithms enable more computationally efficient and illumination robust characterization of the spin pole axis and shape of small bodies. This enhances mission autonomy by reducing the need for ground teams for processing. More autonomous missions reduce overall costs and can allow for more bodies to be visited. By doing so, comprehensive characterization of more bodies is possible which leads to improved knowledge of planetary formation, more informed planetary defense capabilities, and may be the key to sustained exploration of the solar system through enabling in-situ resource utilization.