Lunar Surface Change Detection

点击翻译按钮获取中文摘要

The objectives for this project are to refine the current scaling laws for crater formation, and luminous efficiency calculations for lunar impacts. In doing so, the nature of the data collected will allow the calculation of lunar impactor flux. Currently, scaling laws for crater formation are approximations, and no consensus has been reached on which of the scaling laws is most accurate. This is partly due to the unknown factor in which other parameters play a role; current scaling laws only take into consideration the kinetic energy, density, and the surface density. This neglects parameters such as the impacting angle and the luminous efficiency of the impacts, which are both known to have an effect on the total kinetic energy of the impact. In order to account for these parameters, research will be performed into how they effect the formation of lunar impact craters. To perform this research both the impact flash needs to be observed, and the resultant crater detected for a large number of impact events - this research is therefore split into two main sections; Observation, and Detection.To observe enough impacts for a large data set an extensive observation campaign will be undertaken, and will be implemented in such a way as to eliminate biases that other observation campaigns do not account for. Observation times will be noted to provide a comprehensive analysis of first of their kind time-based analyses. Further to this, observations will be made during all lunar phases, eliminating biases intrinsic to observing only the lunar night side. This is not trivial however as the detection of lunar flashes against the lunar day side requires the development of new observing methods. The development of these techniques form a large part of this project; extensive testing will go into the viability of detecting flashes on the lunar day side in thermal and near infrared, and visible light, as well as exploring other methods such as ejecta cloud detection. Detection of the resultant craters will form the second half of the project. While the PyNAPLE software does a great deal towards automating crater detection, it only acts as an automated processing pipeline with a human-in-the-loop approach to crater identification and error prevention. As part of this research time will be dedicated to the development of a more robust and sophisticated method of image matching, and the development of a neural network crater identification component. This is important to the research for two reasons; it is time consuming to view each final (60000x8000) image to identify craters, and misidentification of craters is easy. Automation will cut down on both the time needed to locate the craters - essential when dealing with a large data set - and lower the number of errant data generated by human evaluation. Further to this, the implementation of a neural network component could identify trends in the data not easily detectable by a human. This research is important for many reasons. By collecting a large, unbiased observational data set, more accurate impact flux calculations can be performed, along with distribution analyses. These are valuable to both Earth based and Lunar infrastructure; the Earth-Moon system shares a meteoroid environment, and the calculation of the lunar impactor flux also calculates the Earth impactor flux. This impactor flux has important implications for satellites and orbital infrastructure, as understanding when greater fluxes can occur allows safeguarding measures to be taken, such as placing satellites into safe modes, or temporarily altering their pointing to minimise the surface area with respect to the incoming meteoroids. The distribution of meteoroid impacts on the lunar surface is of high value to the development of lunar infrastructure - minimising the risk of damage from meteoroid impacts is vital for the longevity of infrastructure and safeguarding the human presence within.