Lunar Surface Change Detection

该项目的目标是完善目前陨石坑形成的定标定律，以及月球撞击的光效计算。在这样做的时候，收集的数据的性质将允许计算月球撞击通量。目前，陨石坑形成的标度律都是近似值，对于哪一种标度律最准确还没有达成共识。这在一定程度上是由于其他参数起作用的未知因素；目前的标度定律只考虑了动能、密度和表面密度。这忽略了诸如撞击角度和撞击的发光效率等参数，这两个参数都已知对撞击的总动能有影响。为了解释这些参数，将对它们如何影响月球撞击坑的形成进行研究。为了进行这项研究，既需要观察撞击闪光，又需要为大量撞击事件检测到由此产生的陨石坑--因此，本研究分为两个主要部分：观察和检测。为了观察到足够多的撞击，将进行广泛的观察活动，并将以消除其他观察活动无法解释的偏差的方式实施。将记录观测时间，以提供对第一类基于时间的分析的全面分析。此外，将在所有月相进行观测，消除仅观测月球夜侧固有的偏差。然而，这并不是一件微不足道的事，因为探测到月面的月光需要发展新的观测方法。这些技术的开发构成了该项目的一大部分；将进行广泛的测试，以检测在热和近红外以及可见光中探测月面闪光的可行性，以及探索其他方法，如喷发云探测。对产生的陨石坑的探测将构成该项目的后半部分。虽然PyNAPLE软件在自动检测陨石坑方面做了很大的工作，但它只是作为一个自动化处理管道，采用人在回路的方法来识别陨石坑和防止错误。作为这项研究的一部分，时间将专门用于开发更稳健和复杂的图像匹配方法，以及开发神经网络陨石坑识别组件。这对研究很重要，有两个原因；查看每一张最终的(60000x8000)图像来识别陨石坑是很耗时的，而且很容易对陨石坑进行错误识别。自动化将缩短定位陨石坑所需的时间--在处理大型数据集时至关重要--并减少由人工评估产生的错误数据的数量。此外，神经网络组件的实现可以识别数据中不易被人类检测到的趋势。这项研究之所以重要，原因很多。通过收集大量的、无偏的观测数据集，可以进行更准确的撞击通量计算以及分布分析。这些对地基和月球基础设施都很有价值；地月系统共享流星体环境，月球撞击通量的计算也计算地球撞击通量。这种撞击通量对卫星和轨道基础设施具有重要影响，因为了解何时可能发生更大的通量有助于采取保障措施，例如将卫星置于安全模式，或暂时改变其指向，以最大限度地减少相对于到来的流星体的表面积。流星体撞击月球表面的分布对月球基础设施的发展具有很高的价值--将流星体撞击造成的损害风险降至最低，对于基础设施的寿命和保护人类在月球内部的存在至关重要。

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.