Lunar Surface Change Detection

该项目的目标是完善目前陨石坑形成的比例规律，以及月球撞击的发光效率计算。在这样做时，所收集数据的性质将使计算月球撞击物通量成为可能。目前，火山口形成的标度律都是近似值，对于哪一种标度律是最准确的还没有达成共识。这部分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和撞击的发光效率等参数，这些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，也需要对大量撞击事件产生的陨石坑进行检测——因此，这项研究分为两个主要部分；观察和发现。为了观察大量数据集的足够影响，将开展广泛的观察活动，并将以消除其他观察活动无法解释的偏见的方式实施。观测时间将被记录下来，以提供首次基于时间的分析的全面分析。此外，观测将在所有月相期间进行，消除了只观测月球夜侧的固有偏差。然而，这并不是微不足道的，因为探测月球白天一侧的月球闪光需要发展新的观测方法。这些技术的发展构成了这个项目的很大一部分；广泛的测试将用于探测月球白天一侧的热、近红外和可见光闪光的可行性，以及探索其他方法，如喷射云探测。探测由此产生的陨石坑将构成该项目的后半部分。虽然PyNAPLE软件在自动陨石坑检测方面做了很多工作，但它只作为一个自动化处理管道，采用人在循环的方法来识别陨石坑和预防错误。作为这项研究的一部分，时间将致力于开发一个更强大和复杂的图像匹配方法，并开发一个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非常耗时的，而且很容易被错误识别。自动化将减少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容易被人类检测到的趋势。这项研究很重要，原因有很多。通过收集大量无偏差的观测数据集，可以进行更准确的冲击通量计算以及分布分析。这些对地球和月球的基础设施都很有价值；地月系统共享一个流星体环境，月球撞击体通量的计算也可以计算地球撞击体通量。这种撞击体通量对卫星和轨道基础设施具有重要意义，因为了解何时可能出现更大的通量，可以采取保障措施，例如将卫星置于安全模式，或暂时改变其指向，以尽量减少相对于来袭的流星体的表面积。流星体撞击在月球表面的分布对月球基础设施的发展具有很高的价值——最大限度地减少流星体撞击造成的破坏风险对基础设施的寿命和保护人类在月球上的存在至关重要。

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.