1:45 PM - 3:15 PM
[O08-P94] Development of an automatic lunar crater measurement program using python
Many people have attempted to determine the depth of lunar craters from photographs, but most of these attempts were manual, time-consuming, and often inaccurate due to the need to measure numerical values with the human eye. I have developed a program using python to automatically measure the depth of lunar craters and height differences of various lunar landforms from lunar photographs, in order to automate these tasks and make it easier for many people to become involved in lunar research. This makes it possible to measure height differences more easily and accurately than conventional manual measurement methods. The program allows the user to enter a photograph of the Moon and the date and time of the photograph, and then simply use the mouse to specify the length of the shadow at the location where the height difference is to be determined. In creating the program, web scraping of the National Astronomical Observatory of Japan's website, which displays the axis of the moon's rotation in its chronological table, was incorporated into the program to eliminate blurring caused by the moon's libration and to achieve accurate measurements that are not easily affected by the moon's age. The program is capable of measuring the height of various lunar landforms, not only craters, as long as there is a photograph of the shadow of the object whose height is to be measured and the length of the shadow is known.
The following is an overview of the program. First, the program requires photo data and the time when the photo was taken (year, month, day, hour, minute, and second). This time is input to the website of the National Astronomical Observatory of Japan (NAOJ), which is called by web scraping, and the positional relationship of the moon, the sun, and the earth is output. Next, the position and relative size of the moon on the photo are measured by Hough transform from the input photo. Then, the user displays the photo and specifies the length of the target shadow using the mouse. The specified shadow length is converted from an apparent length to an actual length using the data obtained from web scraping. Finally, the program ends with a trigonometric function that outputs the height of the object using all the data obtained earlier.
How accurate is this program? The accuracy of the output height of the object depends greatly on how accurately the length of the shadow can be specified. Also, the resolution of the photograph has a large effect on the accuracy, since the length of the shadow is specified in pixels. In addition, the length of the shadow per height of the object, i.e., the angle of incident light and the lunar coordinates, also affect the accuracy. Therefore, it is impossible to discuss the accuracy of this program in general, since the values may vary depending on many factors. However, if the photograph is sufficiently shadowed and the length of the shadow is more than 10 pixels on the photograph, the error will not exceed 1km.
About using this program for research. I am using this program to study the relationship between the depth and width of craters. Also, as mentioned above, this program can be used for terrain other than craters as long as the length of the shadow is known, so I am looking for ways to use it to study grooves and mountainous terrain on the Moon.
What is the outlook for the future? We plan to create a GUI so that various people can use this program easily. We are considering adding a program to automate the process of measuring length, to automate the adjustment of Hough transform parameters, and to automatically crop and process images to be used. In addition, since the program is currently running on Visual Studio Code provided , we plan to create an independent program that can be used by the general public.
Finally, we would like to express our deepest gratitude to Professor Kazuhito Dobashi of Tokyo Gakugei University for his great cooperation in conducting this research.
The following is an overview of the program. First, the program requires photo data and the time when the photo was taken (year, month, day, hour, minute, and second). This time is input to the website of the National Astronomical Observatory of Japan (NAOJ), which is called by web scraping, and the positional relationship of the moon, the sun, and the earth is output. Next, the position and relative size of the moon on the photo are measured by Hough transform from the input photo. Then, the user displays the photo and specifies the length of the target shadow using the mouse. The specified shadow length is converted from an apparent length to an actual length using the data obtained from web scraping. Finally, the program ends with a trigonometric function that outputs the height of the object using all the data obtained earlier.
How accurate is this program? The accuracy of the output height of the object depends greatly on how accurately the length of the shadow can be specified. Also, the resolution of the photograph has a large effect on the accuracy, since the length of the shadow is specified in pixels. In addition, the length of the shadow per height of the object, i.e., the angle of incident light and the lunar coordinates, also affect the accuracy. Therefore, it is impossible to discuss the accuracy of this program in general, since the values may vary depending on many factors. However, if the photograph is sufficiently shadowed and the length of the shadow is more than 10 pixels on the photograph, the error will not exceed 1km.
About using this program for research. I am using this program to study the relationship between the depth and width of craters. Also, as mentioned above, this program can be used for terrain other than craters as long as the length of the shadow is known, so I am looking for ways to use it to study grooves and mountainous terrain on the Moon.
What is the outlook for the future? We plan to create a GUI so that various people can use this program easily. We are considering adding a program to automate the process of measuring length, to automate the adjustment of Hough transform parameters, and to automatically crop and process images to be used. In addition, since the program is currently running on Visual Studio Code provided , we plan to create an independent program that can be used by the general public.
Finally, we would like to express our deepest gratitude to Professor Kazuhito Dobashi of Tokyo Gakugei University for his great cooperation in conducting this research.