1:45 PM - 3:15 PM
[O11-P124] Development of an automatic observation device for meteors and analysis of meteor showers
Keywords:Meteor, Comet, Machine learning
Research Background and Objectives
Since 1953, the school's Astronomy and Meteorology Club has conducted visual observations of the Perseid meteor shower. However, the burden of nightly, all-night observations led to the development of an automatic meteor observation system using video and radio waves. The video observation identifies various meteor orbits, showing where meteors originate and how they move relative to Earth. The radio observation enables 3D simulations of reflection regions and supports orbit identification through head echo observations.
Research Methods
Figure 1 summarizes the two main observation methods—video and radio—and the insights gained from each.
Development of a Video Observation Device and Detection Program
We built a wide-area video observation device using five low-cost security cameras (Figure 2), and we are testing an automatic sunshade for camera protection. For meteor detection, we developed a machine learning program (Figure 3) that detects meteors more accurately than human observation.
Development of a Radio Wave Observation Device and Detection Program
We installed antennas on the school roof to receive radio waves from Fukui, Toyokawa, and others, applying the traditional HRO method. Using the free software MROFFT, we converted echo signals into images, then used a machine learning-based program to detect meteors automatically (Figure 4).
Orbit Identification of the Perseid and Phoenix Meteor Shower
Using observation data and the SonotaCo network, we identified meteor shower orbits (Figure 5). We successfully determined 3D ground and diurnal orbits for the Perseids. The Phoenix meteor shower, a rarely observed shower since being reported by the Japanese Antarctic expedition in 1956, was detected in November last year during a predicted dust trail approach. The diurnal orbit matched that of the proposed parent body, suggesting it was indeed the Phoenix meteor shower.
3D Simulation of Radio Reflective Area
To align radio and video observations, we simulated the radio wave reflection region, which is a rotating ellipsoid (Figure 6). While earlier studies calculated this in 2D using Excel, our original Python program produced 3D outputs (Figure 7), improving data matching. We successfully mapped the meteor's path and its associated reflection region using data from two radio stations and video (Figure 8), showing that echoes occur when meteors pass through the reflection region.
Meteor Orbit Identification in Head Echo Observations
Identifying meteor orbits from radio echo images is challenging. Recent studies, such as those by Kazuhiro Suzuki, have explored using Doppler-shifted head echoes to determine orbits. We applied a method based on solving simultaneous equations and found strong agreement between predicted and actual orbits (Figure 11), suggesting this technique is effective for linking radio and video data.
Summary and Future Prospects
We successfully developed a real-time meteor detection system using video and radio technologies enhanced by machine learning. We identified 3D orbits from video data and visualized 3D radio reflection regions. Our new method for interpreting head echoes also proved useful for orbit identification. Continued data collection is necessary to further verify the accuracy of these techniques.
References
1)Astronomy and Meteorology Club (2024) “Production of Meteor Observatory ‘TenGu’ ver2.0” “13th High School and College of Technology Weather Observation Instrument Contest
2)ATOM. “ATOMCam2
3)International Project for Radio Observation of Meteors
4) Yosuke Utsumi (2002). Calculation of Observation Area of HRO Meteor Radar.
5) Ogawa Hiroshi (2003). Discussion of Reflection Region in Meteor Radio Observations.
6) Hasegawa Hitoshi (2024). Meteor Detection and Location Measurement Detected by ATOMCam,” ‘Meteor Radio Observation Report Meeting 2024’.
7)SonotaCo.com
8)Saito K., Nagasawa T. (1984). Meteors I. Observations in Practice, Koseisha-Koseikaku.
9)Kazuhiro Suzuki (2024). Estimation of Fallen Satellite Paths by KRO.
Since 1953, the school's Astronomy and Meteorology Club has conducted visual observations of the Perseid meteor shower. However, the burden of nightly, all-night observations led to the development of an automatic meteor observation system using video and radio waves. The video observation identifies various meteor orbits, showing where meteors originate and how they move relative to Earth. The radio observation enables 3D simulations of reflection regions and supports orbit identification through head echo observations.
Research Methods
Figure 1 summarizes the two main observation methods—video and radio—and the insights gained from each.
Development of a Video Observation Device and Detection Program
We built a wide-area video observation device using five low-cost security cameras (Figure 2), and we are testing an automatic sunshade for camera protection. For meteor detection, we developed a machine learning program (Figure 3) that detects meteors more accurately than human observation.
Development of a Radio Wave Observation Device and Detection Program
We installed antennas on the school roof to receive radio waves from Fukui, Toyokawa, and others, applying the traditional HRO method. Using the free software MROFFT, we converted echo signals into images, then used a machine learning-based program to detect meteors automatically (Figure 4).
Orbit Identification of the Perseid and Phoenix Meteor Shower
Using observation data and the SonotaCo network, we identified meteor shower orbits (Figure 5). We successfully determined 3D ground and diurnal orbits for the Perseids. The Phoenix meteor shower, a rarely observed shower since being reported by the Japanese Antarctic expedition in 1956, was detected in November last year during a predicted dust trail approach. The diurnal orbit matched that of the proposed parent body, suggesting it was indeed the Phoenix meteor shower.
3D Simulation of Radio Reflective Area
To align radio and video observations, we simulated the radio wave reflection region, which is a rotating ellipsoid (Figure 6). While earlier studies calculated this in 2D using Excel, our original Python program produced 3D outputs (Figure 7), improving data matching. We successfully mapped the meteor's path and its associated reflection region using data from two radio stations and video (Figure 8), showing that echoes occur when meteors pass through the reflection region.
Meteor Orbit Identification in Head Echo Observations
Identifying meteor orbits from radio echo images is challenging. Recent studies, such as those by Kazuhiro Suzuki, have explored using Doppler-shifted head echoes to determine orbits. We applied a method based on solving simultaneous equations and found strong agreement between predicted and actual orbits (Figure 11), suggesting this technique is effective for linking radio and video data.
Summary and Future Prospects
We successfully developed a real-time meteor detection system using video and radio technologies enhanced by machine learning. We identified 3D orbits from video data and visualized 3D radio reflection regions. Our new method for interpreting head echoes also proved useful for orbit identification. Continued data collection is necessary to further verify the accuracy of these techniques.
References
1)Astronomy and Meteorology Club (2024) “Production of Meteor Observatory ‘TenGu’ ver2.0” “13th High School and College of Technology Weather Observation Instrument Contest
2)ATOM. “ATOMCam2
3)International Project for Radio Observation of Meteors
4) Yosuke Utsumi (2002). Calculation of Observation Area of HRO Meteor Radar.
5) Ogawa Hiroshi (2003). Discussion of Reflection Region in Meteor Radio Observations.
6) Hasegawa Hitoshi (2024). Meteor Detection and Location Measurement Detected by ATOMCam,” ‘Meteor Radio Observation Report Meeting 2024’.
7)SonotaCo.com
8)Saito K., Nagasawa T. (1984). Meteors I. Observations in Practice, Koseisha-Koseikaku.
9)Kazuhiro Suzuki (2024). Estimation of Fallen Satellite Paths by KRO.