CAMS PIPELINE

The Ancient Trails of Long-period comets

team

Researchers:

Surya Ambardar, Alfred Emmanuel, Julia Nguyen, Sahyadri Krishna, Chicheng Ren, Chad Roffey, Jesse Lash

Faculty:

Siddha Ganju, Meher Anand Kasam, Anirudh Koul, Peter Jenniskens

Posted by:

details

Program:

FDL US

Institution:

NASA Planetary Defense Coordination Office (PDCO), The Seti Institute

Year:

2017

Challenge area:

planetary defense

Need And Challenge:

  • Establish an end-to-end pipeline such that CAMS can exist independently and autonomously at Technology Readiness Level 9
  • Establish the state of the art model for detecting models with a novel augmentation strategy
  • Enable active learning in the established CAMS pipeline to automate insight generation and reduce manual workload. Includes expert-guided testing to verify all the low confidence meteors and retrain on unique data points to improve and localize training.
  • Provide standard guidelines on how to process the decade long meteor dataset
  • Upgraded web tool which enables global access to the meteors detected from the previous night with a new state-of-the-art user-centered design, an upgraded data ingestion pipeline to index incoming streaming data, and several new features like:
  • “export timeline” that allows scientists to highlight and export important trends, aiding in effective communication of ideas and results to a diverse audience, 
  • ability to zoom in or out of a meteor shower, thereby allowing the scientists to view showers that are hidden or overlaid by others, 
  • responsive latitude and longitude coordinates and 
  • constellation markings serving as location identifiers

Results

1.CAMS data used to measure the orbital period of a comet:

In April 2019, CAMS New Zealand station detected a brief outburst of 5 meteors from comet C/1907 G1 (Grigg-Mellish). The timing of the outburst was used to refine the comet orbit. Based on the 1907 observations of the comet orbit alone, valid solutions showed a strong correlation between node and orbital period due to an uncertain distance between comet and observer.

2.CAMS data used to predict the existence of a previously unknown comet by detecting its trail of crumbs:

In April 2020, the Phi Serpentid shower (International Astronomical Union Shower number 839), a shower in the Serpens constellation was traced back to the orbit of an unknown long-period comet

3.CAMS discovered new meteor showers

4.CAMS Monitoring of unusual activity meteor showers

2021: Brief outburst of Delta Mensid

2020: 

-A significant meteor activity from A-Carinid, an otherwise weak annual shower was detected by CAMS

-CAMS discovered meteors of chi Phoenicids, a new long-period comet

-Outburst of Ursids caused by the 1076 A.D. dust of comet 8P/Tuttle

-CAMS detected delta Mensid shower and identified a minor planet 2006 CS as the probable parent body

-CAMS detected rho Phoenicids

-CAMS picked early sightings of chi Cygnids in August, which resulted in speculation that the shower will return, since its last sighting in 2015. The shower indeed returned and was observed in detail in September

2019:

-CAMS detected an outburst of 15 Bootids, whose orbital elements resemble those of bright comet C/539 W1. Researchers also concluded that this meteor shower was caused by the same bright comet as was described in Histories of the Wars, an 553 A.D. book, and in 2019 the comet appeared to be on its way back.

-CAMS captured an outburst of June epsilon Ophiuchids, with parent body as periodic Jupiter Family comet 300P/Catalina

-CAMS detected an outburst of Phoenicids from comet Blanpain

-The alpha Monocerotid outburst ("the unicorn shower") was observed by the CAMS but its intensity was less than expected. 

2018:

-CAMS captured an outburst of Draconids and the October Camelopardalids

2017:

-The automation of the data processing allowed researchers to manage the camera stations better, which led to the detection of the highest number of meteors in a single night (including 3003 Geminids & 1154 sporadic meteors) in December 2017.

-Earth traveled through the 1-revolution dust trail of a long-period comet C/2015 D4 (Borisov). CAMS noted that "Only about once every 25 years is such an intermediate long-period comet discovered that passes close enough to Earth's orbit to have dust trail encounters. This one passed perihelion in 2014." CAMS South Africa network captured 167 meteors.

CAMS In The Media:

https://blogs.nvidia.com/blog/2020/08/20/cams-meteor-showers/ 

https://www.youtube.com/watch?v=GiZ7kyrwZGQ 

Method

ALCAMS: Active Learning for Camera Allsky Meteor Surveillance:

-Train a meteor classifier to generate predicted labels and associated confidence levels for our entire decade’s worth of data. 

-Using active learning, high-confidence predictions are treated as effectively labeled, while lower confidence predictions are sent to the scientist for manual confirmation. 

-Numerical improvements and continual monitoring for model and data drifts such as weather changes, smoke, and cloud patterns affect the view of the night sky and are regularly incorporated into the CAMS AI model. 

-The AI pipeline is automated and runs regular tests, production of benchmarks, and sends low confidence meteors for verification to scientists. Manual intervention can be triggered when needed, such as sending low confidence meteors for verification to CAMS scientists.

Status And Future Work

Paper: Under Review

Success of the project has encouraged expanding the CAMS network leading to a five-fold global expansion of the camera network, including first light in the Southern Hemisphere, and preparing for first light in the North-Eastern Hemisphere post which the project will achieve 100% global coverage

Our contributions also encourage the use and distribution of available data and models by making data globally and publicly accessible through daily generated maps of meteor shower activity posted at https://meteorshowers.seti.org/