Asteroid Discovered! | Scitech

Kritti Sharma and Kunal Deshmukh, two undergraduate students in their third and fourth year respectively, have made a discovery that has made the entire institute proud. They discovered the closest asteroid flyby of Earth ever recorded without impacting the planet! The discovery was made just hours after the flyby using the data from the Zwicky Transient Facility (ZTF) at Caltech.

Kunal and Kritti are a part of the IITB wing of GROWTH collaboration led by Prof. Varun Bhalerao from the Physics Department. Caltech leads the GROWTH collaboration which stands for Global Relay of Observatories Watching Transient Happen. It is an international scientific collaborative project in astronomy studying the physics of fast-changing events in the cosmos like supernovae, neutron stars or black-hole mergers and near-earth asteroids. Kunal and Kritti work on the objective of finding asteroids as part of the ZTF Solar System Working Group led by Caltech. This group has members from India, Taiwan, UK, and the US who collectively work on finding asteroids.     

A look into how the asteroid was detected:

The process begins in the Palomar Observatory, California, United States. Here, the Zwicky Transient Facility, a wide-field sky astronomical survey, uses a camera attached to the Samuel Oschin Telescope to survey the skies with the massive field of view of 47 square degrees. For comparison, the moon is about 0.2 square degrees when viewed from the earth. To capture such a wide field of view, the camera uses 16 CCDs (a CCD is an array of photosensitive capacitors) of 6144 X 6160 pixels each. Such wide-field facilities are designed to detect transient changes in the night sky—moving objects such as comets and asteroids, collisions between neutron stars, gamma-ray bursts and supernovae. These phenomena are referred to as transients.

Now, the asteroid detection process consists of three key steps.

1) Image capturing and processing – This step involves capturing a multitude of images, processing them into science images and vetting them using Machine Learning and Deep Learning algorithms to look for transients. The two classification problems here are real/bogus separation to separate astrophysical sources from artefacts and separation into different classes among the astrophysical sources. The challenge of real/bogus classification is the construction of a training set that is representative of nightly data across filters, sky location, and in the case of multi-CCD surveys like ZTF, possible variations between CCDs, as well as cross-talk. The real/bogus separation algorithms are trained using pictures of actual asteroid streaks, artificially injected streaks and false positives. This is done to remove biases which may creep into the classification during training and to ensure that no real transient goes unnoticed. The ML and DL algorithms eliminate a large number of images, which do not have any transients in them for sure. 

2) Scanning of images – This is where the team at IIT Bombay comes in. Experts are needed to inspect the shortlisted images manually and try to figure out whether it was a false positive. Close examination of these images involves looking for streaks, diffraction effects, reflections, reverse image subtraction errors, edges and cosmic rays. Other streaks may correspond to either satellites or asteroids. A satellite usually has a blurry end to the streak, whereas an asteroid has more well defined and clear endpoints. Such details are used to differentiate between satellites and asteroids. Thus the students at IIT Bombay manually scanned the data to find a near-earth asteroid which was the size of an SUV.

3) Verification through Minor Planet Center – MPC is a worldwide body responsible for verifying and cataloguing data about minor planets. This involves confirming their presence, calculating their orbits and their distance from the earth by astrometrical analysis. This third and final step confirms the discovery of an asteroid

0