[Paper Review] The DASCH Data Processing Pipeline and Multiple Exposure Plate Processing

E. Los, J. E. Grindlay+3 more|arXiv (Cornell University)|Feb 23, 2011

Astronomical Observations and Instrumentation1 references3 citations

TL;DR

This paper presents a specialized data processing pipeline for the DASCH project that enables accurate photometry and astrometry of 525,000 historical astronomical plates, with a focus on multiple-exposure plates. It introduces the Pickering Wedge Filter and Multiple Exposure Loop to detect ghost objects and resolve overlapping exposures, improving photometric accuracy and quality control by identifying problematic plates and misaligned optics.

ABSTRACT

Digital Access to a Sky Century @ Harvard (DASCH) is a project to digitize the collection of approximately 525,000 astronomical plates held at the Harvard College Observatory. This paper presents an overview of the DASCH data processing pipeline, with special emphasis on the processing of multiple-exposure plates. Such plates extended the dynamic range of photograph emulsions and improved photometric accuracy by minimizing variations in plate development procedures. Two approaches are explored in this paper: The repetitive use of astrometry.net (Lang et al. 2010) and local correlation searches. Both procedures have yielded additional quality control checks useful to the pipeline.

Motivation & Objective

• To enable high-accuracy photometry and astrometry for 525,000 historical astronomical plates digitized by the DASCH project.
• To address the challenge of multiple-exposure plates that degrade photometric precision due to overlapping sources and development variations.
• To develop automated methods for detecting ghost objects from Pickering Wedge and grating plates, improving data quality and pipeline reliability.
• To implement a robust data processing pipeline capable of handling plates with complex exposure histories and optical artifacts.
• To provide quality control by identifying plates with poor astrometric solutions or misaligned optics through repeated astrometric solutions.

Proposed method

• The pipeline uses a multi-stage approach: plate transcription, mosaic generation from 60 tiles, and WCS fitting via astrometry.net, WCStools, and polynomial fitting.
• The Pickering Wedge Filter identifies ghost objects by analyzing spatial correlation patterns around the 300 brightest stars, detecting characteristic double peaks in object distribution.
• The Multiple Exposure Loop iteratively removes catalog-matched objects and re-applies astrometry.net to detect unresolved or blended exposures, flagging ambiguous cases.
• Photometric calibration uses annular bins and LOWESS curve fitting to link instrumental magnitudes to catalog magnitudes, with local 50×50 grid corrections for emulsion and sky variations.
• Defect filtering distinguishes real stars from emulsion defects by comparing PSF characteristics between matched and unmatched sources.
• A companion algorithm flags plates with inconsistent solutions or multiple exposure blends, preventing infinite loops and improving data integrity.

Experimental results

Research questions

• RQ1 How can multiple-exposure plates with overlapping sources be reliably identified and processed to preserve photometric accuracy?
• RQ2 What automated methods can detect ghost objects from Pickering Wedge and grating plates within historical photographic data?
• RQ3 How effective is the combination of astrometry.net and local correlation searches in improving pipeline quality control?
• RQ4 To what extent do repeated astrometric solutions reveal misaligned optics or poor plate conditions in the DASCH dataset?
• RQ5 Can the Pickering Wedge Filter and Multiple Exposure Loop detect and flag plates with poor photometric quality due to optical artifacts?

Key findings

• The pipeline successfully detected 31% of double exposures and 18% of triple exposures in a sample of 87 fully transcribed multiple-exposure plates.
• The Pickering Wedge Filter flagged 857 out of 11,461 scanned plates as containing ghost objects, indicating widespread use of this technique.
• The Multiple Exposure Loop identified 292 plates with poor photometric quality due to grating plates or bad astrometric solutions, enhancing data quality control.
• The method detected potential optics misalignment in the 'dsy' telescope series through inconsistent astrometric solutions across exposures.
• The use of local correlation searches and repeated astrometry.net runs provided additional quality control checks beyond standard pipeline processing.
• The pipeline achieved reliable photometric calibration using annular bins and LOWESS fitting, with local 50×50 grid corrections reducing sky and emulsion variation effects.