The SuperCOSMOS Sky Survey. Paper II: Image detection, parameterisation, classification and photometry

Nigel Hambly,Mike Irwin,Harvey MacGillivray

DOI: https://doi.org/10.1111/j.1365-2966.2001.04661.x

2001-08-17

Abstract:In this, the second in a series of three papers concerning the SuperCOSMOS Sky Survey, we describe the methods for image detection, parameterisation, classification and photometry. We demonstrate the internal and external accuracy of our object parameters. Using examples from the first release of data, the South Galactic Cap survey, we show that our image detection completeness is close to 100% to within 1.5 mag of the nominal plate limits. We show that for the Bj survey data, the image classification is externally > 99% reliable to Bj = 19.5. Internally, the image classification is reliable at a level of > 90% to Bj=21, R=19. The photometric accuracy of our data is typically 0.3 mag with respect to external data for m > 15. Internally, the relative photometric accuracy in restricted position and magnitude ranges can be as accurate as 5% for well exposed stellar images. Colours (B-R or R-I) are externally accurate to 0.07 mag at Bj = 16.5 rising to 0.16 mag at Bj = 20.

Astrophysics

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to develop and describe the methods of image detection, parameterization, classification and photometry in the SuperCOSMOS Sky Survey (SSS). Specifically, the authors detail how to detect celestial body images from scanned Schmidt photographic plates and perform parameterization, classification and photometric calibration on these images. The following are the main issues involved in the paper:

1. **Completeness of image detection**: Ensure that as many celestial body images as possible are detected within a given magnitude range. For example, the paper mentions that the completeness of image detection in the BJ band is close to 100% until the magnitude is approximately \( B_J \sim 19.5 \).

2. **Accuracy of image classification**: Ensure the reliability of image classification. For example, for data in the BJ band, the external classification reliability is greater than 99% up to \( B_J \sim 19.5 \), and the internal classification reliability is greater than 90% up to \( B_J \sim 21 \).

3. **Photometric accuracy**: Ensure the accuracy of photometric data. For example, for celestial bodies with magnitudes greater than 15, the photometric accuracy is typically approximately 0.3 magnitudes. For well - exposed stellar images, within a specific position and magnitude range, the relative photometric accuracy can reach approximately 5%.

4. **Accuracy of color measurement**: Ensure the accuracy of color measurement. For example, at \( B_J \sim 16.5 \), the standard deviation of color measurements for \( B - R \) and \( R - I \) is 0.07 magnitudes; at \( B_J \sim 20 \), the standard deviation rises to 0.16 magnitudes.

### Formula Examples

- **Transmission and intensity calibration**:
\[
\gamma \log_{10} I \propto D \propto \log_{10}(1/T)
\]
where \(\gamma\) is the slope of the response curve in the \(D-\log_{10}I\) plane.

- **Threshold setting**:
\[
I_{\text{THRESH}} = \left(1 + \frac{\text{PCUT}}{100}\right) \times I_{\text{SKY}}
\]
where \(\text{PCUT} = 100 \times \frac{2.3 \sigma_{\text{SKY0}}}{I_{\text{SKY0}}}\), and \(I_{\text{SKY}}\) is the local sky estimate obtained by bilinear interpolation based on the filtered sky background map.

- **Contour area profile**:
\[
I_{\text{AP}_n} = I_{\text{THRESH}} + 2^{n - 2} \times \sigma_{\text{SKY0}}, \quad n = 1, 2, \ldots, 8
\]

Through these methods, the paper demonstrates the high precision and reliability of SSS data in image detection, classification and photometry, and verifies its consistency with other external data sources.