III.D. Quality Checking

ISSA Explanatory Supplement
III. PROCESSING
D. Quality Checking


Chapter Contents

Table of Contents | Index | Previous Section | Next Section


  1. Pre-Production
  2. Production
  3. Post-Production
  4. Types of Anomalies
    1. Data Anomalies
    2. Processing Anomalies

D. Quality Checking

ISSA data were subjected to quality checks during the pre-production, production and post-production processing stages.

D.1 Pre-Production

Prior to image production, a machine-readable file of previously identified anomalous scans was compiled. The file contains start and stop times of the scans to clip completely from processing. The various sources from which this clipping information was culled include the SkyFlux images, telescope pointing anomalies discovered during production of the IRAS Faint Source Survey and scans identified through individual research efforts. A summary list is presented in Appendix C.

D.2 Production

During image production, the global basketweaver corrections were applied to each scan in a given field. There were several conditions, however, that could prevent the global corrections from being applied to a portion or all of a scan (Appendix D). In these cases either the scan was completely ignored in downstream processing or was turned over to the local destriper to derive a fit to the local background.

D.3 Post-Production

The final step of the quality checking was to inspect each image visually to identify any anomalous data not removed in the previous quality checking. Individual HCON images were examined to identify anomalous features. When found, the end points of the scan portion containing the anomaly were identified and that entire scan portion was removed. The main criterion for removal of an anomaly at this stage was that it not confirm in another HCON image for that field. Identified anomalies were removed from both the individual HCON images and the co-added images. Once an anomaly was removed, the field was reprocessed, creating a new set of individual HCON and co-added images which were again inspected to verify removal of the anomaly. In general, anomalies found during this process consisted of detections of nonconfirming space debris or showers of energetic particles. As mentioned in §III.C.4, comet tails and trails were not removed. The tail of comet Iras-Araki-Alcock is seen in fields 416 and 418. If an anomaly appeared faint or was not covered by another HCON it was generally not removed. Anomalies not seen in the co-added image or near a field boundary were not removed. To maintain consistency during a somewhat subjective process, one person, Gwen Johnson, performed the inspections and identifications of anomalies for all images during this post-production quality check. The amount of data removed during the post-production quality checking of the || > 20° sky is shown in Table III.D.1(a). The amount of data removed from the ISSA Reject Set is shown in Table III.D.1(b).

Table III.D.1(a) Amount of Data Removed in Anomaly Processing
(|| > 20°)
Wavelength
(µm)
% Data Removed
(||>50°)
% Data Removed
(||<50°)
120.180.51
250.300.31
600.140.24
1000.070.26

Table III.D.1(b) Amount of Data Removed in Anomaly Processing of the ISSA Reject Set (|| <20°)
Wavelength
(µm)
% Data Removed
120.34
250.30
600.27
1000.47

D.4 Types of Anomalies

An attempt was made to characterize the anomalies found by visual inspection. Anomalies fell into two main groups, data anomalies and processing anomalies, which are described below. Most of these anomalies were removed through the visual inspection process described in §III.D.3. All processing problems were corrected in the software except those that caused the improper handling of saturated data.

D.4.a Data Anomalies

Figure III.D.1 Distribution of Focal Plane Anomalies Plotted in Equatorial Coordinates
larger largest
-- Focal Plane and Partial Focal Plane Anomalies
All or a subset of the detectors in the focal plane jumped to a higher intensity for a time then fell back to approximately their original intensity. Both the rise and fall were fairly sharp. This was likely due to either a particle or paint flake in the near field of the telescope or by a shower of secondary energetic particles from the observing structure. Figure III.D.1 shows the distribution of focal plane anomalies.

Figure III.D.2 Distribution of Mini-Streak and Detector-Streak Anomalies Plotted in Equatorial Coordinates
larger largest

-- Detector Streaks/Ministreaks
One or a few detectors showed nonconfirming spikes or raised intensity. Generally the mini-streaks were due to orbital debris in the field of view, whereas detector streaks were due to calibration problems. The distribution of detector streaks and ministreaks is found in Figure III.D.2.

D.4.b Processing Anomalies

Figure III.D.3 Distribution of Local Destripe Anomalies for || > 50°, Plotted in Equatorial Coordinates
larger largest
-- Local Destriper
These anomalies were shown to appear only after the local destriper processing. They were caused by an error in the local destriper software that did not account for data gaps in the time-ordered detector data. A number of local destriper anomalies were left in these images because they were not bright enough to stand out visibly. This error was corrected prior to processing the || < 50° sky. Distribution of local destriper anomalies for the || > 50° sky is found in Figure III.D.3.

Figure III.D.4 Occurrences of Saturated Data for Entire Sky, Plotted in Equatorial Coordinates
larger largest

-- Saturated Detector Data
An error was found in the algorithm for handling saturated intensity values. This error affected the SkyFlux images as well as the entire set of ISSA images. The algorithm eliminated the wrong detector when saturation occurred. This resulted in the inclusion of saturated intensity values in making the images while erroneously eliminating some nonsaturated intensity values. Figure III.D.4 shows that the problem occurred mainly in the Galactic plane where 60 and 100 µm detectors saturate. Table III.D.2 provides a list of fields along with the number of occurrences in each field. The total number of occurrences throughout the mission is 6,289. Each occurrence reflects a single detector saturation. There may be several detectors saturated within a second of data. Assuming that on an average ten detectors saturate per second, the total number of occurrences is about <0.005% of the survey data.

Table III.D.2 ISSA and ISSA Reject Fields Affected by Saturated Data
Field #OccurrencesField #Occurrences
17 14 153 387
18 12 170 338
32 194 171 103
33 289 182 22
34 24 * 183 22
35 30 189 278
36 76 * 190 278
37 57 206 436
52 113 207 196
58 17 226 136
59 369 227 14
60 119 248 12
77 32 * 249 12
78 27 262 8
86 37 263 163
87 448 284 10
104 8 297 77
* 105 8 298 115
17 374 331 2
118 1603 360 6
119 30 361 4
137 8 390 9
* 138 8 391 59
152 27 407 6
*Overlapping area with adjacent field not included in total.


Chapter Contents

Table of Contents | Index | Previous Section | Next Section