Fast, computer-assisted detection of dust and debris impact craters on Stardust interstellar foils

The NASA Stardust Interstellar Dust collection provides our current best sample set for direct laboratory analysis of dust grains from the contemporary interstellar dust stream. While a handful of likely interstellar dust grains were identified within the silica aerogel collection media, interstellar dust also impacted Al foils covering the collector frame. Locating these rare impacts requires labor-intensive collection and examination of tens of thousands of high-resolution SEM images. Here, we implement a Python-based algorithm to dramatically reduce the human time investment needed to locate impact craters. The algorithm employs a circular Hough transform to identify circular features in the foil images, followed by several tests to detect characteristic morphological features of impact craters-a dark center and a bright rim, with inclusion of multi-core processing capabilities to significantly increase processing speed. For most data sets, the code produced a pool of potential crater candidates in 1-5% of the input images, producing a more manageable subset of images for a human expert to review. We used this code to locate 31 impact craters across 12 Stardust interstellar foils, 25 of which were located on three adjacent foils, I1008W,1, I1009N,1, and I1020W,1. Many impacts on these foils formed shallow, oblique craters, with residue compositions consistent with solar cell glass and orientations consistent with debris ejected from the spacecraft solar cells. The code can be integrated into future searches for Stardust interstellar grain impacts and can be implemented as a general utility for dust impact studies on spacecraft materials.