Results from simulated data: (1) one of the ten simulations with a detection density of 27.59 × 10⁻⁴ det. px⁻¹. The identified tracks are shown as gradient lines from red to yellow. The background image serves merely as a visual aid, showing the rough locations of detections (it is not a stack of images created to run the detection algorithm on; instead the detections were simulated first and the background image was created retroactively). (2) the combined miss-rate distribution from the ten simulations with a detection density of 27.59 × 10⁻⁴ det. px⁻¹. (3) the velocity distribution of the same track population, showing a clear tendency of the algorithm to create more fast spurious tracks, especially when compared to velocity distributions from real data (Fig. 17.2). The probability density functions were created with Gaussian kernel density estimation.