Next: MakeCatalog, Previous: ImageStatistics, Up: Image analysis [Contents][Index]
Once raw data have gone through the initial reduction process (through the programs in Image manipulation). We are ready to derive scientific results out of them. Unfortunately in most cases, the scientifically interesting targets are deeply drowned in a sea of noise. NoiseChisel is Gnuastro’s tool to detect signal in noise. Infact, NoiseChisel was the motivation behind creating Gnuastro and has a full journal article79 devoted to its techniques. Following the explanations for the options in NoiseChisel options should also give you a relatively good idea of the steps. Currently the paper does a very thorough job at explaining the concepts and methods of NoiseChisel with abundant demonstrations for each step. However, the paper cannot undergo any futher updates, so as the development of NoiseChisel evolves, this section will grow.
Detection is the process of separating signal from noise. In other words, after detection is complete, one set of data elements (pixels in an image) will be distinguished as signal and another set of the data elements will be noise. Segmentation is the process of labeling the detected pixels into possibly multiple components (objects). For example when two galaxies lie sufficiently close to each other to be detected as one object.
NoiseChisel was the first software to make use of a noise-based concept to detection and segmentation. In this method, instead of emphasizing on the signal and trying to guess the properties of the to-be-detected targets prior to detection (for example assuming that it is an ellipse), the emphasis is put on the noise in the image and it imposes statistically negligible requirements on the signal. The name of NoiseChisel is derived from the first thing it does after thresholding the image: to erode it. In mathematical morphology, erosion on pixels can be pictured as carving off boundary pixels. So what NoiseChisel does is similar to what a wood chisel or stone chisel does. It is just not a hardware but software.
| • Invoking astnoisechisel: | Options and arguments for NoiseChisel. |
It is currently under production by the Astrophysical Journal Supplement Series. It can also be read in arXiv: http://arxiv.org/abs/1505.01664.
Next: MakeCatalog, Previous: ImageStatistics, Up: Image analysis [Contents][Index]
Read in other formats.
GNU Astronomy Utilities manual, November 2015.