can be used to detect particles in an image. In general, particle identification and detection is a process involving the following steps:
- First calculate a background mask if there are large areas in the image not to be analyzed. Using a background mask may significantly speed up the detection of particles.
- You have to create an image which specifically shows the interesting particles. This can be done by something as simple as the intensity of a single spectral line, or by more elaborate approaches such as specific classifiers which are selective for certain compounds.
- The particle image is then dichotomized by applying a threshold to the image value, i.e. all pixels showing values below the threshold are set to zero, all other pixels show a non-zero value.
- Next, the dichotomized image is subjected a particle recognition routine which detects neighboring pixels and combines them into individual particle areas.
- These areas are then processed, calculating their characteristic parameters (width, length, direction, etc) and entering the found parameters into a list of particles.
- Optionally you can enter quantitative information using the particle editor in order to be able to create a calibration model (for example, to estimate the constituents of a particle).
- The generated particle list can then be inspected and edited by the particle editor.
Dichotomization and particle recognition can be accomplished by using almost any type of image displayed in . Just right-click the particular image and select the command "Detect Particles" from the context menu. This will bring up a dialog showing the intensity distribution of the image data. You can then set a threshold which is used for the dichotomization when you click the button "Detect Particles".