Ron Brecher. AstroCATS May 3-4, 2014

Transcription

1 Ron Brecher AstroCATS May 3-4, 2014 Observing since 1998 Imaging since 2006 Current imaging setup: Camera: SBIG STL-11000M with L, R, G, B and H-alpha filters Telescopes: 10 f/3.6 (or f/6.8) ASA reflector; 80 mm f/6 refractor Mount: Paramount MX Software: TheSkyX, MaximDL, FocusMax, CCDCommander, PixInsight Reach me at rbrecher@rogers.com 1

2 Part 1 Results and Part 2 Workflow Overview of software options Advanced Image Processing Steps Using Masks Deconvolution Multiscale Processing Noise Reduction Morphological Transformation RGB Image of NGC4565 was processed in steps: Crop Gradient removal Colour balance Stretch Adjust Contrast and Saturation 2

3 The following techniques will be used in Part 2: Crop Gradient removal Colour balance Deconvolution Stretch Noise Reduction Stretch Multiscale Processing Morphological Transformation Adjust Contrast and Saturation 3

4 PixInsight (PC, Mac, Linux) Designed for astrophotography Documentation improving, but still limited Excellent user group Frequent significant updates Photoshop (PC, Mac) Powerful, but not specific to astrophotography Expensive Layers allows 3D image processing ImagesPlus, Nebulosity and MaximDL (PC) Integrated acquisition, reduction and processing software Masks used to selectively modify portions of an image Deconvolution a method of recovering detail lost to atmospheric blurring Noise reduction get smoother images without losing significant detail Multiscale processing a technique to reveal detail in the core of nebulae and galaxies Morphological transformation shrink stars; fix elongated stars 4

5 Mask protects parts of an image during processing Only parts visible through the mask are changed White reveals, black conceals Masks are black-and-white images: Star Mask Stars only Range mask selected brightness range Luminance mask variable masking depending on luminance of an image region Masks can be inverted or blended White reveals Black conceals Source Image Star Mask Range Mask Luminance Mask 5

6 Pure signal gets degraded convolved by imperfect seeing and optics Mathematical Point Spread Function (PSF) can be used to back out the degradation and restore detail Gaussian PSF assumed, or can be measured directly from stars in the image many iterations applied (30-50 typical) masking and deringing support to prevent artifacts Create support files Luminance mask to protect background De-ringing support to protect bright stars PSF Image file Adjust Deconvolution settings by experimenting Apply to (masked) image 6

7 Deconvolution Recovers detail in high dynamic range areas of an image Bright nebulae Galaxy cores Mask used to protect stars (avoid pinholes ) 6 wavelet layers 6 and 4 wavelet layers Original 7

8 Use Local Histogram Equalization tool to recover contrast 6 and 4 wavelet layers Local Histogram Equalization 8

9 Make a star mask to protect bright stars Set parameters and apply HDR Multiscale Transform tool ( Wavelets ) RB uses 6 then 4 layers most often Make a luminance or range mask to protect dark areas Apply Local Histogram Equalization Use a low strength setting 9

10 Multiscale Processing Noise appears as a grainy texture in images Chrominance (colour) and luminance (brightness) noise are different Eye/brain notices detail mostly in the luminance Colour noise can be managed more aggressively (often without a mask) Noise occurs at multiple scales Run NR multiple times using different scales Different masks are needed for different scales 10

11 PixInsight has many noise reduction options RB uses: SCNR for chrominance great for reducing Green cast or green pixels ACDNR or TGVNoise separate luminance and chrominance noise reduction built-in adjustable masks can be used together with usergenerated masks Use Histogram Transformation to reset the black point after noise reduction Noise Reduction 11

12 Morphology = shape Transformation = change Morphological Transformation = Shape Change Technique for modifying star shapes Fix elongated stars Shrink stars bright stars to make them less prominent PixInsight Morphological Transformation process Need a mask to protect non-stellar parts of image Background and structures must be protected Adjust shape of morhphological transformation using editing tool Adjust intensity to taste 12

13 Morphological Transformation There are powerful tools and techniques to tease out detail from your images Get comfortable with masks Learn pixel math to make complex masks Each process contributes a little Use them in the right order Experiment Back up your data first! Seek resources on the internet 13

14 14