Saturation And Value Modulation A New Method For

Saturation And Value Modulation: A New Method For Integrating Colour And Grey-Scale Imagery David Viljoen & Jeff Harris Geological Survey of Canada 1/30

Overview • Why integrate colour and grey-scale data? • Conventional methods and problems. • Saturation-Value-Modulation concept & Sat. Val. Mod Arc. GIS implementation. • Geoscience examples. 2/30

Integrating Colour & Grey-scale imagery Enhancing the visualization of a dataset. } DEM with shaded relief of DEM for Mt. Logan 3/30

Integrating Colour & Grey-scale imagery Visualizing the relationships between two datasets. } K-Th-U (RGB) Gamma ray spectrometer with Landsat TM 7 4/30

Layer transparency 20% 40% Tra nsp (Co lou Result is independent of layer order are ro ver 80% ncy sha of de dr Co lou elie f) r 60% Tra 80% ns pa (Sh ren ad ed cy rel of s ief ov had er co ed lou r) Layer transparency control in Arc. Map rel 40% ief 20% 5/30

Red-Green-Blue coordinates RGB used to define colours in computers and image files. Integrating colour and grey scale imagery involves manipulation of colours in RGB & HSV coordinate spaces. 6/30

Hue–Sat–Value coordinates Saturation Hue V = 100 120º 0 Value 0º S = 100 240º 0 0º 359º Hue 7/30

Conventional Value Replacement H R G B RGB 2 HSV R S V HSV 2 RGB STOP G B 8/30

Problems with Value Replacement • Loss/corruption of colour where: 1. Pixel colour is differentiated on Value (HSV) 2. Pixel colour has low saturation (achromatic axis of RGB cube) 9/30

Problems with Value Replacement 1. Pixel colour is differentiated on Value R, G, B H, S, V 0, 110, 0 120, 99, 43 0, 188, 0 120, 99, 74 0, 253, 0 120, 99 10/30 Replace V with pixel values from shaded relief

Problems with Value Replacement 2. Pixel colour has low saturation (achromatic axis of RGB cube) Using Values from grey-scale image … … to replace the Values of colours with low saturation … … results in poor image integration. 11/30

Conventional RGB Modulation R Multiply R’ G Multiply G’ B Multiply B’ 12/30

Problem with RGB Modulation Overwhelming majority of pixels in the output raster will have a darker colour than original. Better than value replacement. Colours are not “lost” but what happened to the yellow. 13/30

Inspiration for Sat. Val. Mod 14/30

Sat. Val. Mod Concept Lower Saturation Lower Value Cutoff 15/30

Sat. Val. Mod Multipliers Lower Saturation Multiplier Saturation multiplier Value multiplier Shade value Lower Value Cutoff 16/30

Sat. Val. Mod Detail H R Multipliers S V * * 1. 00 0. 75 = Sm = Vm HSV 2 RGB G B Pixel value in grey-scale image determines multiplier 17/30

Sat. Val. Mod in Arc. GIS Sat. Val. Mod written in VBA Runs in Arc. GIS 8. x or 9. x (Arc. Map – Arc. View or Arc. Info) 18/30

Sat. Val. Mod Colour Raster Options CLR file pixel value and RGB’s R G B 8 -bit image planes R G B 3 separate 8 -bit raster files 19/30

Sat. Val. Mod Parameters 20/30

Sat. Val. Mod parameters (Cut. Off) Cut. Off = 255 Cut. Off = 180 Cut. Off = 128 Cut. Off should be the maxima of the shade raster histogram. For shaded relief images, this is sin(Altitude). 21/30

Sat. Val. Mod parameters (Vmin) Vmin = 0. 3 Vmin = 0. 6 Vmin > 0 is best so shadows are not completely black. 22/30

Sat. Val. Mod parameters (Vexp) Vexp = 0. 3 Vexp = 1 Vexp = 3 Vexp 1 is recommended so shadows are not too extensive. 23/30

Sat. Val. Mod parameters (Smin) Smin = 0. 3 Smin = 0. 6 Smin < 0. 4 is recommended. Higher values do not provide enough decrease in saturation. 24/30

Sat. Val. Mod parameters (Sexp) Sexp = 0. 3 Sexp = 1 Sexp = 3 Sexp 1 provides better results 25/30

Geology + Shaded relief aeromag. Sat. Val. Mod does not lose or corrupt original colours. Value replacement Sat. Val. Mod RGB Modulation 26/30

Pan sharpening Colour critical? Use Sat. Val. Mod. Sharpening critical, colour unimportant? Use value replacement Value replacement Sat. Val. Mod RGB Modulation 27/30

Ternary images Value replacement Sat. Val. Mod RGB Modulation U-Th-K RGB composite image integrated with shaded relief DEM 28/30

Conclusions q Sat. Val. Mod provides better integration of colour and grey-scale imagery compared to many traditional methods. q Colour is more accurate. q SVM Parameters provide flexibility on integration. Value replacement Sat. Val. Mod RGB Modulation 29/30

Availability • Arc. GIS 8. x and 9. x MXD VBA • Arc. Map – Arc. View or Arc. Info • e-mail for download instructions viljoen@nrcan. gc. ca 30/30