Merge pull request #42 from aquatiko/rgb_hsv

added demonstration: RGB to HSV and thresholding

docs/make.jl

@@ -21,6 +21,7 @@ makedocs(format = :html,
                      "Demos" => Any[
                          "demos.md",
                          "demos/color_separations_svd.md",
+			 "demos/rgb_hsv_thresholding.md",
                      ],
                      "function_reference.md",
                      "api_comparison.md",

docs/src/demos.md

@@ -21,6 +21,21 @@
 </div>
 ```
 
+```@raw html
+<div class="card-200">
+<div class="card-img">
+```
+[![rgb_hsv_thresholding](assets/demos/rgb_hsv_thresholding.png)](@ref rgb_hsv_thresholding)
+```@raw html
+</div>
+<div class="card-text">
+```
+[RGB to HSV and thresholding](@ref rgb_hsv_thresholding)
+```@raw html
+</div>
+</div>
+```
+
 ```@raw html
 </div>
 ```

docs/src/demos/rgb_hsv_thresholding.md

@@ -0,0 +1,38 @@
+# [RGB to HSV and thresholding](@id rgb_hsv_thresholding)
+
+This example illustrates how RGB to HSV (Hue, Saturation, Value) conversion
+can be used to facilitate segmentation processes.
+A simple segmentation of the image can then be effectively performed 
+by a mere thresholding of the HSV channels.
+
+
+```julia
+using Images, TestImages, LinearAlgebra, Interact, ImageView
+
+rbg_img = testimage("lighthouse")
+hsv_img = HSV.(rbg_img)
+channels = float(channelview(hsv_img))
+
+hue_img = channels[1,:,:]
+value_img = channels[3,:,:]
+saturation_img = channels[2,:,:]
+binary_img = zeros(size(hue_img))
+@manipulate for hue in 0:255, saturation in 0:255, value in 0:255
+    fill!(binary_img, 0.0)
+    for ind in eachindex(hue_img)
+        if hue_img[ind] <= hue && saturation_img[ind] <= saturation/255 && value_img[ind] <= value/255 
+            binary_img[ind] = 1
+        end
+    end
+    colorview(Gray, binary_img)
+end
+```
+
+Here's the result in IJulia:
+
+![lighthouse](../assets/demos/rgb_hsv_thresholding.png)
+
+You can click on the slider bars to change the threshold
+of hue, saturation and value.
+The obtained binary image can be used as a mask on the
+original RGB image.