NavList:
A Community Devoted to the Preservation and Practice of Celestial Navigation and Other Methods of Traditional Wayfinding
From: Bill Lionheart
Date: 2016 Jan 29, 06:51 -0800
I made a preliminary forray in to how to automatically analyse the photo. After rotating so that the horizon was horizontal I summed the red channel of the photo along rows of pixels. I think this makes it easier to pick out the centre of teh sun and the horizon. In this case I did it only on a range of x pixels between 1000 and 1500 which includes the sun but not the yacht. The sea surface is actually clearer in teh red chanel further from where the sun light is reflected so that suggests a wider shot with a calear horizon is best. Anyway so far I stil get 2403 pixels
I did this in Matlab, but it would work as well in Octave I think which is free. The commands I used are
I = imread('DSLR-Sun-Hs.jpg');
I=I(:,1000:1500,1);% Red channel
subplot(1,2,1)
colormap(gray)
imagesc(I(size(I,1):-1:1,:));
set(gca,'YDir','normal')
subplot(1,2,2)
plot( sum(I(:,:),2),size(I,1):-1:1)