Faster imhist

[yha]

Converting each element to real makes the searchsortedlast call dispatch on the faster method specialized for ranges.

Benchmarks:

using BenchmarkTools
using TestImages
img = testimage("lighthouse")
img_gray = Gray.(img)
img_normed = real.(img)
img_float = float.(img_normed)

Before:

@btime imhist($img)           # 24.367 ms (5 allocations: 386.06 KiB)
@btime imhist($img_gray)      # 23.378 ms (3 allocations: 1.92 KiB)
@btime imhist($img_normed)    # 12.228 ms (3 allocations: 1.92 KiB)
@btime imhist($img_float)     # 14.518 ms (3 allocations: 1.92 KiB)

After:

@btime imhist($img)           # 14.108 ms (5 allocations: 386.06 KiB)
@btime imhist($img_gray)      # 12.681 ms (3 allocations: 1.92 KiB)
@btime imhist($img_normed)    # 12.379 ms (3 allocations: 1.92 KiB)
@btime imhist($img_float)     # 13.913 ms (3 allocations: 1.92 KiB)

[codecov]

Codecov Report

Merging #820 into master will increase coverage by 21.09%.
The diff coverage is n/a.

@@             Coverage Diff             @@
##           master     #820       +/-   ##
===========================================
+ Coverage   59.36%   80.46%   +21.09%     
===========================================
  Files          10       10               
  Lines        1228      906      -322     
===========================================
  Hits          729      729               
+ Misses        499      177      -322

Impacted Files	Coverage Δ
src/exposure.jl	95.02% <ø> (+28.98%)	⬆️
src/labeledarrays.jl	76.92% <0%> (-6.42%)	⬇️
src/Images.jl	33.33% <0%> (ø)	⬆️
src/algorithms.jl	74.2% <0%> (+11.06%)	⬆️
src/convexhull.jl	95.34% <0%> (+14.95%)	⬆️
src/edge.jl	67.7% <0%> (+18.09%)	⬆️
src/bwdist.jl	59.57% <0%> (+21.73%)	⬆️
src/connected.jl	82.22% <0%> (+31.18%)	⬆️
src/corner.jl	82.85% <0%> (+32.85%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 48c4571...75543c2. Read the comment docs.

[yha]

Rebased

[timholy]

Thanks!