Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Changing radar factor to use a correct serialization strategy #611

Closed
wants to merge 1 commit into from
Closed

Changing radar factor to use a correct serialization strategy #611

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
33 changes: 20 additions & 13 deletions examples/marine/asv/rex/RadarFactor.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,6 +13,7 @@ arp2 = AlignRadarPose2(sweep[10], sweep[11], MvNormal(zeros(3), diagm([5;5;pi/4]
"""

using TensorCast
using Images, FileIO, Base64

import IncrementalInference: getSample
import Base: convert
Expand Down Expand Up @@ -61,30 +62,36 @@ function (rp2::AlignRadarPose2)(res::Vector{Float64},
end

struct PackedAlignRadarPose2 <: PackedInferenceType
im1::Vector{Vector{Float64}}
im2::Vector{Vector{Float64}}
im1::String
im2::String
PreSampler::String
p2p2::PackedPose2Pose2
end

function image2string(image::Array{Float64, 2})::String
io = IOBuffer()
save(Stream(format"PNG", io), image)
return base64encode(io.data)
end

function string2image(stringdata::String)::Array{Float64, 2}
io = IOBuffer(base64decode(stringdata))
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Just to help string together all the different image encoding methods -- here is a binary store via png (ImageMagick.jl):
https://juliarobotics.org/Caesar.jl/latest/concepts/interacting_fgs/#Adding-Data-Blobs

I also created this in Caesar.jl

Caesar.jl/src/images/imagedata.jl

Line 26 in 7ec2d11

function fetchDataImage(dfg::AbstractDFG,

return load(Stream(format"PNG", io))
end


function convert(::Type{PackedAlignRadarPose2}, arp2::AlignRadarPose2)
TensorCast.@cast pim1[row][col] := arp2.im1[row,col]
TensorCast.@cast pim1[row] := collect(pim1[row])
TensorCast.@cast pim2[row][col] := arp2.im2[row,col]
TensorCast.@cast pim2[row] := collect(pim2[row])
PackedAlignRadarPose2(
pim1,
pim2,
return PackedAlignRadarPose2(
image2string(arp2.im1),
image2string(arp2.im2),
string(arp2.PreSampler),
convert(PackedPose2Pose2, arp2.p2p2))
end

function convert(::Type{AlignRadarPose2}, parp2::PackedAlignRadarPose2)
TensorCast.@cast im1[row,col] := parp2.im1[row][col]
TensorCast.@cast im2[row,col] := parp2.im2[row][col]
AlignRadarPose2(
collect(im1),
collect(im2),
string2image(parp2.im1),
string2image(parp2.im2),
extractdistribution(parp2.PreSampler),
convert(Pose2Pose2, parp2.p2p2))
end
21 changes: 12 additions & 9 deletions examples/marine/asv/rex/ScanMatcher.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -98,7 +98,7 @@ using LinearAlgebra
using Optim

startsweep = 5
endsweep = 10
endsweep = 25
graphinit = false

# newfg = initfg()
Expand All @@ -112,23 +112,26 @@ for i in 1:(endsweep-startsweep)
end

# Run the initialization (very slow right now)
# ensureAllInitialized!(newfg)
ensureAllInitialized!(newfg)

# solving will internally call ensureAllInitialized!(newfg)
tree, smt, hist = solveTree!(newfg)

# Factor debugging
# fs = getFactorFunction.(getFactor.(newfg, lsf(newfg)))
# fs = filter(f -> f isa AlignRadarPose2, fs)
# pf = convert.(PackedAlignRadarPose3, fs)
# convert.(AlignRadarPose2, pf)
# using ImageView
# imshow(fs[5].im1)
# sim = image2string(sweeps[5])
# orig = string2image(sim)
#
# pf = convert(PackedAlignRadarPose3, fs[5])
# back = convert(AlignRadarPose2, pf)

# Save the graph
saveDFG(newfg, "$dfgDataFolder/segment_test.tar.gz");

lsf(newfg)
# this should run the radar alignment
pts = approxConv(newfg, :x0x1f1, :x1)

# solving will internally call ensureAllInitialized!(newfg)
tree, smt, hist = solveTree!(newfg)

## Looking at the results
using Plots
Expand Down