7 Reading OBJ and STL Meshes

Once you’ve translated a CAD file through the Model Derivative API, you’ve got an OBJ or STL file sitting on disk. Now the fun part: loading it into R.

rgl (Adler, Murdoch, et al. 2024) and Rvcg (Schlager 2024) both read OBJ and STL files. rgl is great for interactive viewing; Rvcg is the one to reach for when you want to compute things. It returns a tmesh3d object that plays nicely with the whole Rvcg family of analysis functions.

flowchart LR
  A[CAD file\nin OSS bucket] --> B[translateObj / translateStl\nModel Derivative]
  B --> C[downloadFile\narial.obj]
  C --> D[vcgImport\ntmesh3d object]
  D --> E[Analyse\nmetrics / visualise / compare]

7.1 Prerequisites

install.packages(c("rgl", "Rvcg"))

7.2 Getting the OBJ File

The full upstream chain, translate, poll, get output URN, download, picks up where the Model Derivative chapter left off:

library(AutoDeskR)

resp     <- getToken(id     = Sys.getenv("client_id"),
                     secret = Sys.getenv("client_secret"),
                     scope  = "data:read data:write")
myToken  <- resp$content$access_token

myEncodedUrn <- jsonlite::base64_enc(Sys.getenv("urn"))

translateObj(urn = myEncodedUrn, token = myToken)

repeat {
  status <- checkFile(urn = myEncodedUrn, token = myToken)
  if (status$content$status == "success") break
  Sys.sleep(5)
}

outputUrn       <- getOutputUrn(urn = myEncodedUrn, token = myToken)
myOutputUrn     <- outputUrn$content$derivatives[[1]]$children[[1]]$urn
myEncodedOutput <- jsonlite::base64_enc(myOutputUrn)

downloadFile(urn        = myEncodedUrn,
             output_urn = myEncodedOutput,
             token      = myToken,
             destfile   = "aerial.obj")

7.3 Reading with Rvcg (recommended for analysis)

vcgImport() parses the OBJ and returns a tmesh3d, a list with a vertex matrix (vb) and a triangle index matrix (it). Every analysis function in the next three chapters works on this object.

library(Rvcg)
mesh_vcg <- vcgImport("aerial.obj")
str(mesh_vcg)
#> List of 4
#>  $ vb     : num [1:4, 1:2847] ...   # homogeneous vertex coords (X/Y/Z/W)
#>  $ it     : int [1:3, 1:1832] ...   # triangle face indices
#>  $ normals: num [1:4, 1:2847] ...
#>  $ material: list()

7.4 Reading with rgl (for interactive viewing)

readOBJ() returns a similar mesh3d object that you can render immediately:

library(rgl)
mesh_obj <- readOBJ("aerial.obj")

Tip

OBJ files from the Model Derivative API sometimes contain multiple named objects (one per CAD layer or body). vcgImport() merges them into a single mesh by default. Use rgl::readOBJ(material = TRUE) if you need to keep them separate for per-layer analysis.

7.5 Reading STL

Exactly the same call works for STL files. vcgImport() handles both formats transparently:

mesh_stl <- vcgImport("aerial.stl")
# Same tmesh3d structure — all subsequent chapters work identically on OBJ and STL

7.6 Quick Inspection

A few fast checks to make sure the mesh loaded correctly:

cat("Vertices:", ncol(mesh_vcg$vb), "\n")
#> Vertices: 2847
cat("Faces:   ", ncol(mesh_vcg$it), "\n")
#> Faces:    1832

# Coordinate ranges — useful sanity check that units are what you expect
apply(t(mesh_vcg$vb[1:3, ]), 2, range)
#>           [,1]      [,2]  [,3]
#> [1,] -142.830  -98.470  0.000
#> [2,]  318.620  241.190 12.500

The mesh_vcg object lives on from here through the Mesh Metrics, 3D Visualisation, and Mesh Comparison chapters, so keep it in your environment.