Okay, so I figured out how to do it, although it's not very elegant:
# load datapoints <- readLAS(input_path)# returns the RGB values for the highest pointsRGBZ <- function(r,g,b,z) { bands = list( R = r[which.max(z)], G = g[which.max(z)], B = b[which.max(z)] ) return(bands)}# create & save orthoortho <- grid_metrics(points, ~RGBZ(R,G,B,Z), res = 0.1)writeRaster(ortho, output_path)