mirror of
https://github.com/alicevision/Meshroom.git
synced 2025-05-29 17:06:35 +02:00
[multiview] Panorama: change default parameters
Use describerQuality=high as it improves quality on many challenging panorama datasets (in particular for the challenging matching between fisheye images) and the number of images to process on a panorama is usually reasonable.
This commit is contained in:
parent
3daf4a26f3
commit
715348f68d
1 changed files with 5 additions and 1 deletions
|
@ -180,6 +180,10 @@ def panoramaFisheyeHdr(inputImages=None, inputViewpoints=None, inputIntrinsics=N
|
||||||
panoramaHdr(inputImages, inputViewpoints, inputIntrinsics, output, graph)
|
panoramaHdr(inputImages, inputViewpoints, inputIntrinsics, output, graph)
|
||||||
for panoramaInit in graph.nodesByType("PanoramaInit"):
|
for panoramaInit in graph.nodesByType("PanoramaInit"):
|
||||||
panoramaInit.attribute("useFisheye").value = True
|
panoramaInit.attribute("useFisheye").value = True
|
||||||
|
# when using fisheye images, the overlap between images can be small
|
||||||
|
# and thus requires many features to get enough correspondances for cameras estimation
|
||||||
|
for featureExtraction in graph.nodesByType("FeatureExtraction"):
|
||||||
|
featureExtraction.attribute("describerPreset").value = 'high'
|
||||||
return graph
|
return graph
|
||||||
|
|
||||||
def panoramaHdrPipeline(graph):
|
def panoramaHdrPipeline(graph):
|
||||||
|
@ -214,7 +218,7 @@ def panoramaHdrPipeline(graph):
|
||||||
|
|
||||||
featureExtraction = graph.addNewNode('FeatureExtraction',
|
featureExtraction = graph.addNewNode('FeatureExtraction',
|
||||||
input=ldr2hdrMerge.outSfMData,
|
input=ldr2hdrMerge.outSfMData,
|
||||||
describerPreset='high')
|
describerQuality='high')
|
||||||
|
|
||||||
panoramaInit = graph.addNewNode('PanoramaInit',
|
panoramaInit = graph.addNewNode('PanoramaInit',
|
||||||
input=featureExtraction.input,
|
input=featureExtraction.input,
|
||||||
|
|
Loading…
Add table
Add a link
Reference in a new issue