Modified Multi-Level Pattern Histograms (MLPH) for SAR image classification.
Some results for different regions in and around New York City area:
Platform tested on: Ubuntu 16.04
Softwares/Packages required:
Open the terminal inside the directory “IP_Project_Group23”
To run MLPH on training images and train SVM & test on hold-out set:
$ python -m sarclf -run_mlph 1 -train_svm 1
$ python -m sarclf
$ python -m sarclf -test 0 -clfimg <image_path>
$ python -m sarclf -test 0 -clfimg ./test_images/2.tif
Example:
$ python -m sarclf -test 0 -clfimg ./test_images/2.tif -modified 1
Following the research on MLPH, we derive a pattern matrix for each pixel based on a threshold value, use these matrices varying bin lengths to get local pattern sub-histograms, concatenated to give local pattern histogram. With multiple thresholds, the concatenation of local pattern histograms gives MLPH for each pixel.
此处可能存在不合适展示的内容,页面不予展示。您可通过相关编辑功能自查并修改。
如您确认内容无涉及 不当用语 / 纯广告导流 / 暴力 / 低俗色情 / 侵权 / 盗版 / 虚假 / 无价值内容或违法国家有关法律法规的内容,可点击提交进行申诉,我们将尽快为您处理。