PHOTOMOD Radar 2.3.1. New version
More news on this topic
RACURS is the winner of the BRICS Solutions Awards
PHOTOMOD 8.0 New version of the photogrammetric platform
PHOTOMOD 7.51 New version of the photogrammetric platform
PHOTOMOD 7.5 New version of the photogrammetric platform
PHOTOMOD 7.4 New version of the photogrammetric platform
The upgraded photogrammetric platform, PHOTOMOD 7.3, is now available: new functions and reliability
The new build of PHOTOMOD Radar 2.0.1 for SAR data processing is available
PHOTOMOD Radar software supports GaoFen-3 SAR images
The upgraded photogrammetric platform, PHOTOMOD 7.2, is now available: new functions and reliability
The Photogrammetric Platform Strategy Approach. GIM - Issue 7 - 2021.
RACURS has announced the PHOTOMOD purchase and upgrade special offer.
Our upgraded photogrammetric platform, PHOTOMOD 7.1, is now available: new functions and software
The new version PHOTOMOD Radar 2.0.1 for SAR data processing.
RACURS has signed a distribution agreement with Capella Space.
2020 Year Review. A year that long be remembered.
PHOTOMOD Radar software for Sentinel-1 data processing
The production project for the processing of more than 100,000 UAS images
Date: 24 december 2024
Type of publication: news
Neural Network Processing of SAR Data module has been supplemented with annotation and training tools.
- Annotation Tool is designed for annotating SAR images, marking up target objects using bounding rectangles, and forming training and control test samples for neural networks.
- Training Tool is designed for training neural networks, as networks identify complex dependencies between input and output data and generalize them in the form of appropriate weights. The features describing objects of interest and classification procedures are also formed during the learning process. A properly trained neural network will be able to give the correct result on data that was initially missing from the training sample.
Both, annotations prepared in the Annotation Tool by PHOTOMOD Radar and external images annotated in COCO-json format, can be used as input to Training Tool.
Files with pre-trained weights of neural networks and training parameters have been prepared for detecting ships, aircrafts, and oil pumps, depending on the detail of the SAR images and the satellites they were captured with.
The pre-trained weights are available for SAR data from the following satellites: TerraSAR-X, Gaofen-3, Condor-FKA, Spacety, Umbra, Sentinel. The weights were obtained on the basis of small samples of the initial SAR survey data and are given solely for introductory purposes.
To speed up Training Tool, it is recommended to use Nvidia Tesla series graphics cards and other graphics computing systems based on the CUDA architecture.
For asking questions about the cost and getting trial licenses, contact the Commercial department: com@racurs.ru
For solving technical issues with PHOTOMOD Radar, please contact the Remote sensing data group: dzz@racurs.ru
Download the update: Ph_Radar_64bit_241223.zip (workable for current users only)
Download pre-trained weights and additional neural network processing files for Windows 10: Radar_Neuro_Win10.zip (workable for current users only)
Instructions for installing additional files to "Neural network processing of SAR data" module