Blog

DeepCore 1.4 Now Supports ONNX

DeepCore supports the Open Neural Network Exchange (ONNX) format, an open standard created by a consortium of major companies including Maxar (www.onnx.ai). Facebook, Microsoft, AWS, Nvidia and others started supporting ONNX to foster portability of neural network models. For the last few years deep learning algorithms were built and trained in frameworks with libraries that …

DeepCore v1.4.0

I’m pleased to announce the release of version 1.4.0 of the DeepCore Framework. DeepCore has undergone some changes for the 1.4 release, including adding support for ONNX models for inference in DeepCore, as well as adding a new Cloud Optimized Geotiff creation tool that is significantly faster than current COG generation tools already on the …

DeepCore Cogifier

DeepCore has a new capability! While the DeepCore team has been primarily focused on it’s machine learning and computer vision mission, we’re also interested in things that directly support that mission. Overhead imagery tends to be large, which makes it more expensive to send across the wire for inference jobs. However, there are image formats …

DeepCore Suite

It has been a while since we’ve posted anything on this site, but the DeepCore development team has been hard at work on the DeepCore Suite of tools. This Suite includes a series of micro-services that include tools for data set curation, chipping, scoring and inference. We have a thick client called DeepCore Workbench that …

2018 GTC-DC – DeepCore, CityBox and xTerrain

For anyone heading in to DC this afternoon, please attend the presentation by Buzz Roberts and Tony Frazier at GTC-DC! DC8250 – Harnessing Artificial Intelligence, Automation, and Augmentation to Build a Better World Growth in the geospatial sector has led to an explosive amount of data being collected to characterize our changing planet. The National …

OpenSpaceNet User Guide

OpenSpaceNet is our command line tool built on our DeepCore SDK that allows a user to download, perform object detection, and manipulate geospatial vector files. We felt it would be a good idea to give examples and descriptions of OpenSpaceNet’s most popular features. We hope that with this guide you can also create detections that look …

Validation and Verification of Machine Learning Detections using Tomnod

In our blog post entitled Discovering Pattern of Life Activity using Machine Learning, we described how the output from a machine learning algorithm can aid in characterizing human activity over time. We did this by counting all the objects from certain categories like planes, trains and automobiles and then view the results in aggregate. This …

YOLOD – Improving neural nets for detecting packed cars in satellite imagery

Neural network research moves very quickly and it can be hard to keep up. In our work at Radiant Solutions, we use a variety of machine learning modalities. My team works with neural networks for segmentation, detection, and other imagery processing tasks, but our most common application is object detection. Detection is basically the task …

AWS re:Invent – Gathering Machine Learning Training Data with Amazon Mechanical Turk

AWS re:Invent is a learning conference hosted by Amazon Web Services for the global cloud computing community. The event features keynote announcements, training, certification opportunities and more than 1,000 technical sessions. This year, Radiant Solutions was provided the opportunity to give one of those technical sessions on gathering machine learning training data for satellite imagery. …

Track GIS Technician Work with the ELK Stack – Part 3

Part 3: Use Kibana to Visualize Data from Elasticsearch In this final part of the series, I’ll show you visualizations that can be done with Kibana using data in Elasticsearch. This is the fun part and there’s no coding required! Note: There are countless visualizations you can make in Kibana, but these are a few …