• Automate Analytics with
    Deep Learning for Faster,
    More Accurate Results

Deep Learning

Harris Geospatial has developed a suite of deep learning-based tools called MEGA™ that are designed specifically to work with imagery to solve geospatial problems. This technology is currently being used to solve real-world problems in industries that include agriculture, utilities, transportation, and defense.

Leverage the MEGA Advantage

MEGA excels at automated target detection, land cover classification mapping, and change detection. By employing a highly tuned process that relies on training models and high-performance computing to train a neural network, MEGA can identify and extract objects of interest or discover specific conditions across vast quantities of images and data. MEGA delivers results faster and more accurately than ever before.

Domain Expertise. Results You Can Trust.

The standard accuracy for pixel by pixel classification is between 75-85%. Harris Geospatial’s investment of time and money in deep learning technology has resulted in classification accuracies of up to 90-95%, and sometimes higher.

Leverage the power of Geospatial Deep Learning

Knowing how to work with imagery is the first step toward getting a good result with deep learning. For example, understanding spectral bands can help with data reduction to make processing more efficient. It is also often necessary to exploit the strengths of multiple data modalities to answer questions. As the image science leader, Harris Geospatial leverages its deep learning technology on optical, SAR, and LiDAR data to create rich geospatial products and answer specific questions for clients.


How Can We Help?

TIMELY TOPICS



MEGA™ Deep Learning Tech On GBDX

RECORED WEBINAR

Harris' MEGA™ deep learning technology on the GBDX Platform.


DEEP LEARNING FOR DEFENSE APPLICATIONS

PrecisionPass™

RECORDED WEBINAR

Mass production of deep learning algorithms for GEOINT is finally within reach.


UAV IMAGERY & DEEP LEARNING FOR WIND TURBINE INSPECTION

RECORDED WEBINAR

Use UAV imagery to automatically identify and locate abnormalities on wind turbine blades.


RELATED CONTENT


  Deep Learning Data Sheet (pdf)

  Deep Learning Whitepaper

  View Related Blog Posts

MEGA at Work

MEGA is offered through a flexible approach that combines the right steps and methods to deliver a deep learning solution to meet project requirements.


Manage Railroad Assets

Manage Railroad Assets with Deep Leraning Technology

Harris partnered with a railroad company to apply its deep learning technology on LiDAR data for asset inventory. The railroad company needed to find a variety of 3D objects (signals, crossings, boxes, poles). Applying deep learning to LiDAR data for 3D feature extraction can be very complex. However, Harris deep learning technology has proven to be very successful, and in this case extracted the features the railroad company was looking for with 90%+ accuracy.

Transmission & Distribution Inspection

Transmission and Distribution Inspeciton and Mega

MEGA is used by utility companies as a dependable and cost-effective solution to augment manual Transmission & Distribution inspections. In one recent customer engagement, Harris Geospatial successfully deployed MEGA to identify specific anomalies on distribution poles that, had they gone undetected, might have disrupted service. MEGA was also used to locate poles and insulators, as well as identify anomalies.

Defense & Intelligence

Mega Operationalizes Big Data and AI for Defense and Intelligence

The modern battlefield pose great challenges for even the most advanced militaries and intelligence organizations. Harris Geospatial has experience implementing solutions that use deep learning technology to address a wide range of GEOINT problems. Whether it’s navigating an armed conflict in an urban area, or sorting through an information minefield, MEGA will identify patterns, links, and anomalies to provide actionable information.

Wind Turbine Inspection

Wind Turbine Inspection Using Mega

Harris Geosptial has partnered with EdgeData to bring imagery expertise and deep learning to wind turbine inspections using Unmanned Aerial Systems (UAS). By using UAS, the risk of climbing a turbine is mitigated, and costs and inspection times are greatly reduced. MEGA is used to discern damage to turbine blades from lightning strikes that require repair verses dirt, paint chips, and damage from bird strikes that does not need to be repaired. Harris Geospatial’s solution can be repeatedly applied to produce correct damage assessment accuracies of greater than 95%.

We Develop MEGA Big Ideas

Have a problem that deep learning might be able to solve? Harris Geospatial can work with you to develop a solution that meets your project requirements.

 
First Name*
Last Name
Company
Email*
Phone
Country*
Tell us about your project*
YES! Send me news and updates from Harris Geospatial Solutions.