宙畑 Sorabatake

Machine Learning

Machine Learning

3rd Prize Approach of The 4th Tellus Satellite Challenge

This article was contributed by citron, the 3rd place winner of the 4th Tellus Satellite Challenge.

Machine Learning

The 4th Tellus Satellite Challenge! Explanation of approaches taken by the winners

"Coastline extraction" was the theme of the 4th Tellus Satellite Challenge. In this article, we will look at the approaches taken by the winners, make comparisons, and summarize their methods so that they can be applied to other (satellite) image data competitions.

Machine Learning

Using SSD to Perform Object Detection on Airplanes

For readers who want to try using machine learning and satellite data to do something interesting, we'd like to suggest trying out object detection. In this article, we attempt to use satellite data in order to perform object detection on airplanes.

Machine Learning

No Image Pair required! Converting To and From Optical and SAR Images With Unsupervised Learning

In this article we use a yet-to-be released data set to convert SAR images into optical images, and vise-versa by using unsupervised learning that doesn't require image pairs.

Machine Learning

Via using a Learned Model, we create a Machine Learning Model to Easily Detect Golf Courses [Code/Data Included]

Creating a machine learning model with transition learning by using a trained model makes it easier to perform image identification prediction. We will go into detail on images both with and without golf courses, and the coding behind them.

Case studies

Automatic Detection of Parking-Lot Space With Satellite Data Challenges and Points of Improvement for the Joint Project by Sakura Internet, akippa, and Ridge-i

During a talk session on August 4, it was announced that Sakura Internet, akippa, and Ridge-i are working together to develop an algorithm that uses satellite data to find space that can be rented out for parking cars, so we decided to talk with them about the challenges they face whilst developing this service, and the success they have made using Sharp's super-resolution technology.

Machine Learning

The 4th Tellus Satellite Challenge is now underway! The theme is “coastline extraction.”

Here is an overview of the 4th Tellus Satellite Challenge, a competition for segmentation using satellite images, with details of the competition, helpful papers, and material.

Machine Learning

The “Extracting Difference Between Two Points of Satellite Data” Challenge — A Look at ABEJA’s Difference Extracting Algorithm

We went behind the scenes to ask ABEJA about the applications and future of their difference extracting algorithm!

Machine Learning

The conversion of SAR to optical image using pix2pix and analysis of another SAR image with this generator

Using GANs, which has become a popular topic in recent years as an image generation algorithm, I tried to convert non-intuitive SAR images into optical images.

Machine Learning

Get super-resolution for satellite images using SRCNN [with code]

In this article, we will try to get super-resolution images of actual satellite data using Tellus.

Machine Learning

Super-Resolution Processing of Satellite Images Using Sharp’s Deep Learning Model

Super-resolution is a technique to artificially raise the resolution of an image. Super-resolution is one of the hot topics in the field of machine learning, but what happens when you combine it with satellite imagery? We went to the Sharp Corporation Research and Development HQ, and asked the manager of the 3rd Research Team for Communication & Image Technology Laboratories, Tomohiro Ikai, and researcher, Eiichi Sasaki about the future of this technology.

Machine Learning

The 3rd Tellus Satellite Challenge! ~ Check out the Winners’ Models ~

The third Tellus Satellite Challenge was held with a mission to "detect the extent of sea ice." In this article, we explain the challenge and introduce the approaches of the winning teams.

Machine Learning

Classify satellite images according to cloud density

AI (Artificial Intelligence) has now become pretty common in much of the business world. In the satellite data field, applications of AI such as machine learning and deep learning are capturing more and more attention every year.

Machine Learning

[Kaggle Competition Commentary Series] Identification of Sea Ice and Ships on Satellite Images

This article explains the analytical approaches that the top three winners took at the data science competition, Kaggle, to identify sea ice and ships from satellite images.

Machine Learning

Data Science competition of Sea ice detection : its purpose points on images

On October 4th, 2019, the "3rd Tellus Satellite Challenge", a satellite data analysis competition, began at SIGNATE. The theme of this contest is "detection of the sea-ice area." This article will explain the purpose and points to be considered on images.

Machine Learning

Vessel Detection— Introduction of the analytical approaches used by the winners of the 2nd satellite data analysis contest

We are going to introduce the analytical approaches used by the top three winners in the “Tellus Satellite Challenge”, a vessel detection algorithm competition using satellite data.

Machine Learning

The First Satellite Data Analysis Contest Report -The answers and what to look forward in the 2nd Challenge

We are glad to share the feedback on the 1st Tellus Satellite Challenge, the satellite data analysis competition, from Shu Saito, CEO of SIGNATE Inc. operator of the contests.