Published on Sat Jun 24 2017

Evolving Spatially Aggregated Features from Satellite Imagery for Regional Modeling

Sam Kriegman, Marcin Szubert, Josh C. Bongard, Christian Skalka

Satellite imagery and remote sensing provide explanatory variables at relatively high resolutions for modeling geospatial phenomena. Regional aggregations are often desirable for analysis and actionable insight. We propose a novel method of inducing spatial aggregations as a component of the machine learning process.

0
0
0
Abstract

Satellite imagery and remote sensing provide explanatory variables at relatively high resolutions for modeling geospatial phenomena, yet regional summaries are often desirable for analysis and actionable insight. In this paper, we propose a novel method of inducing spatial aggregations as a component of the machine learning process, yielding regional model features whose construction is driven by model prediction performance rather than prior assumptions. Our results demonstrate that Genetic Programming is particularly well suited to this type of feature construction because it can automatically synthesize appropriate aggregations, as well as better incorporate them into predictive models compared to other regression methods we tested. In our experiments we consider a specific problem instance and real-world dataset relevant to predicting snow properties in high-mountain Asia.

Fri Jan 31 2020
Machine Learning
Improving the Detection of Burnt Areas in Remote Sensing using Hyper-features Evolved by M3GP
The aim is to understand the relationship between feature spaces and the predicting ability of the models. We find that training on datasets built from more than one image provides models that generalize better. These results are explained by visualizing the dispersion of values on the feature space.
0
0
0
Sat Sep 07 2019
Machine Learning
Unsupervised Image Regression for Heterogeneous Change Detection
Change detection in heterogeneous multitemporal satellite images is an emerging and challenging topic in remote sensing. One of the main challenges is to tackle the problem in an unsupervised manner. In this paper we propose a framework for bitemporal change detection based on the comparison of affinity matrices and image regression.
0
0
0
Tue Jul 03 2018
Computer Vision
SpaceNet: A Remote Sensing Dataset and Challenge Series
SpaceNet is an open source network of satellite imagery. It is being used to train machine learning algorithms to update maps faster.
0
0
0
Mon Jan 20 2014
Neural Networks
An Evolutionary Approach towards Clustering Airborne Laser Scanning Data
The original purpose of LiDAR systems was to determine the altitude of ground elevations. Newer full wave systems provide additional information that can be used on classifying the type of ground cover and the generation of maps.
0
0
0
Tue Jul 31 2018
Computer Vision
Remote sensing image regression for heterogeneous change detection
Change detection in heterogeneous multitemporal satellite images is an emerging topic in remote sensing. We propose a framework, based on image regression, to perform change detection in these images.
0
0
0
Wed Aug 21 2019
Machine Learning
Importance of spatial predictor variable selection in machine learning applications -- Moving from data reproduction to spatial prediction
Machine learning algorithms find frequent application in spatial prediction of biotic and abiotic environmental variables. However, the characteristics of spatial data, especially spatial autocorrelation, are widely ignored. This results in models that can reproduce training data but are unable to make spatial predictions.
0
0
0