Introduction Aspect-slope maps are an interesting way to visualize reliefs. In essence, these maps are created by first discretizing both metrics, slope and aspect, and secondly combining the resulting values into a single one. In a third step, a custom… Read More »Creating Aspect-slope Maps with Python
Recently, I had to frequently share deep learning model outputs with a client. However, with each dataset being several GB in size, this increasinly became a challenge. By default many outputs, such as binary segmentation results, are stored as floating… Read More »GDAL Float to Binary Conversion
Introduction The introduction of AlexNet back in 2012 can be considered one of the breakthrough events that triggered the ongoing evolution of the deep learning (DL) subdomain of artificial intelligence. In that year, AlexNet greatly outperformed competing algorithms by a… Read More »A Brief Overview of AI-driven Image Analysis
Introduction In a series of earlier posts, we have discussed how geospatial imagery can be segmented using a deep neural network. When applying such a trained model to imagery data, noise is a common and often not desirable phenomenon. Figure… Read More »Noise Removal with Connected Components Analysis using GDAL and OpenCV
Introduction In the previous post, we have successfully trained a model for lake segmentation. In this post, we are going to have a look at how its performance can be quantitatively and qualitatively assessed as well as how it can… Read More »Semantic Segmentation of Geospatial Imagery with Deep Learning Part 6 – Model Inference and Evaluation
Introduction In the previous post, we have defined the U-Net architecture and defined a class to access our training samples. Hence, we are ready to train the model. This first requires to split up our image crops into different datasets… Read More »Semantic Segmentation of Geospatial Imagery with Deep Learning Part 5 – Training the Model
Introduction In the previous post, we have successfully created image crops that we are going to use for training the U-Net model. In this post, we are going to talk about how the U-Net architecture looks like and how it… Read More »Semantic Segmentation of Geospatial Imagery with Deep Learning Part 4 – U-Net
Introduction In the previous post of this series, we have acquired the relevant datasets and transformed the lake vector dataset into a raster file that aligns with the SWISSIMAGE dataset. In this blog post, we are going to use these… Read More »Semantic Segmentation of Geospatial Imagery with Deep Learning Part 3 – Training Samples
Introduction In the previous post, we have set up our working environment. In this blog post, we are going to have a look at from where we can obtain suitable training datasets. Training a deep learning model is only possible… Read More »Semantic Segmentation of Geospatial Imagery with Deep Learning Part 2 – Data Acquisition
Introduction Feature segmentation, i.e., the association of a pixel to a certain feature class such as a tree or a car, is a common task when processing remote sensing data. In recent years, this task has been dominated by deep… Read More »Semantic Segmentation of Geospatial Imagery with Deep Learning Part 1 – Setup