- 50
- 27 598
Courage Kamusoko
Japan
Приєднався 19 кві 2021
My name is Courage Kamusko, and I am a geospatial consultant and instructor at Ai.Geolabs, in Tokyo, Japan. My consulting, training and research interests include remote sensing, GIS, land cover change and environmental modeling, and machine learning. I apply big geospatial data, machine learning, open-source software (R and python), and cloud computing (e.g., Google Earth Engine, Planetary Computer) to map and analyze environmental changes.
This channel features different categories or playlists of videos to help you work with machine learning and geospatial data analysis. Here you will find and access our tutorials, scripts, and other resources. You can find more information about our training programs at aigeolabs.com/.
For questions, please email me at: cou.kamusoko@aigeolabs.com
This channel features different categories or playlists of videos to help you work with machine learning and geospatial data analysis. Here you will find and access our tutorials, scripts, and other resources. You can find more information about our training programs at aigeolabs.com/.
For questions, please email me at: cou.kamusoko@aigeolabs.com
Top 10 Geospatial Python Libraries Every Beginner Should Know
Summary
This video explores ten essential Python libraries every beginner should know for geospatial data analysis. These libraries are powerful tools for handling, analyzing, and visualizing vector and raster geospatial data. From GeoPandas for vector operations to Rasterio and Earthpy for raster data handling, these tools streamline complex workflows. Libraries like Geemap integrate Google Earth Engine, while Scikit-learn enables advanced image classification and analysis. These libraries provide a complete toolkit to help you tackle geospatial challenges with confidence and precision.
Additional resources:
1. Python script
github.com/ck1972/Python-Geospatial_Model1/blob/main/4h_Ten_Essential_Geospatial_Python_Libraries_Demonstrations_GitHub.ipynb
2. Access courses at Ai. Geelabs
aigeolabs.com/courses/
3. Buy Data-centric Explainable Machine Learning eBook
4. Subscribe to my newsletter
substack.com/@couragekamusoko
This video explores ten essential Python libraries every beginner should know for geospatial data analysis. These libraries are powerful tools for handling, analyzing, and visualizing vector and raster geospatial data. From GeoPandas for vector operations to Rasterio and Earthpy for raster data handling, these tools streamline complex workflows. Libraries like Geemap integrate Google Earth Engine, while Scikit-learn enables advanced image classification and analysis. These libraries provide a complete toolkit to help you tackle geospatial challenges with confidence and precision.
Additional resources:
1. Python script
github.com/ck1972/Python-Geospatial_Model1/blob/main/4h_Ten_Essential_Geospatial_Python_Libraries_Demonstrations_GitHub.ipynb
2. Access courses at Ai. Geelabs
aigeolabs.com/courses/
3. Buy Data-centric Explainable Machine Learning eBook
4. Subscribe to my newsletter
substack.com/@couragekamusoko
Переглядів: 75
Відео
National Land Cover Mapping: High-Resolution Planet-NICFI Data and Machine Learning
Переглядів 5012 годин тому
Summary This UA-cam video demonstrates the production of a detailed, high-resolution (5 m) land cover map for Zimbabwe using Planet-NICFI base maps and a random forest classifier. The map integrates satellite imagery and training data prepared from Google Earth images and field checks conducted in Harare and Mazowe District. Additional resources: 1. GEE App ee-kamusoko-test.projects.earthengine...
How to Calculate SDG Indicator 11.3.1 in GEE
Переглядів 7014 годин тому
Summary Learn how to calculate SDG Indicator 11.3.1: the ratio of land consumption rate to population growth rate, a critical metric for measuring urban land use efficiency and sustainability. This step-by-step tutorial uses Google Earth Engine and datasets like GISD30 for impervious surface mapping and LandScan for population data. Follow along as we calculate Nigerian state land consumption a...
Seasonal NDVI Animation of the Zambezi River Basin (2013-2024)
Переглядів 4216 годин тому
Summary This video presents a mesmerizing animation of NDVI (Normalized Difference Vegetation Index) trends over the Zambezi River Basin from 2013 to 2024. The NDVI data is derived from the Suomi National Polar-Orbiting Partnership (S-NPP) NASA Visible Infrared Imaging Radiometer Suite (VIIRS) Vegetation Indices (VNP13A1) product. This dataset provides vegetation indices at a 500-meter resoluti...
Modeling Forest Biomass with GEDI L4A Data, Planet Imagery, and Machine Learning
Переглядів 114День тому
Summary This tutorial demonstrates countrywide (wall-to-wall) aboveground biomass density (AGBD) modeling using GEDI Level 4A data, Planet-NICFI imagery, and a random forest regression model. The workflow begins by defining the area of interest (Zimbabwe), loading Planet-NICFI imagery, and computing NDVI to enhance vegetation analysis. GEDI data is processed by applying quality, error, and slop...
Modeling Forest Canopy Height with GEDI L2A and Planet-NICFI Data
Переглядів 106День тому
SummaryThis video tutorial demonstrates how to model country-wide (wall-to-wall) forest canopy height using GEDI L2A and Planet-NICFI satellite imagery data within Google Earth Engine. We guide you through the entire process, from preprocessing the data (including NDVI computation, quality filtering, and masking by land cover) to training a random forest regression model. The model predicts can...
Deep Learning for Mapping Building Footprints: Using Grad-CAM for Enhanced Explainability
Переглядів 100День тому
In this tutorial, we will extract the building footprint using satellite imagery and the U-Net model. We will also apply the gradient-weighted class activation mapping (Grad-CAM) technique to explain the U-Net model. Grad-CAM helps us gain insights into how the U-Net model segments the building features. We use the Dar es Salaam dataset from OpenEarthMap, a benchmark global high-resolution land...
Creating a Training Dataset for Machine Learning | Part 2: Combining with Sentinel-2
Переглядів 6114 днів тому
In Part 2 of this tutorial series, we build on the filtered GEDI Level 4A Aboveground Biomass Density (AGBD) data prepared in Part 1. The script overlays the GEDI L4A data with Sentinel-2 spectral bands to create a comprehensive training dataset for machine learning modeling. We start by loading the project boundary (Mavuradonha Wilderness) and importing the AGBD sample points. Next, we retriev...
Creating a Training Dataset for Machine Learning | Part 1: GEDI L4A Data Preparation
Переглядів 10714 днів тому
In this tutorial, I demonstrate a complete workflow for preparing GEDI Level 4A Aboveground Biomass Density (AGBD) data using Google Earth Engine. This is Part 1 of a two-part series focused on creating a high-quality training dataset for machine learning applications. We start by defining the study area, which is the Mavuradonha Wilderness in Zimbabwe, and then proceed to load the GEDI L4A dat...
Deep Learning for Flood Mapping: Using Grad-CAM for Enhanced Explainability
Переглядів 25614 днів тому
In this tutorial, we will explore how deep learning can be applied to flood mapping, overcoming the limitations of traditional methods. Using a U-Net model for image segmentation, we will leverage free Kaggle datasets to identify flood-affected areas. To make our model more trustworthy and transparent, we implement Grad-CAM, a powerful tool that visually explains which areas in the image influe...
Explainable Geospatial Machine Learning for Modeling Forest Canopy Height
Переглядів 33121 день тому
In this tutorial, we dive into explainable geospatial machine learning techniques to model forest structure, explicitly focusing on forest canopy height. Using a random forest model, we will explore how to predict canopy height based on the GEDI canopy height dataset, eight Sentinel-2 spectral bands, and spectral indices such as the normalized difference vegetation index (NDVI) and canopy chlor...
Exploratory AGBD Modeler 1 01
Переглядів 4492 місяці тому
The Exploratory AGBD Modeler application models aboveground biomass density (AGBD ) using the Global Ecosystem Dynamics Investigation (GEDI) Level 4A (L4A) dataset, optical and synthetic aperture radar (SAR) imagery, and selected spectral indices. The app allows you to: • Select a study area from the global protected areas database; • Load and visualize optical (Landsat and Sentinel-2) and SAR ...
Introducing the AGBD Modeler App: Simplifying Biomass Mapping in Google Earth Engine
Переглядів 6172 місяці тому
The Exploratory AGBD Modeler application models aboveground biomass density (AGBD ) using the Global Ecosystem Dynamics Investigation (GEDI) Level 4A (L4A) dataset, optical and synthetic aperture radar (SAR) imagery, and selected spectral indices. The app allows you to: • Select a study area from the global protected areas database; • Load and visualize optical (Landsat and Sentinel-2) and SAR ...
Improving AGBD Models: Combatting Overfitting with a Data-Centric Approach in Machine Learning
Переглядів 2262 місяці тому
In the previous tutorial, we identified overfitting issues in our random forest model. In this tutorial, we will tackle this challenge by improving both the quality and quantity of our training dataset. We will demonstrate how to use the quality_mask, error_mask, and slope_mask functions to filter out unreliable Global Ecosystem Dynamics Investigation (GEDI) Level 4A (L4A) aboveground biomass d...
Modeling AGBD Using GEDI, Sentinel-2, and Machine Learning: An Introductory Guide with Python
Переглядів 4672 місяці тому
The tutorial models aboveground biomass density (AGBD) using the Global Ecosystem Dynamics Investigation (GEDI) Level 4A (L4A) dataset, Sentinel-2 (S2), spectral indices, Shuttle Radar Topography Mission (SRTM) digital elevation model (elevation and slope), and a random forest method. We use Mafungautsi Forest Reserve in Zimbabwe as the test site. Course, script, and blog post links aigeolabs.c...
Monitoring Water Quality using NDSSI in Manyame Catchment Area, Zimbabwe
Переглядів 2502 роки тому
Monitoring Water Quality using NDSSI in Manyame Catchment Area, Zimbabwe
Mapping Countrywide Wall to Wall Forest Cover Using GEDI and Sentinel Data in Earth Engine#3
Переглядів 1,7 тис.2 роки тому
Mapping Countrywide Wall to Wall Forest Cover Using GEDI and Sentinel Data in Earth Engine#3
Prepare Training Data using GEDIs Level 2A Data in Earth Engine#1
Переглядів 7742 роки тому
Prepare Training Data using GEDIs Level 2A Data in Earth Engine#1
Mapping Woodland Cover using GEDI and Sentinel Data GEE
Переглядів 6952 роки тому
Mapping Woodland Cover using GEDI and Sentinel Data GEE
Modeling Forest Canopy Height using EO Data and GEE
Переглядів 4 тис.2 роки тому
Modeling Forest Canopy Height using EO Data and GEE
Pixel-based Urban Land Cover Mapping using Rainy Season Sentinel 2 Imagery
Переглядів 1912 роки тому
Pixel-based Urban Land Cover Mapping using Rainy Season Sentinel 2 Imagery
Net Primary Productivity Southern Africa
Переглядів 4652 роки тому
Net Primary Productivity Southern Africa
great. thank you Dr.
That's so useful. Thank you very much.
you did not prepare the testing and training image and mask And its better to show the procedure of prepare the mask from the image.Thanks
Dear sir, I am Dinesh from Sri Lanka, I am interested to learn and practicing this script belongs to Sroi Lanka. I got a problem this related to Sri Lanka , North Central Province, No layers display. Please help to get better picture.
Please send me your script link.
nice tutorial, Thanks Courage
You are welcome
😂 i got a lot of error , ❤❤❤❤
Please send me your code link. I will look into your error.
that was great
The code is the description below. Please confirm if you can access the code.
yes
Hello, Honorable. I will be grateful if you share the code.
Which study region were you focusing on ?
I used a free dataset from Kaggle. However, the study area is not defined.
Super
Wow Appreciate with the most precious course.
it is so useful and professional video, I want to perform this method for my study area, is it CHM?
Yes, it is CHM
hi, thank you so much for useful video, this is CHM or FCH?
This is CHM (canopy height model).
many thanks for this video. I have a question. What is the different between extracting HEGHT from GEDI in the Googleearth engine and nasa hub from aspect accuracy?
I do not know. I have not yet tried the GEDI canopy height at the NASA hub.
Wow, that was such a wonderful lecture. How can I access the GEE codes?
aigeolabs.com/courses/course-4/
hello Is it possible please share the code
Thank you! Could you please provide and video related to the Flood Susceptibility mapping using Machine or Deep Learning methods?
This sounds like an incredibly useful and powerful application for researchers and conservationists👍
One of the best instructors I have ever met in the geospatial space
Could you show us how to use Partial dependance plots PDPs in GEE. many thanks
Cool ; I appreciated
Very Nice
can you provide the link of the javascript
Here is the link (code.earthengine.google.com/e7363e3fb08cbf6097978f456e49d5f6). Note that all links and tutorials are found at aigeolabs.com/
i am not able to access the link for the script. is there any possibility of providing the script?
Hi. The scripts are available for free for members.
I have subscribed your channel and liiked this video. Very helpful
Thanks a lot for this usefull video ! I'll like to know how to map forest AGB or AGBD using Forest Canopy Height map ?
I do not know how to model AGBD using only forest canopy height.
@@couragekamusoko5689 Thank you very much for your answer. I wanted to use tree canopy height and Sentinel 2 variables (bands, vegetation indices) and SRTM (slope, altitude, aspect) to model the AGB. But I'll try something on my own first and then eventually come back to you for help. Thanks !
Nice tutorial! I have one question. The projection of training and validation is 4326. Whether it should be reprojected to 32735, same as the other data.
Yes, it should be reprojected to 32735.
Good videos for beginners. But you never shared your code
Non of the links are working
how can I download the image zip folder with all the other files, not just the .tiff file that it makes available in the drive? I need the other vv, vh, iw tapes etc... Right now I just have the .tiff image with the 44 bands. Thanks
Hello Courage, It is a wonderful video and a great work. While working on the code, when i select a different area and try to project it, am receiving the error "Projection error: Image.divide: If one image has no bands, the otger must also have no bands. Got 0 and 1." I changed the date range and area, however, the issue persist. Could you kindly help? Thanks in advance.
Hi, i face the same problem. Any solution to this problem?
Very useful channel. Thank you for sharing!
Can i ask why are you using GEDI level 4b product and not level 4a and what is the difference?
The main difference between GEDI L4a and L4b AGBD datasets is that L4a provides footprint-level predictions, while L4b offers gridded estimates of AGBD with greater accuracy. I will also prepare a tutorial with Level 4a in the near future.
this is great work absolutely amazing
Thank you for this helpful video.
Hey Courage! what a great tutorial! this is incredibly useful.
Are these datasets pre-processed like Thermal Noise/Border Noise removed?
Great work, thank you. I've got this type of error after changing the starting and ending dates. "NPP: Layer error: reduce.mean: Error in map(ID=2022_01_01): Image.multiply: If one image has no bands, the other must also have no bands. Got 1 and 0". Could you please help me to solve this issue.
Please share your Earth Engine code link so that I can check it.
thanks for sharing good work
I can not find your previous video on loading GEDI level 2A in earth engine that you mentioned in this video.
Excellent work presented. My question is whether these same results can be extrapolated to forage crops for animals. I would like to estimate the biomass of that forage crop. Thank you very much. I'm from Formosa, Argentina
Greetings, when running the code I have this error, please guide me how to solve it. Thank you "Line 223: Can't transform (-892096.5,-483859.0)"
Please email your code
Thank u professor
Thank you very much for this wonderful tutorial. Can I get the code, please?
Puedes compartir el Scritp?
how to fuse sentinel 1 and 2 in gee
cant even acess your website .requires username and password. pls provide a script link
My sincere apologies. The website was down (maintenance). You can now access the free courses.
Thanks a lot!! really useful!!!!
Thank you for your course. I wonder if there are any relevant papers for reference about this method. Thanks a lot
Please check on science direct!!