Table of Contents
- Table of Contents
- Introduction to the platform
- Projects and Detectors
- Uploading images
- Creating your first detector
- Running your detector
- Viewing and exporting your results
- Generating a report
- Recap with our onboarding webinar
Introduction to the platform
Go through our introduction to the platform video to learn the basics of Picterra.
Projects and Detectors
Picterra consists of two main views. In the Projects view, you can see your list of projects. Projects are where:
- You upload your images
- You run your detectors to analyze your images
- You generate reports and export your results
On the other hand, in Detectors, you'll find your detectors which are used to detect objects on your images
Uploading images
To upload images, create a new Project or open an existing one. Then, click on Add Imagery and select Upload your files. You can then drag'n'drop your images - preferably in Geotiff format.
Creating your first detector
To create your first detector, click on Detectors from the top bar to go in your Detectors library. From there you can create a new detector by clicking on the blue Train a new detector button.
After you have given your detector a name, you need to add training images to it. To do that, open one of your project where you uploaded your images and select Add to training next to the image you want to use to train your detector. You can then click on Next in the top-left to start working on your detector. The training of the detector is explained in the in-app tutorial you can access from the blue help button at the bottom left of the detector workspace.
Here are few additional tips regarding the drawing of annotations and training your detector:
- The different Annotation Areas (training/accuracy/testing) are listed in the small window in the right. You can easily filter by name and order it to easily access a specific area.
- The different Annotation Classes are listed in the small window at bottom right of the window. You can create additional classes and change their colors.
- The different drawing tools are in the left toolbar. You need to distinguish between Annotations Areas and the Annotations themselves. The first are the zone provided to train the model, the latter are what the model is trained to detect or segment. For example, an image that you would entirely annotate would have a training area corresponding to the entire image.
Running your detector
To run your detector, you have to go in Projects and open a project. You then need to do two things:
- Assign your detector to this project. So that it will appear in the detectors list on the right sidebar of your project
- Select your detector from the right sidebar and run it on an image by clicking on the play button corresponding to the image.
Viewing and exporting your results
After your detector has finished analyzing an image, you can view results by clicking on the view results button corresponding to the image. From the results view, you can export your results in a variety of format such as GeoJSON, CSV and Shapefile.
Generating a report
From your project in Projects you can click on the Reports tab in the top right to access your reports for your results. From there you can generate various types of reports:
We currently support Heatmaps, Change Detection, Polygon Visualisation and PDF reports.
Once you've created the report you can share them using the share icon. Note that PDF reports can only be downloaded, not directly shared via a URL like the other reports.
Recap with our onboarding webinar
We also have a lot more help resources accessible from our main documentation page.