- Enabled the option to upload multiple GeoTiff pieces of a large orthomosaics into one dataset. Now it is possible to select many files during initial upload and also add later through “Layers->Map->Upload orthophoto”
- Introducing the option to browse between datasets of the same project using the project's timeline.
To display use the “Timeline” button in the bottom left corner of the map
We adapted Maps and Images to smartphone screens and tablets.
Now it is even more convenient to share results of your survey with others without laptops and desktop computers
- Project->DataSet->Map. Annotation toolbar now has the following sections:
- Objects - to annotate objects of interest
- Background - to define background for each object type. Previously we called it “Training zone”, but according to user feedback we simplified naming. Backgrounds now can be both rectangular and polygonal
- Working area - optional areas limiting the detection area. Using working areas may save time and megapixels during processing
- Filters - now it is possible to filter annotations by size (in square meters) and detection confidence
- Project->DataSet->Images. Annotation tools are the same as for maps, except the working areas which are not useful for raw imagery. Also filters allow you to select images by detection confidence.
- Project->Dataset->Maps. Right bar redesigned.
- Objects are now represented as a simple list for easy access and visualization. Each object may have its own color code adjusted by user
- List of working areas can be found below the objects. Each working contains number of objects of different types within it
- Project->DataSet->Images. Right bar redesigned. Objects are now represented as a simple list for easy access and visualization
- Project->DataSet->Map. Dataset now displays vector data available for the project. Now it is possible to show DXF on top of an orthophoto map and compare planned construction progress with the actual state.
- Maps and images checkboxes in the right bar now are memorized for users I.e. after returning back to the same view in the same browser you’ll get the same object/layer setup
- Minor bug fixes and usability improvements
DXF visualizer in Project->Vector data->Map section now supports more styles: line thickness and colors
- Atlas is publicly available for sign-up
Following plans available: https://atlas.ugcs.com/#pricing
- Operations for training and detection are simplified
Flag “Detected objects have a convex shape” and threshold field has been removed
- Possibility to minimize Megapixels spending
Now it is possible to make test detection runs on the selected subset of images or selected area on the map, to minimize Megapixels spending
- Better progress indication for all long-lasting operations
Possibility to draw self-intersecting figures during manual annotation implemented
- Image downloading now applies to filtered images
- Overall stability improvements and minor bug-fixes
- Bug: Fixed non-latin symbols handling in DXF Georeferencing
- AI has become a bit “smarter”
Newly trained detectors are more robust in different light conditions
- For annotation, without any training zone, an invisible training zone with the same shape as annotation has been added
That may reduce the number of training faults due to the absence of a training zone. But for better detection results it is still very important to add training zones manually. Training zones help to identify background.
- Sometimes on maps, there was experienced some detections outside of the map area
Fixed cases leading to such behavior
- Problem with insufficient zooming in Images fixed.
It is possible to “zoom in” to physical image resolution to see small details
- Image size in pixels added to Image details.
To display, click “Image details” button in the bottom right corner in the image viewer
- Added feature to download images as a single ZIP archive.
To download click “Download raw images”
- “Download” button added to [Your project] -> Survey data -> [Your dataset] -> Images.
- “Export detection report” operation allows to download CSV files for each label assigned to files in the dataset. CSV will contain image names associated with a label and coordinates from EXIF
- Significantly improved the speed of object counting after the selection on the map
- All cards (project, dataset names) now accept long names and do not stretch
- Renamed “Run” to “Detect” which better reflects the meaning of this operation - start detection with an existing detector.
- Applies to Images and Maps
- Now Download image operation in Image Details opens a separate browser's window
Experimental! The Vector data section is now enabled. It is possible to import 2D DXF data and georeference it to global coordinates
- Slightly changed structure of Projects. In future a Project will consist of “Survey data” and “Vector data” sections.“Survey data” leads to datasets with raw images and maps. “Vector data” is for project-level technical drawings in DXF format. The roll-out of “Vector data” section will be done in several days.
- The section [Your project]->Survey data->[Your dataset]->Images is updated. It enables to annotate data here and train the detectors on raw imagery
- Added support for large maps. Now it’s possible to upload GeoTiffs larger than 5Gb to [Your project]->Survey data->[Your dataset]->Map
- The upload in Images section of a Dataset is limited to JPG and PNG files only. This is because we noticed that some users try to upload large GeoTiff maps in this section
Improved Zoom function in Images section of a Dataset.