Originally published by UP42. See original blog here
One of our core missions here at UP42 is to enable access to geospatial data and analytics. In the past, we have covered how our products connect you to the data you need. But ordering and downloading data is often not enough - you need to process the data to make it analytics-ready, and then subsequently get actionable insights from it.
The UP42 analytics platform complements our data offering. It hosts a collection of processing blocks that lets you run algorithms on geospatial data, and obtain accurate results for your specific use case. As we continue to look for new ways to enrich your experience on UP42, we’ve made a significant step forward by partnering with a major analytics solution provider: we’re joining forces with Trimble to provide their eCognition services for our analytics partners.
Advanced Processing Capabilities With Trimble eCognition
Solution builders and developers that use eCognition are often faced with a variety of technical and business challenges. In order to build an end-to-end workflow that can produce high quality analytics solutions, they require easy and flexible access to very high resolution (VHR) geospatial imagery. As they grow their operations, they also need access to reliable backend infrastructure that can reliably support processing at scale.
With this new partnership between UP42 and Trimble, we offer a novel approach for eCognition developers to create and showcase their products on our platform. If you’re already building analytics products powered by eCognition, you now have the ability to become an UP42 partner, create enhanced processing capabilities using our analytics platform, and start offering your solutions to our diverse geospatial customer base through the UP42 marketplace.
We offer multiple ways to kickstart this partnership and grow your business. Our extensive documentation guide walks you through the process of creating your own custom block built on eCognition (A block is a component within the UP42 workflow that retrieves specific algorithms). You can also get in touch with our team directly at firstname.lastname@example.org to learn more about the process of becoming an UP42 partner. Once your analytics solution is ready, you can publish it on the UP42 marketplace to reach a global audience, thereby unlocking an additional revenue channel for your business.
Multi-resolution Segmentation Using eCognition on UP42
We want to enable our partners to learn more about the capabilities of eCognition and develop their own blocks. With this in mind, we have worked together with Trimble to release our own Multi-resolution Segmentation (MRS) block on the UP42 platform. We have published the code for this block and made it public at this Github repository, making it easier for you to get started with using your own eCognition rulesets and build a new block.
Multi-resolution is a general segmentation algorithm used specifically for very high resolution (VHR) geospatial imagery. It’s effective for a range of different use cases, such as land cover classification, vegetation management, urban area analysis, and detection workflows. The MRS block ensures high quality and accuracy by locally minimizing the average heterogeneity of image objects–hence, this block can be used to accurately detect a range of different objects (of varying sizes, shapes, and spectral bands). Apart from the use cases already mentioned above, this block is useful in any scenario that calls for accurate object detection or segmentation.
How can you start using the MRS block on UP42? Start by signing up for free on the UP42 platform and entering your relevant organization information. Once that’s completed, you’re ready to get started. Using the UP42 catalog, you can search for your desired AOI, find the list of available imagery, and then place an order based on your requirements.
Once the order is complete, the images are delivered to the storage folder within the UP42 platform.
After you find the downloaded imagery within storage, you can use the Processing from Storage functionality within UP42 to create a new workflow for processing the downloaded images. At this stage, we add the MRS block to perform object-based image analysis on our downloaded Pleiades imagery.
Input image from Pleiades sensor
Output using the MRS block
By using the MRS block, we've just completed object-based image analysis on the input Pléiades imagery. This workflow can be applied to a variety of different use cases that call for object detection on very high resolution imagery. If you are a geospatial solution provider, this block can be used as the first processing step to get the specific kind of data insights that you need. As we continue to bring additional functionalities and new partners on board, watch this space for more analytics platform stories coming soon!