311
Deep learning CV and Earth Observation data. What satellites can see? Monitoring of Environment, Land management; Urban planning and Mapping, Mapflow.ai etc. Сontact us at hello@geoalert.io or @godnik0
Mapflow maintnance service announcement!
Our web application app.mapflow.ai will be unavailable from Friday, September 27th 4:00 UTC to Monday, September 30th 6:00 UTC due to the planned update.
We highly recommend you plan your work in the web app so that your processings would not be affected during this time, otherwise they may be canceled.
The Mapflow #API, including Mapflow-QGIS plugin will be down for 2 hours during this period, and available at all other times. The processings launched via API / plugin will continue without interruptions.
On the positive note, this is not just a maintenance, and will bring you new capabilities, so stay tuned to see the update!
One more #opening #role in our RnD team. We are looking for Frontend developer (Middle)
Our primary product:
Mapflow.ai
Our stack:
React, Next.js, Node.js
effector
react-query, GraphQL
sass/scss, tailwindcss, chakra-ui, styled-components
Required skills:
2+ years and confident skills in web development and javascript/typescript
React, Next.js, Rest API, HTML/CSS experience, and skills
Good communication skills in English and Russian (B1+)
As an advantage:
UX/UI prototyping in Figma
GIS and Maps APIs tech experience
Docker
Node.js
You are welcome to promote any additional skills that might be useful for the startup team
What do we offer:
We have a lot of opportunities for professional growth and career in our company.
💲Salary depends on the technical level and experience, salary bump is TBD every 6 months
📍Remote work. Office: Tashkent, in the city center.
Contacts to send your CV:
Email: hello@geoalert.io
Tg: @godnik0
Check for more vacancies at Geoalert on our web-site.
📢 We are looking for a full-time cartographer to join our AI training team.
You will support us with data labeling/validation, conducting mapping tasks, and contributing to the development of Geoalert’s platform for AI mapping using aerial and satellite imagery.
You would probably be a perfect candidate if:
You have 1-2 years of experience in cartography
You are experienced to work in QGIS
You are open to new technologies, methods, and approaches
What do we offer:
You will join the small team of cartographers under the supervision of ML engineer and a business manager.
We communicate in Russian and English.
We have a lot of opportunities for learning about new professional tools and growing your skills as a geospatial professional.
💲Salary depends on the technical level and experience, salary bump is TBD every 6 months
📍Remote work. Office location: Tashkent
Contacts to send your CV:
Email: hello@geoalert.io
Tg: @godnik0
Check for more vacancies at Geoalert on our web-site.
While conducting some CV #research in Arizona, U.S. – chatting with a friend who is a dedicated pilot of light aircraft.
– Wow, It's not bad in Arizona. Landed next to your house, go chilling out in the swimming pool.
– It's bad in Arizona. Check temperatures. 😅
⚠️ #alert #work-in-progress ⚠️ Last couple of days we have been solving temporary issues with our infrastructure / public apps. We tried to reach out to those who might have experienced the issues using #Mapflow. If you have any issues – don't hesitate to contact our support at help@geoalert.io or via the chat in the app.
Читать полностью…
#GeoQuiz
The unexpected application of Segment-Anything (#SAM, Zoom 18) using Mapflow.ai.
Try to guess what, where? 🤔
Users who send us the correct answers will receive a promocode for 500 Mapflow credits. 😉
Some users keep experimenting with SAM for field boundary delineation. And they even get relevant results (even using an input image of not a very high quality). 😎 Don't hesitate to leave your feedback like this user as we are collecting more evidence to tune SAM for specific object classes. Mapflow powerful capability to run analysis on a large scale + tuned SAM can become a game changer for your mapping tasks. ✨ #India #fields #SAM #Mapflow
Читать полностью…
#Mapflow 3D #building footprints generated from 2D 0.5m RGB satellite image.
錦田, Luoyang, 泉州台商投資區, Quanzhou, Fujian, China
В эти дни проходит конференция «Современные геоинформационные технологии 2024», организатором которой выступает компания Дата Ист.
Коллеги продемонстрировали интеграцию с #Mapflow AI, которая предоставляет возможности обработки космических и аэрофотоснимков совместно с инструментами интерактивной карты платформы CoGIS.
Вкратце, как это работает:
1) Создание нового пользователя Mapflow (необходимо получить API токен)
2) Создание новой обработки
3) Получение результатов от сервера Mapflow и дальнейшая работа с ними в CoGIS.
✍️
Демонстрационный прототип интегрированного решения Mapflow <> CoGIS.
Документация Mapflow #API.
Обращайтесь к нам на hello@geoalert.io, если у вас появились вопросы по разработке и внедрению решений на основе AI картографии.
Geoalert's workshop at the Mapacon Online conference (in Russian) 👆
Читать полностью…
We got to share a new customer success story by Island Green Power whose mission is to help countries increase their solar energy usage, making more renewable energy whilst drastically reducing carbon emissions. So they design solar plants and map the projects.
"Mapflow Qgis integration assists us in swiftly identifying hedgerows and forests during the preliminary analysis of environmental and landscape constraints. Thanks to Mapflow, we have significantly streamlined the preliminary design processes for solar and BESS projects." - says Antonio Saponaro, Project Designer.
Thank you for sharing your story 🤗
This week our #team is in Istanbul for the teambuilding event ❤️. There are nine of us, not all could come. Istanbul is beautiful as always even through the bloody and dark times in history.
On Friday, 10.30 am – 1.30 pm, we will hold a seminar about Mapflow AI for our local partners (we will talk about some insightful user feedback and present the new experimental workflow to analyse imagery with the text prompts). Let's meet in real life by this chance. Feel free to join, we'll send the details about the event - register here https://lnkd.in/e6-cpbEt
Fields model deprecation 👎
🚜 Fields (hi-res) - agricultural fields boundaries detection model - has been deprecated. It didn't meet our standards and we decided to remove it from our default models list in #Mapflow, so we could focus more on our best performing models and get back to this one as soon as we have more resources.
From now on users will not be able to create new and copy old processings using this model, however old results will be kept and available.
If you want to keep using this model (in some territories it performs good enough and was used for research and commercial purposes) - drop us a message and we will connect it to your Mapflow account.
Sometimes you are supposed to be AI chatting with the visitors. Otherwise they don't want to continue talking to you. 😅
BTW, if anyone is experienced in developing #ArcGIS plugins and eager to integrate #Mapflow API - we are ready to discuss and support you.
Today we launched a PoC project with geocloud.work to provide Mapflow-QGIS through the #Geocloud's geospatial software marketplace.
Mapflow-QGIS is a preinstalled plugin for #QGIS to leverage geospatial imagery segmentation and classification with #Mapflow AI.
The PoC project aims to collect more evidence about using remote infrastructure while Mapflow is connected as an external instance. So users pay for uptime and virtual machine configuration in Geocloud. Mapflow credits are used to run the analysis (each user gets 250 credits for free upon registration).
However, the deeper integration by running the Mapflow AI on Geocloud's instances requires more effort and will make sense after collecting more evidence and requests from the clients.
The essential point is that you can configure VM with different licensing and free software to work in a single workspace.
Follow our guide on how to run Mapflow-QGIS in Geocloud. Contact us if you have any feedback or requests.
#Mapping #Geospatial #DataAnalysis #aerialmapping #Cloudcomputing
From the user #feedback. ⭐️⭐️⭐️
As we started experimenting with the segment-anything model in Mapflow we provided users with the "zoom" option because interpretation by the model depends on the input resolution and the window size.
Here is a good explanation of how the results may vary depending on the zoom level (zoom 18 - user rated ⭐️ VS zoom 16 - user rated ⭐️⭐️⭐️⭐️).
👉 BTW thanks to all who rate the results - we are trying to reach out to anyone if we know the way to improve.
Geoalert backend Team and CEO in the historical part of Khiva
Читать полностью…
Mapping scenarios featuring several AI models makes your analytics more accurate and cost-effective.
E.g. #Mapflow enterprise solution for #Telco maps leverages combo scenario for “3D building footprints” and “Forest with heights” models - https://geoalert.io/solutions/telecom.
Our clients also use other combinations like we told here about pvDesign use case.
Sharing more visual interpretations:
1. “Forest with heights” for fast and accurate research and design before the installation of solar panels.
2. Audit of already installed panels.
BTW. Total area of solar panels from the imagery - 9.53 Ha (Hall Farm Solar, Northwold, UK). 👆
#UrbanMapping – The Tashkent region construction areas dataset has been updated (github link). This detection was conducted using satellite imagery basemap and is mostly relevant to summer and autumn 2023.
Some extracts:
- The number of detected constructions has almost not changed compared to the year 2021 (around 500 objects covering the agglomeration area of 1000 sq.km)
- The biggest area, given 2023 layer is the construction of the new Uzbekistan's Olympic City which started in 2022.
We hope that there are many more insights that are meaningful for urban researchers and authorities.
Of course, the results are produced fully automatically and may contain errors (there are false positives, construction site contours are not as accurate as might be required...). But the most important that you can use and edit this data, and if you like to contribute back – write an issue and propose the updated vector data that was validated.
If you are interested in the more recent data and the data updates or validation for your favorite city – you are welcome to contact us.
Geoalert RnD #Team is growing. 🚀
Doniyor has joined us (UZ office) as a software engineer. Welcome on board! 🤚
#Mapflow #lab
Experimenting with AI mapping of rooftops.
Here we combined "Buildings" and "SAM Aero" ("segment anything") to get more details on the rooftops using aerial imagery of Australian residential area.
If you think it might be useful for your project you can use this or another custom AI mapping - contact us to join Mapflow mapping lab.
#Buildingfootprints + Vegetation = #Mapflow #custom
Here we tested Bing aerial imagery for some areas in Australia to map buildings and trees in a one-shot processing. The joined output has a valid topology and the mapping class is indicated as the "object type". Behind these experiments is a user story to produce 3D animations for infrastructure development.
BTW we seek more cases in which Mapflow models can be applied in different combinations. 😊
#Mapflow #How-to #Hint
How to use Nearmap as an imagery provider with Mapflow Custom URL.
Nearmap provides access to its Vertical and Panorama Imagery via a Tile API. If you have a subscription to their service you can use it easily with Mapflow Custom URL.
The URL format should be:
https://api.nearmap.com/tiles/v3/{tileResourceType}/{z}/{x}/{y}.{format}?apikey={YOUR_API_KEY}
Playing around with Q2threeJS for #powerline vegetation reports.
Read more about the solution for analysis of vegetation risks and work control near powerlines.
#QGIS #Mapflow #Solutions
A special combined scenario can be created to apply multiple models to the same image – to avoid many uploads/downloads and optimize processing time. The Geojson output contains the topology valid geometries with the attributes.
In this preview two scenarios were combined: Buildings with heights (aka 3D #Buildingfootprints) + Vegetation with height classes (low, mid, high).
You can find more about #Mapflow pipelines in our docs.
Presenting about Mapflow user feedback at the friendly coworking place in Istanbul #Team #Geoalert
Читать полностью…
#GeoQuiz Guess the city.
The right answer will come in the next posts. And the dataset as usually will be availalbe in the urban #mapping open repo.
Mapflow-QGIS Plugin 2.6.0 released!
A neat feature that was always presented in #Mapflow API is now promoted to plugin - you can now create and manage projects, which is useful to organize your processings when it comes to many items.
For every processing you can now download AOI to re-use it in your further research and visualization.
And you can use the "magnifying glass" button on the left menu to preview every Data source that is selected for the processing.
Another #research work featuring #Mapflow plugin for #QGIS at the stage of #buildingfootprints collection. This time a group of researchers from the Institute of #Renewable Energy leverages a machine learning approach for the analysis of PV installations in urban areas. The authors come up with the demonstration of ML technology as a cost-effective way to obtain data for decision-making in the fields of energy and urban planning.
While in this work the building footprints were combined with the DSM data available for the study area, it's worth noticing that Mapflow is capable of generating 3D building footprints (by predicting building heights with ±3m RMSE using mono images only).
One of the most interesting applications the research work refers to is to find out the cost-effective approach to define the best possible location for the solar panel installations given the rooftop area, its insolation capacity, and the other factors that are significant for estimation of the overall potential in the specific urban area.
Check out the publication. 👈