Mapillary Blog

Our team has been hard at work redesigning the platform inside and out. These changes include new features, but also improvements under the hood to help support the future of Mapillary.

This month we're requiring all users to update their passwords to ensure compliance with our new policies.

In this story, Trek View shares their experience trying to map Africa's highest mountain during a time when international travel has been significantly curtailed.

The extent to which pedestrian data has been mapped in OpenStreetMap varies significantly. In this post we'll look at 5 distinct cities and how closely OpenStreetMap matches the reality on the ground.

By April 18th, 2021, it is critical that every Mapillary user has an updated email address that is fully functional and can receive important account emails. Read on for details.

Over the coming months we will be launching version 4 of our API and will be phasing out version 3. This post contains some key information for developers and how they can handle the transition.

2020 was a year many of us will never forget. Before 2020 gets too far behind in the rear-vision mirror, we'd like to reflect on some of the highlights for Mapillary and our community. We hope you enjoy these reflections.

Today we are making available Mapillary Vistas 2.0, a major semantic annotation upgrade for our street-level image dataset of 25,000 images from around the world. We are increasing the label complexity by almost doubling the amount of semantic categories, and additionally provide an approximate depth order of objects shown in the scenes.

Today, we’re introducing a new capability in the Mapillary plugin for JOSM, which allows for adding AI-detected map features using an accelerated workflow. Objects that were detected from Mapillary street level imagery can now be added to the map with just a few clicks.

Tesla vehicles have often been described as “computers on wheels”. They also include cameras for assisted driving features. In the Mapillary community, these features have raised the question whether the imagery from the car’s cameras could be collected and uploaded to Mapillary. In this post, we present a solution to do exactly that.

Over the last couple of years the Microsoft Open Data Team in Serbia has been very busy. They've contributed over 1,000,000 images on 8,000 km of roads. They've also validated 300,000 traffic signs and made more than 4,000 traffic sign related edits in OSM.

Finding and navigating Mapillary images in iD Editor has been significantly improved. Chris walks through these latest improvements.

Today we are announcing the introduction of our machine-learned Quality Score that allows you to filter images on mapillary by their quality.

The hardware landscape for street-level image capable cameras has improved substanially since Mapillary began. We'd like to share our latest thoughts on different camera types and the models we recommend.

Students in chapters across the YouthMappers network are finding innovative ways to use geospatial data and tools to improve Sierra Leone’s electrical grid.

Mapillary and OpenStreetMap contributor Peter Leth shares his experiences encouraging open data contribution and usage on the golf course.

Today, we are announcing the extension of machine-generated detections to scene classes. The new scene classes cover transportation infrastructure such as gas stations, toll stations, and parking lots and will help cities and community groups speed up their mapping efforts.

We recently made changes to how we handle image blurring for street level images. This post provides an update and explains those changes for our Mapillary users and contributors.

Anyone can now access Mapillary's map features directly via our API. Read on to see the ways you can filter and download map features programmatically.

Map features are now available in both iD Editor and JOSM. Let us know how they fit into your mapping workflows.

The Mapillary Planet-Scale Depth dataset is unique in its scale and diversity, thanks to the millions of images that our community uploads to Mapillary every day.

Today we're announcing that Mapillary has joined Facebook on the quest to improve maps everywhere and for everyone.

After almost 3 months of mapping, groups from around the world have submitted their Map2020 reports. One project has been selected to present at the 2020 Humanitarian OpenStreetMap Summit which will be held later this year.

Thanks to our community that verified over half a million machine-generated object detections, we’ve developed an efficient approach to object recognition that helps improve map data quality. We used the verifications to include partially annotated images in our training data, leading to much higher detection accuracy compared to using only fully annotated images. This is a scalable way to get diverse training data for developing models that perform well on real-world object recognition tasks globally.

Geocortex adds out-of-the-box support for Mapillary imagery and data in their latest Geocortex Essentials release, making it easier than ever for Geocortex's 1500 partners to use street-level imagery for building better GIS solutions.

Today we’re releasing the Mapillary Street-Level Sequences Dataset, the world’s most diverse publicly available dataset for lifelong place recognition. Mapillary Street-Level Sequences is one of our three papers that will be published at CVPR later this year.

Better capacity planning needs better maps. By adding hospital bed count to OpenStreetMap, we can all make a difference.

Boost your OpenStreetMap editing with new features in Mapillary's plugin for JOSM: more flexible filters, enhanced changeset tagging, support for organizations, and a number of UI and performance fixes.

As the world takes action against the Covid-19 outbreak, here is how you can help.

Mapillary and Humanitarian OpenStreetMap Team are launching a new mapping campaign, #map2020, to solve one of the greatest developmental challenges for low- and middle-income countries: maps and navigation. Here’s how you can take part.

Alban Vivert from Nomad Maps went on a six-month bike trip in Latin America. He captured his journey on Mapillary and used the imagery to edit OpenStreetMap. This is his story.

ArcGIS Urban, Esri’s new product for smart city planning, is now featuring a Mapillary integration. Bring street-level imagery side by side with 3D models of projects for a better understanding of how cities will evolve.

We spoke to hundreds of delivery drivers across the US to find out how maps impact their work on a daily basis. It turns out that broken maps are a $6-billion-problem to logistics companies in the US alone.

Jupyter Notebook is popular among geospatial engineers and it's often used in conjunction with platforms like ArcGIS. This is how you can make the most out of Mapillary data in Jupyter Notebook.

We're back with the first CompleteTheMap of 2020! Here's how you can take part.

Mapillary has long been integrated into OpenStreetMap editors, and we are evaluating how to increase the value of both the imagery contributed by our community as well as the data extracted from that imagery. As it stands, Mapillary’s world-renowned computer vision capabilities provide an excellent tool for fixing maps in an unusual way. Outside of Mapillary imagery, we are seeing machine learning applied toward satellite imagery and telemetry data to help map roads and buildings. The coupling of computer vision and street-level imagery, however, offers a better view of the detailed reality maps have often lacked.

In Latin America’s rapidly growing cities, policymakers have struggled to keep up with changes in public transportation—especially around private services that have developed spontaneously for decades. With tools like Mapillary, communities have the ability to help collect the data that highlights their own experiences and needs, allowing their governments to develop regulations that can ultimately make for safer and more environmentally friendly transportation systems.

Street-level imagery from the Mapillary platform is not only helping to build better maps all over the globe, it has also found its way into new and exciting research projects. We recently spoke with three researchers who have found innovative uses for the Mapillary platform in salmon habitat restoration, documenting the experience of a Victorian cemetery, and the fight against invasive pests.

In our latest work we reveal five key techniques for improving optical flow prediction — the task of estimating apparent 2D motion of every pixel in two consecutive images from a video. Our findings are the result of carefully analyzing shortcomings in existing works and thus help to improve a wide range of them. We quantitatively and qualitatively surpass the performance of directly comparable works and set new records on challenging optical flow benchmarks.

We are introducing a new way of doing 3D object detection from single 2D images. The architecture is called MoVi-3D and is a new, single-stage architecture for 3D object detection. Starting from a single 2D image, it uses geometrical information to create a set of virtual views of the scene where the detection is performed using a lightweight infrastructure.

The Mapillary platform now hosts one billion high-resolution images from around the world, all open and available for improving maps. With 2019 drawing to a close, we are looking back on some of the events that helped us reach this important milestone.

Access to high-quality training data is one of the most important requirements to push the boundaries with machine learning in computer vision. Today we’re unveiling our latest piece of research, where we roll out an entirely new way to generate training data for multi-object tracking and segmentation. The approach turns raw street-level videos into training data with unprecedented quality—even compared to results based on human-annotated data. By allowing machines to generate training data, the cost for training computer vision models can go down substantially. We validate our approach for multi-object tracking and segmentation and obtain new state-of-the-art results. Here is how.

From artificial intelligence for mapping in low resource settings to encouraging women to map, there were several reoccurring topics when Understanding Risk West and Central Africa and State of the Map Africa recently met in the Ivory Coast.

We have published 20 Mapillary datasets on the AWS Marketplace to make it easier than ever for AWS customers to access the map data they need.

We are building the first open dataset for maintaining and updating HD maps with Zenuity, AstaZero, RISE, and AI Innovation of Sweden. Together, we will collect map data in a highly controlled environment through cheap dashcams, lidar, and radar, in an effort to build a cost-effective way of updating HD maps and teaching autonomous vehicles to understand their surroundings through an HD map — even when the surroundings have changed remarkably.

We’re celebrating GIS Day by unveiling a new partnership with IWorQ, the cloud-based software provider that helps 1300 US municipalities to manage their assets, facilities, licenses, and more. Together, we will help local governments all over the US to capture street-level imagery and understand their streets and roads through computer vision. With the help of computer vision technology, municipalities will get access to more and better data, which will improve decision-making and cut costs.

Downloading the map data that you’ve subscribed to is now much more flexible. You can choose the object types you wish to download, which also means your download files will get prepared faster. There are two new file types available, and notifications for when your data is ready.

Mapillary is heading to ICCV for a week packed with activities. Here’s where you will find us.

August and September proved to be a busy and exciting time of year for Mapillary, with travel and presentations to some of the most important conferences in the open geospatial world. Our team was present in Bucharest, Minneapolis, and Heidelberg to talk about Mapillary’s role in fixing maps and growing the largest collect of global street-level imagery. Read on to hear about what we presented at State of the Map, State of the Map US, and FOSS4G!

Mapillary is releasing the world’s first end-to-end encrypted mapping dashcam to help delivery and ride-sharing fleets access the map and navigation data they need to serve their customers. The camera is light, flexible, and while it requires no extra work from drivers, it can capture up to 150,000 high-quality images in one eight-hour session. The images and the data are available on the Mapillary platform within hours of uploading. Finally, a quick and easy way for fleets to get the map data they need to fix maps.

Our newest feature in the iOS app lets you see the milestones you’ve reached as a Mapillary contributor, and the next goals that await.

Mapillary is teaming up with Siemens, Germany’s Federal Ministry of Transport and Digital Infrastructure, and others to make driverless streetcars in a self-operating depot a reality. The project takes place in Potsdam and will over the course of three years teach a driverless streetcar to get from A to B with the help of sensor fusion and street-level imagery that Mapillary is turning into map data to allow the streetcar to see.

Esri Education Manager, Joseph Kerski, explains how teaching with Mapillary is helping to train the spatial awareness of students who will go on to create the communities of tomorrow.

Verification projects help train the algorithms that identify objects in street-level imagery. More verifications mean more accurate detections, and that means better maps for everyone. Join us as we strive to complete one million verifications and compete for cameras and other prizes.

With Verification Projects, you can do quality review of the automatic object detections that underlie Mapillary’s machine-generated map data. In addition to verifying whether detections are correct, you can now also check images for missed objects, as we release a new type of task that’s now available in all verification projects. This will help improve our algorithms even more, enabling you to scale up mapping with high-quality map data.

Roughly 2 million images are uploaded to Mapillary every day. Mapillary’s computer vision algorithms automatically anonymize all images by blurring sensitive information like faces and license plates. Today we’re happy to reveal that our blurring algorithms are the best available for anonymizing street-level imagery. By uploading imagery to Mapillary, you get all the data you need without compromising on privacy.

Mapillary’s Community Operations Manager, Said Turksever, introduces us to his passion project—mapping accessibility issues and wheelchair access. Through ICT4Society, Said introduces young people from across Europe to open data and mapping tools. In the process, they build accessibility maps to help government officials make better decisions for their citizens.

Today we’re announcing hundreds of millions of new images being added to Mapillary. Following lvl5 moving on to new adventures, all images they collected through paid drivers are now coming to Mapillary—and made available to everyone.

Last summer, a team of researchers from the University of Chicago embarked on an ongoing project to map the Galápagos Islands using Mapillary. Annually mapping the islands with street-level imagery contributes to a better understanding of how people affect the natural environment and our very own Chris Beddow had the opportunity to assist the team this summer with their important work.

Every July, Esri hosts its annual User Conference in sunny San Diego, California. Rumors say that attendance was at a record level this year, and it certainly set records for Mapillary. Six members of our team gathered from five different cities around the world to represent Mapillary. Aside from the world-class team at our exhibition booth, our latest demos and case studies helped make for a very busy week talking to new and potential customers.

There were 33 projects submitted for the #map2020 challenge to build better maps in undermapped regions. Mapping teams joined in from Europe, Asia, and Africa for the chance to demonstrate how street-level imagery can play a role in addressing a humanitarian challenge. Two projects were selected to present at the HOT Summit in Germany.

Mapillary is a great tool for student mapping projects since the platform allows anyone, anywhere, to map what is important to them. Whether traveling over rugged mountain terrain or through the jungle, anywhere you can bring a camera, you can map with Mapillary.

Many businesses and organizations, ranging from governmental to geospatial and mapping, already take advantage of our global street-level imagery platform to provide solutions for their customers. Today, we are making it easier than ever by announcing the latest and greatest iteration of our partnership program.

At CVPR this year, Mapillary won two computer vision benchmarking challenges. We will always keep pushing the boundaries of what is possible in computer vision, and it is our award-winning models that allow us to produce the highest quality map data possible.

Today we’re releasing the Mapillary Traffic Sign Dataset, the world’s most diverse publicly available dataset of traffic sign annotations on street-level imagery that will help improve traffic safety and navigation everywhere. Covering different regions, weather and light conditions, camera sensors, and viewpoints, it enables developing high-performing traffic sign recognition models in both academic and commercial research.

Whether it’s OpenStreetMap, HD maps, our mobile navigation and map applications, the up and coming story is often the same theme: everything is about to change. Mailing list discussions and State of the Map presentations have both mourned and praised OpenStreetMap in this light. The arrival of cutting edge methods of collecting or extracting map data has left many wondering if there is space for a community and the touch of human hands.

From Ghana to Iraq, Zambia to the Philippines, we take a look at some of the 33 mapping projects that are taking part in #map2020 to build better maps in undermapped regions. The mapping participants are collecting street-level imagery to improve things like waste management, natural disaster response, and damaged roads, addressing some of the most pressing issues in undermapped regions.

Mapillary uses semantic segmentation to understand the contents of each image on the platform. We visualize the segments with different colors overlaid on the image objects. This is quite the challenge when done on 360° images, which are rendered as spheres in the Mapillary viewer. Let's take a look at our approach to solve the spherical visualization problem.

We sit down with Peter Kontschieder, the Director of Research at Mapillary, to talk about “Disentangling Monocular 3D Object Detection”, the latest academic paper to be published by Mapillary’s Research team. Peter tells us about how 3D object detections made in single 2D images have the ability to improve mapmaking and push down the cost of autonomous vehicles, and how the team unveiled a fundamental flaw in the metric used by the most dominant benchmarking dataset in this area.

Dashcams are the perfect type of camera for hours of imagery capture without needing much interaction from the driver. That’s why we developed the Mapillary edition of BlackVue DR900s, a high-resolution and high-performance dashcam, explicitly optimized for effortless mapping. This custom dashcam makes image collection more efficient while simplifying uploads, enabling both individual mappers and entire fleets to scale up fresh imagery collection for better maps everywhere.

Mapillary is a distributed team of 55 employees in eight different time zones and twice a year we all get together at company offsites. We are all mappers at heart, so after the last gathering Tobias Ollmann, one of Mapillary’s Computer Vision Engineers, decided his trip home to Graz, Austria would be the perfect time to capture imagery for the Mapillary platform. For this reason, he decided to ride his bicycle 1,500 km from Sweden to Austria, and even though all of the odds seemed stacked against him, he was determined to finish his ride.

In 2016, Geochicas was created by a small group of women in OpenStreetMap who noticed a structural issue in data communities due to the lack of female participation and project leadership. By promoting work that analyzes how women are represented in geospatial and technological spaces, they are helping to improve the overall diversity and quality of the data that goes into OpenStreetMap.

Mapillary has been processing street-level imagery with computer vision in order to extract traffic signs for several years, but in 2019 we’ve started to offer an expanded dataset. Point features such as fire hydrants, crosswalks, and manholes are also now available, with their location approximated on the map thanks to 3D reconstruction of scenes combined with segmentation of the images where these objects are recognized. To date, Mapillary has identified more than 191 million of these map features, and one of the results is GeoJSON data that can be used to enhance a map.

Today, we launch the world’s first demand-driven marketplace for generating street-level imagery and map data. Mapmakers, cities, and NGOs can post mapping requests that anyone can pick up and help complete. Mapillary’s collaborative contributor network already spans 190 countries and is open for everyone to join and work on map data projects to improve maps, cities, and mobility everywhere.

With our recently released Capture Projects, organizations like the City of Detroit have collected thousands of kilometers of fresh street-level imagery by having a team of drivers cover the area, task by task. Today we’re making Mapillary Driver, the app that drivers use for completing their tasks, available on Android. This improves access to Capture Projects as teams are now able to use a broader range of devices in the field to see what they need to map.

#CompleteTheMap is back for the Northern Hemisphere summer. We’re encouraging location submissions from around the world with the goal to capture dense street-level imagery within the challenge area. The top three participants will be awarded cameras for the contributions at the end of the challenge.

Mapillary and Humanitarian OpenStreetMap Team are joining forces to accelerate map data collection in undermapped regions. Local mappers are invited to use street-level imagery in mapping projects that address humanitarian or developmental challenges. Two of the projects submitted to the #map2020 campaign will be selected for a fully-funded trip to HOT Summit in Heidelberg, Germany this September.

When building a global platform for street-level imagery, there are lots of moving parts in the system that need to work together. At Mapillary, we pay a lot of attention to quality assurance and have built a testing framework that is reliable, developer-friendly, and scales as we keep developing the platform further. Here is how we do it.

Road safety is a big problem in the US and a key focus for states across the country. In a big step towards ensuring road safety for their residents, Departments of Transportation in Utah, Florida, Arizona, Connecticut, and Vermont have now all made their state photologs publically available on the Mapillary platform. These images cover 270,000 miles of roads and are automatically processed with computer vision to detect guardrails, traffic signs, and pedestrians’ interactions with vehicles at scale.

We've just enhanced the Mapillary viewer with the Combined Panning feature that lets you pan between overlapping regular images just like you would do with a full panorama, simply by dragging with your mouse. This provides a more natural and convenient viewing experience as you navigate among Mapillary imagery to look around in an area.

We just hit another milestone for our platform. More than 500 million images have been uploaded to Mapillary by our amazing community, customers, and partners worldwide.

With Verification Projects, map makers and GIS experts can review machine-extracted map data by adding human validation to the workflow. Comprising a project dashboard and a simple game-based tool for verifying object detections, Verification Projects enable quality assurance of the data available in any area, as well as help to improve Mapillary’s machine learning algorithms with human feedback.

Approaching the 500 millionth images on the Mapillary platform, we invited our community to guess when the 500 millionth image would be uploaded and enter the race to win a BlackVue camera and Mapillary swag.

Today we’re announcing that Mapillary will publish four papers at CVPR this year. In this post, we’re looking at the paper named Seamless Scene Segmentation, which, as a world-first, rolls out a new computer vision model that slashes up to 20% computing powers when teaching machines to distinguish between people, cars, and map data like traffic signs, together with its overall environment.

Toyota Research Institute (TRI) is focused on developing state-of-the-art machine learning algorithms for autonomous driving to realize safe and accessible mobility for the future. In this blog post, Jie Li, Research Scientist at TRI, outlines how TRI utilizes the Mapillary Vistas Dataset as a benchmark for driving scene understanding algorithms providing geometric and semantic diversity at scale.

Uploading thousands of images got even easier with the latest release of the Mapillary Desktop Uploader, which adds functionality for previewing images and managing your upload history. Spend less time handling your uploads and more time improving maps!

A lesser-known OSM tool, Deriviste allows creation of new point data by clicking in the Mapillary viewer, or referencing a point on the map with a marker in the viewer. In this post we’ll learn where it came from and how it works.

Today we’re announcing Capture Projects to help cities, mapmakers, and transportation agencies everywhere ramp up imagery and map data collection. With the release of a new project dashboard and an accompanying app, called Mapillary Driver, organizations can now manage an unlimited number of drivers to capture street-level imagery at any given point, rapidly escalating how quickly maps can be updated at scale.

As technology evolves, there is a growing need for more detailed information in maps. Curbs are a great example of a street asset that has a wide variety of purposes, which puts increased pressure on maps and the details they do—or rather, don’t—provide. In this post, Daniela Waltersdorfer J joins Mapillary's Christopher Beddow to explore the potential for OpenStreetMap to embrace the curb.

Since 2016 when we opened our AI lab in Graz, the research team has been busy publishing papers, winning benchmarking competitions, and developing the building blocks that power Mapillary. Now we are celebrating the opening of a brand new Graz lab and looking back at how it all came together.

Today we’re announcing 42 new types of map features that Mapillary extracts from street-level imagery and automatically positions onto the map. Ranging from utility poles and streetlights to mailboxes and manholes, this will help cities, mapping companies, and transportation agencies keep their maps up to date using cameras.

2018 is coming to a close, and it’s that time of year where we look back at some of the best moments.

The latest addition to the Mapillary toolset is a desktop application that lets you upload thousands of geotagged images with little time and effort.

After a few months in hiding, CompletetheMap is back.

With 2018 coming to a close, we are highlighting some of the outstanding community projects from this year.

The Mapillary community is spread across the world, with significant contributions coming in from Finland in the north, Argentina in the south, and—if we’re slicing the world in half by the international dateline—the USA in the west and Japan to the east.

After the labyrinth of Stone Town’s alleys and a ship wreck on the west coast, our story heads east. This blog post is a continuation of Pt 1 - The Chief Officers’ Logbook. If you haven’t read part 1 of our adventures in Zanzibar, start there. If you have, welcome to part 2!

Following FOSS4G 2018 in Dar Es Salaam, Mapillary's own Chris and Ed went on a mapping adventure across Zanzibar. This is part one of a two part series about their mission to help create one of the most accurately mapped areas in Africa.

In many one-person GIS shops across the United States, collecting and managing local assets can be a daunting task. For Steven Hewett of Clovis, New Mexico, Mapillary presented a way to streamline and automate the process of completing a city-wide traffic sign inventory.

The Mapillary mobile app SDKs enable anyone to build a street-level imagery capture component into their app with custom features. This makes it easier and more convenient for people to collect Mapillary imagery, which they can then use to update and enrich maps.

We have just launched an upgraded version of our website that features notable improvements (up to 80%) in page loading times. In addition, we’ve implemented the latest MapillaryJS release which helps straighten up distorted images and make transitions smoother between images. With this, we’re offering a much faster and better experience when using the platform.

With more than 200,000 km mapped, JB Brown has dominated the Mapillary contributions leaderboard since joining two years ago. We recently caught up with him to hear about what motivates him, and how he explains Mapillary and mapping when he gives members of the Amish community a ride in his van.

For this International Day for Disaster Reduction we are profiling some of the different ways people and organizations are using street-level imagery to spring into action before, during, and after disaster strikes.

This week marks five years since we started this little “project” of ours. It seemed controversial at first but now, five years later, it’s become a completely normal and sane thing. Time for a short reflection on this half-decade.

Radial distortion is a common problem when the 3D world is represented in 2D images. With the latest MapillaryJS release, we are now undistorting every image. This leads to a better 3D representation and improves the viewing experience of Mapillary images.

Map coverage in the Alps needs improving. That’s why Mapillary, HERE, and Garmin are offering prizes for mappers who increase map coverage in the given areas. The Alps are visited by over 120 million people annually, so up-to-date maps are important. Here’s how you can help (and how you win one of the prizes!).

Managing parking infrastructure is a billion-dollar problem for cities all across the US. There has been no easy way for cities and Departments of Transportation to access parking sign data, resulting in poor decisions around parking infrastructure and planning. Today, Mapillary and Amazon Rekognition are introducing a scalable way to help US cities get a handle of their parking infrastructure.

Earlier this year, COCO and Mapillary invited computer vision researchers to participate in a workshop co-located with this year’s European Conference on Computer Vision. Today we’re announcing the winners.

Mapillary just scored another success in computer vision benchmarking. This wouldn't be possible without our community: by contributing more than 350 million images from all over the world, they've created a dataset that helps build robust algorithms for making sense of any street scene that a self-driving car may encounter.

At Mapillary, we believe that the best way to visually represent our planet is through people and organizations working together. As of today, our community has collectively mapped five million kilometers!

The newly released extension of the Mapillary Vistas Dataset provides over 60 new object classes and even more granularity, including traffic light states. With a total of more than 150 classes, it remains the most diverse publicly available training dataset covering street scenes from around the world.

This week we're happy to welcome Till Quack as Mapillary's first VP of Product

Learn how our GIS customers use Mapillary for infrastructure maintenance, traffic sign inventory, and asset management—now with support for working with your organization on the Mapillary platform.

This is a guest post by a group of engineering students at the Costa Rica Institute of Technology, showing the importance of individual mappers in getting the less common perspective when evaluating mobility conditions in cities around the world.

We’re releasing an update to Mapillary’s traffic sign recognition, featuring wider support of traffic sign classes globally, improved recognition accuracy, and traffic sign taxonomy.

Mapillary has won the Semantic Segmentation Challenge as part of the 2018 CVPR Robust Vision Workshop. Here's how we did it.

Mapillary now supports organizations to easily create and share street-level imagery and automatically extracted data to keep maps and geospatial datasets up to date.

Citizens and the city council in Valencia, Spain, are working together to increase the safety, efficiency, and popularity of cycling. They use street-level imagery to provide a visual reference of the current situation and extract data for updating maps. This improves information exchange between stakeholders and helps the city government decide how to allocate the €4.5M budget assigned to developing bicycle infrastructure.

The updated Mapillary command line tools make uploading and preprocessing your imagery easier—and not just for programmers.

After the recent update of MapillaryJS, you can now compare any pair of images, regular as well as panoramas. It also features a simpler design to make MapillaryJS integrate smoothly in various applications.

Learn how you can add points of interest and road information in Map Creator with help of Mapillary imagery.

Put your mapping skills to the test and participate in HERE’s Game of Maps with awesome prizes!

At Mapillary, we are all in for research and innovation. In this guest post, Levente Juhász describes his academic research about Volunteered Geographic Information and asks for some help from our community.

Our partnership with HERE goes global to reinforce the power of community mapping.

Peter Leth, our active contributor and educational advisor for Creative Commons Denmark, represented Mapillary at the Creative Commons Global Summit. He shares his insights (and an unexpected meeting) in this blog post—published under the CC BY license, giving everyone the ability to share, copy, distribute and reuse it.

The second Global CompletetheMap Challenge will launch on 1 May. Nominate your location and compete with Mapillary contributors across the world—and win awesome prizes!

We have developed an accurate and fast algorithm for privacy blurring on street-level images. The new model, based on semantic segmentation, detects 99.9% of identifiable faces and license plates and is optimized for large-scale processing.

BMW i Ventures, Samsung Catalyst Fund, and NavInfo join our existing investors to support growing the world’s largest collaborative street-level imagery platform for extracting map data using computer vision.

To keep the competitive spirits up, we're publishing a series of monthly leaderboards. Take a look at who contributed with most images, edits, blurs, and verifications in March.

Knowledge and supervision from humans, speed and scale from machines. This is how we can keep up with the freshness, coverage, accuracy, and high level of detail that we require from our maps.

The Vermont Transportation Agency is the first State Department of Transportation (DOT) in the U.S. to bring their photo logs to Mapillary. Five million images collected over five years are now easy to integrate into the agency’s workflows, and also available to the public as open data.

Dan Joseph, GIS Officer at the American Red Cross, gives us an insight into using street-level imagery in disaster response and the challenges they face when time is critical.

GIS specialist Sanna Jokela from Mapillary's partner Gispo was invited to give a speech on AI and GIS. This is what she found out, exploring the unfamiliar topic of artificial intelligence.

To keep the competitive spirits up, we're publishing a series of monthly leaderboards. Take a look at who contributed with most images, edits, blurs, and verifications in February.

With the recent update of MapillaryJS, we have added controls for fast sequence navigation and better sequence overview. Now you can easily navigate to the beginning, end, or any other position of interest in a sequence.

The first edition of the Global Complete the Map challenge has drawn to an end with remarkable results despite the hardships of the winter in the Northern Hemisphere. Here’s a recap of what was achieved and what’s planned next.

Conversations at the first Mapillary GIS Meetup had a recurring theme. GIS work can include some cumbersome processes, but the experts working in the field have lots of creative solutions. Mapillary’s role is to provide the tools that help bring these to life.

To keep the competitive spirits up, we're publishing a series of monthly leaderboards. Take a look at who contributed with most images, edits, blurs, and verifications in January.

Mapillary for ArcGIS Pro is now available as a Public Beta with considerable performance improvements compared to the previous version. We invite you to join directly from the Mapillary for ArcGIS page.

Can you hike in unknown wilderness and make maps at the same time? Absolutely. Here is how I did it in Morocco last summer, gathering lots of details that are useful for the explorers coming after me.

The City of Turku together with Mapillary partner Gispo used Mapillary for reviewing school routes of students to discover potential safety hazards and other issues.

In this guest post, Anna Siprikova from Project for Public Spaces explains how they used Mapillary to make the process of analyzing public spaces easier, faster, and available to everyone.

Join us for an evening of GIS talk and demos to learn how street-level imagery can help automatize road and street inventories.

Autonomous Intelligent Driving GmbH has licensed the Mapillary Vistas Dataset to develop technology for self-driving cars across the whole Volkswagen Group.

Helsingborg, a forward-thinking Swedish city, makes visual geospatial data available on Mapillary for both staff and the public. Automatically extracted map features complement the city's geospatial records and help get an overview of Helsingborg's assets.

Mapillary Research ranks no. 1 for semantic segmentation of street scenes on the Cityscapes and Mapillary Vistas leaderboards.