Since Waze was acquired by Google with a huge amount of money, another startup that made a big fuss about the crowdsourcing model has emerged. This company, called Mapillary, was founded in 2014 and can help users record their personal life trajectories. It can also work with other users to create a large-scale street view map, which will exceed the Google Street View map, and it has a tendency to challenge it. Not only that, more than three months ago, it teamed up with Esri to provide real-time street view images.

Different routes have the same goal, but the meaning is very different

In a sentence from Jan Erik Solem, co-founder and CEO, Mapillary is a popular version of Google Street View. To be precise, Mapillary is a mobile application that allows users worldwide to produce, contribute, and view geotagged Street View photos.

Google makes street view maps by itself. Why can the public provide photos and synthesize them into popular street view maps? So Jane Eric Solim and others tried to establish an open crowdsourcing map platform. Users can record various types of places and watch the world with a richer vision.

Obviously, Mapillary and Google have reached the same goal by different paths, after all, in order to present the real world of the Internet, but Google has provided street view services for many years. Why does Mapillary try to do the same thing? In fact, the two directions are not the same, and the meaning is quite different. ——Jane Eric Solim explained this. Google starts at the macro level, focusing on the busy arterial roads in the human transportation system, just like the aorta of the human body, Mapillary looks at the micro level, focusing more on the streets and lanes-places that make people feel deeper-focusing on providing higher Time resolution, higher image resolution, and more extended spatial information content are just like the human body (capillary)-in fact, the name of the Mapillary company evolved from this.

This is an application based on spatial information that can record life

A person takes a photo in a certain place and places the photo on the map where the photo was taken. This process is done automatically. To be precise, every time a photo is taken, the “EXIF metadata” in the photo file "Longitude and latitude are marked, and Mapillary's application makes use of this technology.

Nowadays, more and more photos are taken through smart phones. As long as the location service function can be enabled on the camera program on the phone, as long as the photo is taken, the image is geo-tagged, and then the photo can be placed on the electronic map. A certain place.

The focus of Mapillary is not to provide users with the ability to view photos in different locations, because Flickr and Panoramio have already been implemented. The Mapillary application can help you automatically and continuously capture massive amounts of photos. For example, when you are walking, a series of geo-tagged photos are generated, which will become a record of your journey, rather than a memento of a place.

Last fall, Mapillary held an online hiking competition, and then the winner completed an expedition to the Canadian forest. Why not use video to share such a journey? This is a breeze, because YouTube is flooded with such a huge amount of videos. However, the metadata used for geotagging on videos is not as standardized as the EXIF ​​format, but for Mapillary, geotagging information is important.

Now this situation has changed. When using the Mapillary app to take geo-tagged photos, the speed of taking photos can be modified. This is an important new feature released after the cooperation between Mapillary and the sports camera manufacturer GoPro. The mobile phone is connected to the neighboring GoPro4 camera via WiFi. The Mapillary application can record the coordinates when the photo was taken, at a recording frequency of almost every second, to achieve synchronization with the GoPro camera. At present, the program cannot be used for underwater shooting.

Everyone is a surveyor, the street view map may exceed Google

However, Solim explained that the importance of increasing the time resolution is far less important than ensuring the high resolution of the image. The high resolution is to achieve another goal of theirs: let these geotagged images serve as a data source for later 3D modeling and streetscape creation. In this way, everyone is a surveyor.

If users often use the Mapillary application to take photos of different angles at various locations on the street, through the participation of a large number of users, the coverage of the street view map created may be wider than that of Google Maps.

This can’t help but recall that when Microsoft released Photosynth, based on the object recognition and matching degree, many single photos were connected together to give users a coherent and free conversion visual experience. Computer scientists at the time also developed A digital point cloud-based algorithm was developed to reconstruct the city’s architecture. The University of North Carolina’s research has also made a similar attempt, using Flickr’s massive photos and the ultra-fast processing speed of computers to disclose how a person can build a digital three-dimensional city in six days.

What Mapillary does is to take on the post-processing task through the cloud, and let the public participate in it. For example, users can contribute their respective field of view through a traffic signal game, while the Mapillary application performs object recognition processing in the background. To facilitate the process of 3D modeling, the Mapillary application allows users to view digital point cloud images of uploaded images. However, the vision is likely to be disappointing, because upload photos in many areas are still limited and the generation will not be ideal. Only when a sufficient amount of users have contributed a wealth of photos, the generated 3D model can enter the normal use state, which can meet the needs of measuring space size and distance.

Everyone can use Mapillary's services for free, and the company plans to eventually make money by selling the license rights for the data generated by its users to various companies. Mapillary application shows the potential. It is reported that Mapillary has been recognized by many organizations around the world. For example, the Global Bank also used Mapillary's technology to process images from Dar es Salaam in Tanzania when dealing with flood disasters.

crowdsourcing model becomes the norm

Ten years ago, the crowdsourcing model of user-generated content was a prominent feature of the 2.0 Internet era, but it is no longer a trend or a phenomenon. Because the times have changed, traditional methods have been difficult to manage massive amounts of data. Individuals participating in the data collection process are no longer new. People are more and more aware of the benefits of participating: individuals can influence and control the timing, content, resolution, interval, and precision. And other content. The crowdsourcing model allows everyone to have control. Once he can gain a sense of achievement from it and is willing to actively participate in it, then who would choose to accept it passively?

People want to share data, not only because we are involved in data collection, but we are trying to find a common sharing mechanism: the public collects data and the data serves the public. Will Mapillary, which relies on the crowdsourcing model, become the second Waze? It is still difficult to make a conclusion, because this has already relied on a single model to package the world.