← Back to writeups
Mapping where Google won't go
WRITEUP GEOINT CHRONOS

Challenge

Since 2012, MilSec Industries has been showcasing military projects powered by its artificial intelligence around the world. But at the time, the public had no idea what the firm was really up to.

Mission: find where and at what time this photo of the weapon was taken.

photo of the weapon to identify

Hint: the photo was taken on May 12, 2019.

Flag format: interiut{City_hh:mm:ss:ms} (example: interiut{Issou_13:56:52:123})

Research Methodology

We start with a reverse image search to get more information about the location.

reverse image search

In the first results, we can find the giant statue present in the background of the image:

giant statue identified

giant statue, second result

Next, a Google search about this huge event gives us more precise information:

search about the event

If we search "Tokyo giant gundam" on Google Maps, we get a precise address:

address found on Google Maps

Japan, 〒135-0064 Tokyo, Koto City, Aomi, 1 Chome−1−10 ダイバーシティ東京プラザ内 2Fフェスティバル広場

proof of the address

So we have an estimated location to investigate. However, the available StreetView images aren't identical:

inconclusive Google StreetView

Based on the challenge's title, "Mapping where Google won't go", it's possible to guess we need to use another source of StreetView/360 images. It's possible to find numerous alternatives to Google Maps with a simple Google search:

search for alternatives to Google Maps

reddit.com/r/degoogle - alternatives to Google Street View

list of alternatives found on Reddit

Here is a list of alternatives to Google Maps:

Since we have the location, we just need to explore the different sites offering StreetView mode. On Mapillary, we get these available StreetView images:

StreetView images available on Mapillary

Looking at them one by one, we find the image we're looking for:

mapillary.com - matching photo

matching photo found on Mapillary

Don't forget that the provided photo is a 360 view! By going into the options (the three little dots), we can download the image to get the full view:

downloading the full image

download options

We also find the time the photo was taken: 12:26:14.398

shooting time found

Flag

Celebration