Jack Sapoch Profile picture
Borders, osint, & digital investigations | working @LHreports; some past research @ForensicArchi | alum @ColumbiaSIPA | he/him

Nov 29, 2022, 17 tweets

1/ Dozens were lost at Melilla's Barrio Chino border crossing on 24 June, but why?

Today @LHreports & partners release an investigation reconstructing the event, making use of a detailed 3d model of the crossing. In this thread, I'll break down our development & use of the model

2/ Since the summer, we have been working with @el_pais @derspiegel @lemondefr @EnassMedia on this investigation on this horrific day on the Spanish/Moroccan border over the summer

3/ Step one in our investigation was archiving visual evidence. We collected 100+ videos from open source channels and received unpublished material from press & private sources.

Over several months, we geolocated the clips and sequenced them chronologically.

4/ Looking at the amount of material and the complexity of events, we saw the need for a detailed model of the crossing to assist with analysis, interviews, and visualisation

5/ LiDAR data from Spain's Instituto Geográfico Nacional served as the foundation for the model. By processing a point cloud of the Barrio Chino area in @CloudCompareGPL, we were able to produce .obj files for base terrain and building layers, which we then worked on in #b3d

6/ At 0.5 points/m2, the accuracy of the point cloud was useful but ugly when converted to mesh. We went in and manually refined the shapes of buildings

7/ Understanding the interior layout of the crossing was complex. It's a maze of fences, gates, and barbed wire.

8/ To better understand the layout of the crossing & surrounding area, we collected more reference materials. Old YouTube videos and new Google Streetview of the Spanish side were useful.

Most important were videos taken inside Barrio Chino by @bearamalhosilva over the summer

9/ After this, we were able to make use of the model for internal communication.

We uploaded a version of the model to @sketchfab to allow access from a web browser for use in meetings.

We also used renders for location referencing in interviews with survivors from the day.

10/ The model also allowed us to fine-tune our understanding of the filming location of various pieces of visual evidence

CW: violence

11/ Now in the publication phase, the model features prominently in our joint-investigation. We produced a video which utilizes various animated sequences in the model.

The sequence here shows Spanish drone recording compared to shots from @JaviBernardoP

12/ I haven't written much about our findings here. To understand these, it's better to read/watch through one of the publications attached to the investigation

Full video in English can be found here, along with links to all partner investigations:

lighthousereports.nl/investigation/…

13/ @el_pais published great pieces for the investigation in Spanish:

video: elpais.com/videos/2022-11…

text: elpais.com/espana/2022-11…

14/ And here is @lemondefr in French, with the @derspiegel piece out later today

lemonde.fr/international/…

15/ This model plays an important role in a larger investigation into what happened in Melilla on June 24th.

Without it would have been difficult understand what happened where, at what time & who’s responsible

16/ @alisonkilling @bash_deeb @arthur_wrb and I worked together closely on the visual reconstruction...

Share this Scrolly Tale with your friends.

A Scrolly Tale is a new way to read Twitter threads with a more visually immersive experience.
Discover more beautiful Scrolly Tales like this.

Keep scrolling