Introduction
With the advent of cheap computing power, widely available high-quality optics, and new developments in artificial intelligence models, there are increasingly endless ways to explore spaces with imagery. Using the location and imagery of Batticaloa as inspiration, a number of these methodologies were used both to capture spaces and imagine new ones. Additionally, a series of workshops were held to ensure that each of these techniques were widely available to all attendees and community members.
From the Air
Drone imagery was used to capture both Dreamspace Hive and the wider Batticaloa area through aerial mapping and photogrammetry. A DJI Mavic 2 Zoom was used to capture the images flown with automated flight plans created in DroneDeploy. Due to the location’s close proximity to Batticaloa International Airport, numerous flight restrictions needed to be observed. In particular, a 60 meter flight ceiling covered much of the area, greatly increasing the number of flights and time necessary for capture. Post-capture, DroneDeploy was used both to process the Dreamspace Hive facility, as well as the wider area that surrounds it. Dreamspace also possesses a DJI Air 2 drone, which could have been used, as well, however the automated flight plan feature of DroneDeploy would have been slightly more complicated to use with this particular model.
From the Ground
For ground imagery, a wide range of objects and spaces were digitized using two distinct methodologies. For most objects, images were captured on either a Google Pixel 6 or Apple iPhone 13 Pro smartphone. Once captured, the photos were uploaded to Polycam Web for processing and exported as object files (.obj).
Additionally, a second workflow was occasionally used to map spaces and large objects utilizing the LIDAR camera on the iPhone 13 Pro. In this workflow, Polycam for iPhone was used to capture both pictures and LIDAR readings. These captures would be processed locally on the phone, resulting in fast, wide-area captures.
From the Mind
Local images were brought to life, creating real animals from the region with Disco Diffusion and input from everyone at Dreamspace. Disco Diffusion, the tools we used, is an open source program that takes an image and text prompt as an input and outputs brand new images. In addition to exploring modifying local images, we also ran workshops on Disco Diffusion text-to-image ML and encryption using latent space to the Dreamspace and Dinacon communities.
Elephants from a local Buddhist temple brought to life
From the Past
Primary sources were used as the basis to transform current day Battilacoa into windows to its past, incorporating machine learning via Disco Diffusion, 3d modeling, and Unity.
Information from the local Muslim museum and the old Portuguese fort was used to recreate a scene of an old (1300’s) market.
Illustration from the Muslim Museum
3-D models and photographs taken from the area around Batticaloa were assembled into an interactive market scene using Unity. Visitors are invited to walk around the market, seeing both textures and images from current-day Batticaloa as well as relics from the local Museum. The market is placed under a large central tree like it would have been.
Early mockups:
