TerrAInav Sim: An Open-Source Simulation of UAV Aerial Imaging from Satellite Data

TL;DR

TerrAInav Sim introduces an open-source, geometry-grounded UAV imaging simulator that generates geo-tagged aerial imagery from satellite data by performing raster searches over user-defined bounding boxes. It computes image footprints from camera parameters ($FOV$, $AR$, $AGL$) and derives map zoom levels and pixel sizes to create scalable datasets, including a Memphis-focused TerrAInav dataset and an NVIG supplementary dataset for comparison. The framework supports single and raster missions, pre-processing with entropy-based filtering, and ML-ready data preparation, aiming to provide a controllable alternative to real flights and augment data for vision-based navigation, environmental monitoring, and urban planning. Limitations include temporal metadata gaps from the Google Maps Static API and idealized sensor modeling, with future work proposed on incorporating time-stamped imagery, sensor noise, and multi-angle capture to enhance realism and applicability.

Abstract

Capturing real-world aerial images for vision-based navigation (VBN) is challenging due to limited availability and conditions that make it nearly impossible to access all desired images from any location. The complexity increases when multiple locations are involved. State-of-the-art solutions, such as deploying UAVs (unmanned aerial vehicles) for aerial imaging or relying on existing research databases, come with significant limitations. TerrAInav Sim offers a compelling alternative by simulating a UAV to capture bird's-eye view map-based images at zero yaw with real-world visible-band specifications. This open-source tool allows users to specify the bounding box (top-left and bottom-right) coordinates of any region on a map. Without the need to physically fly a drone, the virtual Python UAV performs a raster search to capture images. Users can define parameters such as the flight altitude, aspect ratio, diagonal field of view of the camera, and the overlap between consecutive images. TerrAInav Sim's capabilities range from capturing a few low-altitude images for basic applications to generating extensive datasets of entire cities for complex tasks like deep learning. This versatility makes TerrAInav a valuable tool for not only VBN but also other applications, including environmental monitoring, construction, and city management. The open-source nature of the tool also allows for the extension of the raster search to other missions. A dataset of Memphis, TN, has been provided along with this simulator. A supplementary dataset is also provided, which includes data from a 3D world generation package for comparison.

Figures (10)

• Figure 1: Image captured at coordinates (35.128039, -89.799163) at above ground level altitude ($AGL$) of 126 m. Image captured (a) by a UAV and simulated (b-d) with a field of view ($FOV$) of 78.8 degrees and aspect ratio ($AR$) of 4:3. Figures b-d represent simulation using three different maptypes.
• Figure 2: Raster mission visualization within UTM zone 16, provided bounding box coordinates.
• Figure 3: The flowchart to process the map in downloading picture/s, using various missions. Each block includes the name of the function used available in the "src/utils" folder within the "geo_helper" package in parantheses. The "Download Raster" Section refers to the flowchart in Figure \ref{['fig:rasterflowchart']}.
• Figure 4: Raster search image capture flowchart.
• Figure 5: Satellite samples. The image with a red border indicates no significant feature.