# How KartaCam powers GrabMaps

[Grab](https://yomu.fyi/company/grab) · Shuangquan Hou · Dec 1, 2022

**Type:** Problem & solution

## Summary

Grab's Geo team required an efficient, low-cost way to collect fresh street-level imagery across Southeast Asia, where professional mapping equipment is prohibitively expensive and smartphone crowdsourcing yields inconsistent quality. To address this, Grab built KartaCam, a custom mapping device equipped with a 12MP sensor, dual-band GNSS, 4G LTE, and onboard edge AI. Edge machine learning models evaluate scene suitability, check image quality, filter for map-relevant objects, and blur personal data directly on the device prior to upload. Deploying a four-camera KartaCam 360 array delivers panoramic coverage and point-of-interest data comparable to commercial mapping rigs at roughly one-twentieth the hardware cost.

## Context

Traditional mapping hardware is too costly to scale across Southeast Asia and leads to outdated maps due to slow refresh cycles, whereas smartphone-based crowdsourcing suffers from inconsistent image quality, inefficient operation, and unreliable positioning.

## Approach / What changed

Grab engineered KartaCam, an edge AI camera with dual-band L1/L5 GNSS, 4G connectivity, and millisecond-synchronized four-unit panoramic setups, running on-device models to filter poor images, blur personal information, and detect map features before upload.

## Takeaways

- Edge AI algorithms run locally on KartaCam to discard low-quality, indoor, or rainy images and blur faces and license plates before any data leaves the device.
- Using a dual-band GNSS module supporting both GPS L1 and L5 frequencies alongside a tuned ceramic antenna significantly mitigates multipath interference and positioning errors.
- Benchmark testing in Singapore and Jakarta showed KartaCam 360 captures point-of-interest information comparable to professional mapping cameras that cost twenty times more.

**Tags:** [Machine Learning](https://yomu.fyi/topic/machine-learning), [Performance](https://yomu.fyi/topic/performance), [Privacy](https://yomu.fyi/topic/privacy)

[Read original post](https://engineering.grab.com/kartacam-powers-grabmaps)