Frequently Asked Questions

Folstag is a professional video data collection company based in Alberta, Canada. We capture custom AI training datasets using certified professional operators — farmers, heavy equipment operators, utility inspectors, and commercial drivers — in authentic field conditions.

Crowdsourced platforms aggregate random footage from unknown contributors. Folstag partners directly with certified professional operators to capture 100% custom datasets during real operations in authentic conditions, tailored to your exact specifications.

We serve six primary industries: autonomous vehicles, agriculture, construction, robotics, manufacturing, and energy/utilities. We can also adapt to other sectors requiring professional video data collection.

Folstag is based in Alberta, Canada. Alberta is our strategic advantage — it offers world-leading industries (oil and gas, agriculture, construction), extreme climate variation (-40°C to +30°C), and diverse environments from urban centers to remote rural areas. While Alberta is our primary base, we serve clients across North America and can deploy globally on request.

Yes. We operate on a 100% work-for-hire model. You retain full ownership of the raw data, all associated IP rights, and any derivative works. We claim no licensing or residual rights to your custom collections.

Yes. We provide comprehensive calibration files (standard YAML/JSON formats) for every sensor configuration. This includes intrinsic parameters for each sensor and extrinsic transformations relative to the vehicle/platform body frame.

Yes. Our datasets include synchronized inertial (IMU) and GNSS data suitable for localization, mapping, and SLAM validation. We can also deploy dual-antenna setups for high-precision heading accuracy.

We utilize hardware-level triggering and PTP (Precision Time Protocol) to ensure tight temporal alignment across cameras, LiDAR, radar, and compute units, minimizing temporal drift to negligible levels.

Yes. We execute targeted "scenario hunting" missions to capture specific environmental conditions, including adverse weather, unique lighting scenarios, complex road geometries, and specific traffic densities.

Yes. We deploy commercially licensed pilots for high-resolution photogrammetry, orthomosaicking, and LiDAR-based aerial surveys for 3D site reconstruction and topographic mapping.

Timelines depend on scope. Small captures: 1-2 weeks. Larger projects with annotation and processing: 4-8 weeks. We provide detailed timeline estimates in every proposal.

Every project is custom-quoted — pricing depends on capture hours, sensor configuration, annotation depth, and delivery requirements. After a discovery consultation, you receive a detailed proposal with scope, timeline, and pricing within 3-5 business days, with no obligation. Bundled packages (capture + annotation + processing) are typically more cost-effective than engaging each service separately.

We deliver annotations in all major formats: COCO JSON, Pascal VOC XML, YOLO TXT, KITTI TXT, TFRecord, CSV, and custom formats. Our proprietary Folstag Anno platform handles the annotation workflow, and we convert to whatever format your ML pipeline requires.

We deploy a range of professional-grade sensors depending on your requirements: high-resolution cameras (up to 4K/60fps), LiDAR systems (32-128 channel configurations), thermal imaging (FLIR-grade sensors), radar, and IMU/GNSS for localization. We also support synchronized multi-sensor arrays with hardware-level triggering and PTP for tight temporal alignment.

Edge cases are our specialty. We execute targeted "scenario hunting" missions to deliberately capture rare, high-impact situations — adverse weather, construction zones, unusual lighting, equipment malfunctions, and other long-tail scenarios specific to your operational domain.

Every dataset goes through a multi-tier QA process: initial annotation by trained specialists, peer review for accuracy and consistency, expert validation for edge cases and complex scenarios, statistical QA with automated checks for completeness and outliers, and client sample review before full delivery.

Folstag Anno is our proprietary desktop annotation platform built in-house for professional video and LiDAR data annotation. It supports multi-modal data (RGB video, thermal imaging, LiDAR point clouds, and telemetry), AI-powered automation (YOLOv8 detection, Deep SORT tracking, SAM 2 segmentation), camera-LiDAR fusion, and vehicle telemetry sync. It operates fully offline — your data never leaves the local environment.

Folstag Anno exports annotations in COCO JSON, YOLO TXT, Pascal VOC XML, KITTI TXT, and CSV formats. It also supports frame extraction with configurable image format, quality, and naming conventions, plus an export preview with statistics showing total annotations, frame coverage, and per-label breakdowns.

Folstag Anno syncs vehicle CSV telemetry data directly into the annotation feed — speed, steering angle, throttle, brake, GPS coordinates, IMU readings — all time-aligned with video frames. Annotators see exactly what the vehicle was doing at every moment, giving annotations rich operational context. This is critical for autonomous vehicle and heavy equipment datasets where understanding machine state transforms annotation quality.