PiiBlur

Street-level imagery powers urban planning, navigation, and surveying. Mapping companies, surveying firms, and municipal agencies capture millions of photos every week — each potentially containing faces, license plates, and other personally identifiable information that must be redacted before publication.

Manual redaction cannot keep up. A single fleet of mapping vehicles generates thousands of images per hour. You need an automated pipeline that detects and redacts PII without human intervention and without delaying delivery.

This guide shows how to build that pipeline with the PiiBlur API.

Why Street-Level Imagery Requires Automated PII Redaction

Street-level captures are dense with personal data. A single panoramic image from a mapping vehicle can contain dozens of faces, several license plates, and visible street signs — all regulated under GDPR and CCPA.

The scale makes manual review impractical. A fleet capturing 50,000 images per day at 30 seconds per image would need over 400 person-hours daily. Automated detection eliminates this bottleneck.

Privacy regulations also demand consistency. Human reviewers miss things, especially under time pressure. An API-based approach applies the same detection model to every image, every time, reducing the risk of an unredacted face reaching your customers or the public.

What PII Appears in Street-Level Captures

PiiBlur detects 13 PII categories in images. Several appear frequently in street-level photography:

Faces — pedestrians, drivers, passengers, cyclists
License plates — parked and moving vehicles
Street signs — address numbers and directional signs that can identify locations of individuals
Screens — visible phone or tablet screens
Documents — visible text on papers, posters, or notices
QR codes and barcodes — on storefronts, delivery packages, or signage

Faces and license plates are the most frequent targets, but a thorough pipeline should cover all visible PII categories.

How to Build an Automated Redaction Pipeline

The PiiBlur REST API accepts images and returns redacted versions. A typical pipeline has three stages.

1. Ingest and Queue Images

As your capture devices upload raw imagery, add each image to a processing queue. Most mapping workflows already use a queue (SQS, RabbitMQ, Redis) for post-processing tasks like stitching or color correction. PII redaction fits as another step in that pipeline.

2. Send Images to the PiiBlur API

For each image in the queue, call the PiiBlur API with the PII categories you want detected. You can redact all supported categories or target only faces and plates.

The API supports both blur and pixelation. Most street-level providers choose blur for a cleaner result that blends with panoramic imagery. For a detailed comparison, see our guide on blur vs pixelation redaction methods.

3. Store Redacted Output

The API returns the redacted image directly. Replace the original in your storage layer or keep both versions, depending on your retention policy.

This pattern scales horizontally. Add queue workers to increase throughput. PiiBlur handles concurrent requests, so you can process hundreds of images in parallel.

Handling Scale: Throughput and Batch Processing

Street-level operations process tens of thousands of images per day. Keep these points in mind for high-volume pipelines:

Parallelize your requests. The API processes each image independently. Run multiple workers against your queue to match your upload rate.

Process during off-peak hours. If your vehicles capture during the day, schedule redaction overnight. Redacted imagery will be ready by morning.

Monitor for failures. Build retry logic into your queue workers. Network interruptions and transient errors are inevitable at scale. A dead-letter queue for repeated failures lets you investigate without blocking the pipeline.

Choose the right plan. PiiBlur plans range from $49 to $499 per month, scaled by volume. Review our pricing page to find the tier that fits your throughput.

Meeting Privacy Regulations for Mapping Data

Privacy laws across jurisdictions require removing PII from imagery before publication. The GDPR, CCPA, and similar frameworks all regulate personal data in public imagery.

Automated redaction meets these requirements by running every image through the same detection pipeline. For more on how image-level PII relates to GDPR obligations, see our guide to GDPR and image compliance.

Key practices for compliance:

Redact before sharing. Process images before they reach third parties, customers, or public-facing applications.
Document your process. Keep records of your redaction pipeline for audits.
Cover all PII categories. Regulations extend beyond faces and plates. Street signs, documents, and screens may also require redaction.

Integrating PiiBlur Into Existing Mapping Workflows

Most mapping and surveying teams already run post-processing pipelines for stitching, GPS alignment, and quality control. PII redaction slots in as another step.

If you use street-level imagery workflows with tools like Mapillary, KartaView, or custom capture platforms, the PiiBlur API integrates through standard HTTP requests. No SDK or proprietary tooling required.

The API accepts common image formats and returns the redacted result in the same format — no conversion steps, no format incompatibilities, and no disruption to your storage and delivery infrastructure.

Start Automating Street-Level PII Redaction

Manual redaction does not scale for street-level imagery. The volumes are too large, PII density too high, and regulatory requirements too strict for inconsistent human review.

PiiBlur's API delivers consistent, automated detection and redaction across all 13 PII categories. Build it into your pipeline, scale it with your fleet, and ship anonymized imagery with confidence.