AI + Human-in-the-Loop  ·  ArcGIS Parcel Fabric

From scanned plats to
Parcel Fabric-ready records

MTerra pairs purpose-built AI with Magnasoft's cadastral experts to turn scanned plats, historical maps, and new development records into validated, ArcGIS Parcel Fabric-ready parcel data — machine speed with survey-grade confidence.

Delivered by Magnasoft | Esri Partner Network | LADM-conformant output
01 · Scanned plat Original scanned subdivision plat — raw input document
02 · AI detection MTerra AI vectorization output — detected parcel boundaries and lot annotations
The Innovation

Two old options. Neither one scales.

For decades, counties have had only two ways to convert cadastral records — and the backlog kept growing. MTerra's breakthrough is to combine the speed of AI with the judgment of experts.

Pure manual digitization

One technician, one plat, hours at a time. Accurate, but the backlog grows faster than staff can clear it.

Pure automation

Fast on a clean scan. Breaks on faded plats and handwritten annotations, and the errors reach your Parcel Fabric unflagged.

MTerra: AI + human-in-the-loop

AI does the heavy lifting. Magnasoft's experts make the call.

1. AI detects

Boundaries, lot lines, COGO values, bearings, parcel numbers, symbols

Flags anything uncertain

2. Expert verifies

Magnasoft's cadastral team reviews, corrects, and confirms every flag

Nothing ships unverified

delivered to
Machine speed, survey-grade confidence
Why it's different: MTerra isn't automated tracing, and it isn't outsourced manual digitization. It's an AI + human-in-the-loop workflow engineered specifically for cadastral data — built to produce parcel records you can review, trace, and defend.
MTerra in Action

A real subdivision plat, end to end

A recorded subdivision plat — from scanned paper to production-ready parcels.

Original scanned subdivision plat — raw input document
Step 01 · Upload

Start with the scanned plat

Thousands of subdivision plats exist only as paper archives. A county uploads the scanned image — no pre-processing, no manual prep.

SOURCE · scanned subdivision plat (JPG / PDF)

AI vectorization output showing detected parcel boundaries and lot annotations
Step 02 · AI Extraction

AI detects every feature

Computer vision classifies boundaries, lot lines, points, textboxes, and COGO values automatically — flagging anything it isn't sure about for review.

DETECTED · boundaries · lines · textboxes · COGO

Human-in-the-loop expert verification of AI parcel detection output
Step 03 · Human Review

Experts review and correct

Magnasoft's GIS experts resolve flagged items — missing segments, undetected origins, unreadable values — directly in the editor before anything is committed.

RESOLVED · ORIGIN · LINE · TEXTBOX

Coordinate adjustment and QA output showing precise vector geometry
Step 04 · Adjustment & QA

Adjusted to the legal record

A least-squares cadastral adjustment aligns geometry to the legal plat dimensions. Legal values are the source of truth — never inversed from geometry — with F-test and W-test checks.

QA · least-squares · F-test ✓ · W-test ✓

Final ArcGIS Parcel Fabric output with attributed lot polygons
Step 05 · Parcel Fabric Export

Appended to production Parcel Fabric

Output is delivered as an LADM-conformant File Geodatabase and appended to the authoritative ArcGIS Parcel Fabric — lots with COGO dimensions, bearings, and stated areas, ready for every downstream client.

OUTPUT · ArcGIS Parcel Fabric · production-ready

Capabilities

What's inside MTerra

Seven AI-driven steps, a human review checkpoint, and Parcel Fabric-ready delivery — so the AI + HITL model is visible at every stage.

AI-Assisted Vectorization

AI

Converts scanned plats, field sketches, maps, and raster records into vector parcel data.

Boundary & Line Detection

AI

Detects parcel boundaries, connected lines, points, and geometry from image-based records.

COGO & Measurement Extraction

AI

Extracts bearings, distances, dimensions, and annotations used in cadastral reconstruction.

Parcel & Attribute Detection

AI

Identifies parcels, lot numbers, labels, and related attributes from the source record.

Symbol & Text Recognition

AI

Reads map symbols, text boxes, and annotations needed for accurate review.

Human-in-the-Loop Review

Human

Keeps trained analysts in the workflow — so the output is verified, not just machine-traced linework.

Least-Squares Adjustment

AI

Reduces measurement discrepancies and improves consistency across geometry and control points.

Coordinate Adjustment & Edge Matching

AI

Aligns parcels with accepted coordinate systems and matches adjoining boundaries for a cleaner fabric.

Parcel Fabric-Ready Output

Output

Verified parcel data delivered ready for ArcGIS Parcel Fabric.

Benefits

By the numbers

Days Minutes

Processing time per subdivision, vs. days of manual work

80–90%

Estimated reduction in processing cost

Accuracy

Geometry 100% true to legal records — driven by legal dimensions

  • A genuinely new approach. AI + human-in-the-loop replaces the old choice between slow manual digitization and unreliable automation.
  • Collapses processing time — days to minutes. What took one to several days per subdivision is completed in minutes.
  • Cuts cost by an estimated 80–90%. AI-assisted extraction turns days of digitization into hours of review.
  • 100% true to legal records. Dimension-driven geometry adjusted to the legal plat tolerances — never inversed from geometry.
  • Clears historical backlog. Systematically work down thousands of undigitized plats.
  • Supports active development. Move new plats and parcel changes into GIS before they become the next backlog.
  • Improves data quality. Automated COGO extraction eliminates manual-entry errors.
  • Strengthens defensibility. Traceable validation produces records you can review, trace, and defend.
  • Frees skilled staff. Analysts move from repetitive tracing to review and higher-value work.
  • Supports Parcel Fabric modernization. Structured, validated parcel data ready for ArcGIS Parcel Fabric workflows.
Built for your teams

One workflow, every land-records stakeholder

GIS Directors

A practical way to reduce parcel backlog and support ArcGIS Parcel Fabric adoption.

Parcel Fabric Teams

Validated parcel data that's cleaner before it reaches production.

County Assessors

More reliable boundaries for assessment, ownership, and tax workflows.

Planning Departments

Better visibility into new subdivisions, growth areas, and planned development.

Land Records Teams

A faster path from recorded plats and sketches to usable GIS data.

Survey & Cadastral Teams

Support for COGO values, legal dimensions, coordinate checks, adjustment, and geodetic validation.

Why Now

Today's new plats are tomorrow's backlog.

Counties manage old records and new growth at once, and the gap only widens with time. The tools that created today's backlog can't clear it — manual digitization is too slow, and automation alone isn't reliable enough to trust. AI + human-in-the-loop is the first approach built to do both at once, so land records, assessment, planning, permitting, and public GIS all draw on parcel data they can trust.

Get Started

Put AI + human-in-the-loop to work on your parcel records.

MTerra is delivered by Magnasoft — bringing AI-assisted plat digitization to ArcGIS Parcel Fabric. Let's talk about clearing your backlog and keeping your parcel system current.