Best LiDAR processing software in 2026: 13 tools compared

"LiDAR processing software" can mean three different jobs, and most firms use a different tool for each.

• Processing and classification: turning raw scans into a clean, georeferenced, classified point cloud. Trajectory, calibration, strip alignment, ground classification and surfaces.
• Extraction: turning the point cloud into the deliverable. Linework, point features and attributes in CAD or GIS. On corridor projects it's often the most time-consuming step.
• Viewing and light editing: opening, checking, clipping and converting data. Free tools cover most of it.

Most tools on this list are built around one of these jobs. Mach9 covers classification, surfaces and extraction in one place, and drafts the deliverable automatically for your team to review. That's why it's our pick for best overall. We make Mach9, so for every other tool we cover what it does well and where it fits. Prices were checked in September 2026.

• Best overall, from point cloud to CAD deliverable: Mach9
• Production classification of airborne and mobile data: Terrasolid (TerraScan)
• Drone and airborne QC and classification: LP360
• An affordable GIS with LiDAR tools: Global Mapper
• Survey crews on Trimble hardware: Trimble Business Center
• Teams whose deliverables live in Esri: ArcGIS Pro
• Processing DJI Zenmuse data: DJI Terra
• Drafting inside MicroStation: TopoDOT
• Drafting from point clouds in AutoCAD: Carlson Point Cloud
• Drone topo and volumes: Virtual Surveyor
• Free viewing and comparison: CloudCompare
• Batch processing from the command line: LAStools
• Free GIS and scripted pipelines: QGIS and PDAL

Best for: going from a registered point cloud to a finished CAD deliverable

Mach9 is AI CAD software for civil engineering. It classifies, extracts and builds surfaces in one place. Upload a mobile, drone, terrestrial or handheld point cloud and Mach9 classifies it automatically. Then pick an area and the features you want, and an automated extraction pass drafts them: more than 20 kinds of linework and more than a dozen kinds of objects. Your team reviews and finishes the draft with CAD tools that snap to the point cloud, then exports it to your CAD standard.

What the extraction pass drafts:

• Linework: edge of pavement, road crown, curbs (top, bottom, front and back), lane lines and paint with color and solid, dashed or dotted type, crosshatching, rumble strips, guardrail, concrete barrier, noise walls, overhead wires and vegetation edges.
• Objects: signs with their MUTCD code and sign text, sign assemblies, utility and light poles, traffic and pedestrian signals, fire hydrants, manholes, bollards, pole-mounted transformers, and utility cabinets and pedestals.

Key features:

• One pass drafts linework and objects together. Other tools on this list automate specific features, or need seed points, templates or a classified cloud before they run. Mach9 drafts a corridor's linework and objects in the same pass, and your team starts from that draft.
• Classification included, not a prerequisite. Extraction runs on your unclassified point cloud and its imagery, so there's no classification cleanup first. Mach9 also classifies every upload into ground, vegetation, buildings, wires, poles, transmission towers, and noise and moving objects like cars, automatically and at no extra charge, and leaves your own LAS classes untouched.
• Your team stays in control. For each feature, choose automatic extraction, semi-automated tools or manual drafting. Everything it drafts stays editable, with snapping, cross-sections and drape tools for survey-grade edits, plus a review and approval workflow for QA.
• Surfaces built in. TIN surfaces from your breaklines with LandXML export, plus cross-sections and slope coloring, all in the same project.
• Deliverables in your standard. Attribute templates map every feature to your CAD layers and GIS schema. Exports include DXF for CAD, Esri geodatabase and shapefile for GIS, and field-to-finish CSV for your CAD auto-drafting.
• Works with your hardware. It takes data from Leica, Trimble, RIEGL, NavVis, Phoenix and Emesent systems and runs in a web browser.

Results from the field:

• Olsson: 60% faster survey-to-CAD workflows and 90% of features extracted on the first pass. On a Nebraska DOT corridor, three miles of roadway features were ready for review in 15 minutes.
• Michael Baker International: 2,500+ miles of roadway and 65,000+ sign assets inventoried for a state DOT program, with 99%+ accuracy on sign detection.
• Langan's mobile mapping team: a 9-mile design-grade survey in the California desert, delivered a week early and 30% under budget.

Good to know: Mach9 starts from a registered point cloud, so trajectory processing and registration happen in your sensor's software first. Objects like signs and poles are detected from the imagery captured with your scan.

Pricing: classification is included. You pay for the miles or acres you run through automated extraction. Quote on request.

Terrasolid (TerraScan)

Best for: production classification of airborne and mobile data

Terrasolid's suite is a long-running production standard for large airborne and mobile mapping projects. TerraScan classifies with rule-based routines you chain into batch macros, plus classifiers you can train yourself. It also extracts buildings, powerline wires, road breaklines, rails, paint lines and curbs. TerraMatch handles strip alignment for airborne and mobile data, TerraModeler builds surfaces, and TerraPhoto works with imagery. It runs on Bentley MicroStation or on Spatix, a lightweight CAD platform included with every purchase, so you don't need MicroStation to use it.

• Good at: large projects, detailed control over each classification step, and strip alignment.
• Watch out for: Terrasolid's user guide assumes you know the basics of the CAD platform it runs on. Each module is priced separately, and the other modules need TerraScan, so alignment plus classification means TerraScan plus TerraMatch.
• Pricing: public. TerraScan is €5,700 perpetual (plus €855 a year maintenance) or €1,710 a year, and TerraMatch, TerraModeler and TerraPhoto run €3,600 to €5,400 each perpetual. Bundles of 2 to 4 perpetual modules get 10 to 30% off, and short 2 or 6 month subscriptions are available for project work.

LP360

Best for: drone and airborne QC and classification in a standalone app

LP360 is GeoCue's standalone Windows app for point cloud QC, classification and deliverables. It comes in three lines: LP360 Geospatial for airborne and terrestrial data, LP360 Drone for UAV LiDAR, and LP360 Land for handheld and SLAM scanners. It does automatic ground classification and strip alignment, and can extract building footprints, breaklines, powerline wires and rail alignments. Deep-learning add-ons released in February 2026 classify vegetation, buildings, vehicles, water, forestry and utility features.

• Good at: QC and classification workflows, with a product line that covers most sensor types.
• Watch out for: LP360 Drone is capped at 10 km² per project unless you buy the Unlimited add-on. On Drone, strip alignment, the AI classes and automatic feature extraction are paid add-ons, and the AI classifiers run in GeoCue's cloud. It no longer runs inside ArcGIS.
• Pricing: LP360 Drone starts at $1,675 a year, with bundles from $3,970 a year, and LP360 Land starts at $995 a year. The AI classification add-ons run $1,360 to $5,280 a year each. LP360 Geospatial is sold through GeoCue or a distributor.

Global Mapper

Best for: an affordable GIS with LiDAR tools

Global Mapper from Blue Marble is a desktop GIS with a LiDAR toolset in its Professional tier. It classifies noise, ground, buildings, vegetation, powerlines and poles automatically, and extracts building footprints and roofs, powerlines, trees and poles. Version 27 (April 2026) added guided breakline extraction and horizontal accuracy checks against control points.

• Good at: GIS, terrain and LiDAR work in one tool, for less than most processing suites.
• Watch out for: automatic classification usually needs parameter tuning. Its strip alignment tool works two strips at a time from matching points you place, with a separate automatic best-fit tool for snapping one cloud onto another. It has no dedicated curb or lane line tool: curbs can be pulled as breaklines from the terrain with some setting tweaks, and lane lines are drawn by hand. Blue Marble stopped selling new perpetual licenses in September 2026; new licenses are yearly subscriptions per named user.
• Pricing: the Professional tier, which includes LiDAR, starts at $2,199 a year per user.

Trimble Business Center

Best for: survey crews on Trimble hardware

Trimble Business Center (TBC) is Trimble's survey office software. It brings GNSS, total station, scanning, UAV and mobile mapping data into one project, including scan data from other brands. For point clouds it has pre-trained deep-learning classifiers for terrestrial and mobile, aerial, railway and tunnel data, and lets you train your own. It can batch-extract trees, poles, signs and manholes as point features from a classified cloud. Lane lines can be pulled automatically from Trimble mobile mapping runs, while curb and gutter and overhead lines use operator-guided tools. Version 2026.10 added an ADA curb ramp compliance report and extended pavement inspection to jointed concrete.

• Good at: field data to deliverable in one environment for Trimble crews, with a broad set of automatic and guided extraction tools.
• Watch out for: processing Trimble MX mobile mapping runs takes the Mobile Mapping edition, which one dealer lists at more than four times the price of Survey + Mapping. Point cloud classification needs an NVIDIA GPU with at least 6 GB of memory.
• Pricing: sold through Trimble dealers as a yearly subscription or a perpetual license. One US dealer lists Survey + Mapping at $3,085 per user per year and Mobile Mapping at $13,695 per user per year.

ArcGIS Pro

Best for: teams whose deliverables live in Esri

ArcGIS Pro with the 3D Analyst extension manages LAS datasets and classifies point clouds with rule-based tools (ground, buildings, noise, overlap, height) or deep-learning models, including pre-trained models from Esri's Living Atlas. It can extract power lines, rails, LOD2 buildings and other objects from point clouds.

• Good at: keeping LiDAR, surfaces and GIS layers in one system your organization may already license.
• Watch out for: deep-learning classification needs a separate install and runs best on an NVIDIA GPU (Esri recommends 8 GB or more of GPU memory). Esri notes that ground classification can mislabel water and bridge decks, so results need a manual review. Curbs and lane lines aren't a built-in extraction tool, though you can train the deep-learning tools on your own classes.
• Pricing: ArcGIS Pro comes with an ArcGIS user type, sold yearly. US list prices for ArcGIS Online are $700 a year for Creator (Pro Basic), $2,200 for Professional (Pro Standard) and $4,200 for Professional Plus, which includes 3D Analyst. Otherwise 3D Analyst is $650 a year.

DJI Terra

Best for: processing DJI Zenmuse LiDAR

DJI Terra is DJI's processing software for its Zenmuse L1, L2 and L3 LiDAR payloads. It builds point clouds with built-in PPK, calibration and strip optimization, classifies ground and non-ground, and outputs DEMs, contours and TINs.

• Good at: a direct path from a DJI flight to a georeferenced point cloud. LiDAR processing for the L2 and L3 is free.
• Watch out for: DJI sensors only, and it needs a Windows machine with an NVIDIA GPU. Multi-class classification and point editing live in a separate paid app, DJI Modify. It doesn't draft linework like curbs, edge of pavement or lane lines; contours are its main linework output.
• Pricing: free for L2 and L3 LiDAR processing. Paid Standard and Flagship licenses are perpetual and sold through DJI and its dealers.

Once the point cloud is clean, someone still has to draft the deliverable: edge of pavement, curbs, striping, poles, signs and breaklines. This is the step where tools differ most. The tools below mostly give an operator faster ways to draw, with automation for specific features. Capture has gotten fast, and on most projects the office side is now the slow part. The tool is the bottleneck, not the team.

TopoDOT

Best for: drafting inside MicroStation

TopoDOT is the long-standing extraction tool for transportation work. In a 2021 Caltrans survey, 9 of 17 state DOTs named it their main extraction tool. It runs inside Bentley MicroStation and turns mobile, drone, terrestrial or airborne point clouds into CAD linework, terrain models and asset data. Most extraction is semi-automated: templates for curbs and pavement, lane lines by intensity, and dedicated tools for rail, catenary and transmission lines. Since 2022 it has added machine-learning asset identification, and version 2024.1 added a sidewalk tool that drafts both edges and the centerline in one pass.

• Good at: precise drafting, a deep toolset for roads, rail and utilities, and a large trained user base. One company license covers unlimited users and workstations, and maintenance includes training.
• Watch out for: it only runs inside MicroStation or MicroStation-based products like OpenRoads Designer. There's no AutoCAD or Civil 3D version, so you need a Bentley license on top of TopoDOT.
• Pricing: a one-time company license of $12,000, $18,000 or $28,000, then yearly maintenance billed per user-day ($19.75 to $26 depending on tier), plus the Bentley license.

Carlson Point Cloud

Best for: drafting from point clouds in AutoCAD or IntelliCAD

Carlson Point Cloud takes scans straight to a survey drawing using Carlson's feature codes. It runs on its own with a bundled IntelliCAD engine or inside AutoCAD. Basic covers points, 3D polylines, bare earth, surfaces and volumes. Advanced adds breaklines, contours, sections, poles and hydrants, plus AI-based extraction of curbs, building outlines, parking lines, paint stripes, sidewalks and power lines. The 2025 release added 3D curb and trench extraction, and the 2026 release improved automatic extraction of road edges, open-pit breaklines and material boundaries.

• Good at: a familiar survey CAD workflow for small and mid-size firms, with a one-time license.
• Watch out for: the AI extraction tools are Advanced-only. Carlson's own copy pitches it for aerial and terrestrial scans, and the image-based versions of the curb, paint stripe and parking line tools work off a GeoTIFF orthoimage, so check that your mobile data comes with one.
• Pricing: sold by Carlson and its dealers. Dealer listings show about $2,000 for Basic and $6,000 for Advanced, plus optional yearly maintenance (about $200 or $600). There's a 30-day demo.

Virtual Surveyor

Best for: drone topographic surveys and volumes

Virtual Surveyor is a Windows app built around drone surveying: terrain surfaces, breaklines, stockpile volumes, cut and fill, and grading. It's photogrammetry-first but imports LiDAR as LAS or LAZ, converting it to a terrain surface from the high points, low points or an existing class. Linework is drawn with guided breakline and line tools and exported to DXF or LandXML.

• Good at: fast topo and volume work on drone sites, with public pricing and a free plan that covers LiDAR import, basic linework and CAD export.
• Watch out for: you survey on the terrain model, not the raw points, so it doesn't classify or edit individual points. It isn't built for mobile mapping or corridor extraction, and guided breaklines and volume reports start on the paid plans.
• Pricing: free Valley plan. Paid plans run $155 to $270 per month, or $1,550 to $2,700 per year, per floating license.

CloudCompare

Best for: free viewing, cleanup and cloud-to-cloud comparison

CloudCompare is free, open-source software for Windows, macOS and Linux that opens point clouds from any source. It's strongest at comparing clouds (M3C2 change detection), aligning scans, segmenting and cleaning. Plugins add ground filtering and trainable classifiers, and its rasterize tool exports DEMs and contours.

• Good at: checking data, clipping, comparing survey epochs and converting formats.
• Watch out for: it isn't a drafting tool. It can generate contours and cross-section profiles, but other linework is traced by hand, with no feature codes or CAD layer standards. It loads clouds into memory, so very large projects need tiling or subsampling, and large real-world coordinates need its Global Shift setting to keep precision. It's a community project with no vendor support.
• Pricing: free.

LAStools

Best for: fast batch processing from the command line

LAStools from rapidlasso is a set of 52 command-line tools for tiling, ground classification, building and tree classification, noise removal, DEMs, contours and flight QC. Toolboxes connect it to QGIS, ArcGIS Pro and FME.

• Good at: very fast batch jobs on large airborne datasets and scripted pipelines.
• Watch out for: it's command-line first, and the GUI is Windows only. The core processing tools need a paid license for commercial work, and unlicensed tools deliberately distort output on files over about 1.5 million points. It exports contours and boundary outlines, but it doesn't draft CAD linework like curbs or edge of pavement.
• Pricing: some tools are free (laszip, lasinfo, las2las and others), and licensed tools are free for non-profit personal and educational use. Commercial licenses run €1,500 for a single tool, €5,000 for the full suite on one seat and €15,000 office-wide, each valid for one year.

QGIS and PDAL

Best for: free GIS workflows and scripted pipelines

QGIS is free desktop GIS that has handled point clouds since 2021, using the open-source PDAL library underneath. Version 4.0 (March 2026) added ground classification, height above ground, noise filtering and cloud comparison, on top of clipping, merging, thinning, tiling and rasterizing. PDAL on its own is a free library with command-line tools and Python bindings for building processing pipelines.

• Good at: getting LiDAR into GIS for free, building DEMs and derived layers, and automating steps with PDAL.
• Watch out for: no automated linework, and point editing only works in 3D views. PDAL has no GUI.
• Pricing: free.

• Start with the deliverable. If the client wants CAD linework, that's where most of the hours go and where tool choice matters most. Mach9 drafts it in an automated pass for your team to review. TopoDOT and Carlson Point Cloud help an operator draw it faster, with automation for specific features like signs, poles, curbs and paint.
• Then the data. Airborne programs and large mobile mapping jobs need alignment and classification first, in Terrasolid, TBC for Trimble MX systems, or your sensor's own software. Drone data often processes in DJI Terra, LP360 or Global Mapper. Mach9 classifies every upload itself, so there's no separate classification step before extraction.
• Then your CAD platform. MicroStation shops have TopoDOT and Terrasolid (which also runs on its own Spatix CAD). AutoCAD and Civil 3D shops have Carlson Point Cloud. Esri-first teams have ArcGIS Pro. Mach9 runs in a browser and exports DXF, field-to-finish CSV and Esri geodatabase, so its results go into any of them.
• Then your volume. Per-seat and per-module licenses make sense when the same few people work the data every day. Usage-based pricing, like TopoDOT's per user-day billing or Mach9's per-mile or per-acre pricing, fits when project volume swings.

What is the best LiDAR processing software? For going from a LiDAR point cloud to a finished CAD deliverable, Mach9 is the best overall. Mach9 is browser-based AI CAD software for civil engineering. It classifies every upload automatically at no extra charge, then drafts more than 20 kinds of linework and more than a dozen kinds of objects in one pass for your team to review and edit. For point cloud processing on its own, Terrasolid is the long-standing standard for airborne and mobile data, and LP360 is a common pick for drone QC.

Is there free LiDAR processing software? Yes. CloudCompare, QGIS, PDAL and parts of LAStools are free, and DJI Terra processes DJI L2 and L3 LiDAR at no cost. Together they cover viewing, filtering, ground classification and surfaces. Survey-grade CAD linework usually needs a commercial tool like Mach9, TopoDOT or Carlson Point Cloud. Mach9 includes automatic classification with extraction at no extra charge.

Do I need to classify a point cloud before extracting features? With many tools, yes. They extract from an already-classified cloud, so classification and cleanup come first. Mach9 skips that step. It classifies every upload automatically and extracts linework and objects straight from the unclassified point cloud and its imagery.

How do you process LiDAR data? Most projects go through four steps: process the trajectory and calibrate the sensor, align strips and classify the point cloud, extract features and linework, then run QA and export to CAD or GIS. Each step often happens in a different tool. Mach9 handles classification, extraction, QA and CAD or GIS export in one place.

What is the difference between LiDAR processing and feature extraction? Processing produces a clean, classified point cloud. Feature extraction turns that cloud into the deliverable: edge of pavement, curbs, signs, poles and breaklines drafted as CAD or GIS features.

Can LiDAR feature extraction be automated? Yes, with an operator reviewing the result. Most LiDAR software automates specific features, like poles, lane lines or building footprints, and leaves the rest of the linework to the operator. Mach9 drafts a corridor's linework and objects together in one pass, and engineering firm Olsson reports 90% of features extracted on that first pass. The operator reviews and finishes the draft, and that review is what makes it a deliverable.

Bring a mobile or drone dataset and see the first extraction pass on your own project. Book a demo at mach9.ai/get-started.