PIX4D 'Closes the Loop' with Gaussian Splatting in PIX4Dmatic. The addition of Gaussian splatting to PIX4Dmatic connects mobile capture, desktop, and cloud processing into a single georeferenced pipeline, and the benefits go well beyond visualization. Gaussian splatting has been building momentum in geospatial circles for a while now, mostly as a visualization technique. The renders are visually striking and high fidelity, and they are gaining quick ground among professionals in many fields for their ability to visualize challenging scenes. But for professional survey workflows, a good-looking model is not enough on its own. Accuracy, reliability, and clean data at every stage of the pipeline are what actually matter to the people on the ground. PIX4D's latest update to PIX4Dmatic, announced on a recent blog, takes a different approach to Gaussian splatting than the broader industry conversation might suggest. Rather than positioning it as a visual upgrade, the company has integrated the technology as a structural improvement to its processing pipeline, one that reshapes how data quality flows from initial reconstruction all the way through to final deliverables. A georeferenced pipeline, start to finish. The context here matters. PIX4D already offered Gaussian splatting in PIX4Dcatch, its mobile capture app, and in PIX4Dcloud, its cloud-based processing platform. What was still waiting was the connection to professional desktop control, and now PIX4Dmatic fills that gap. These results can be shared to PIX4Dcloud via a Share to Cloud option. The entire path from field capture to processed deliverable now runs through a connected, georeferenced ecosystem rather than a set of isolated tools. That is the loop being closed. For survey professionals, this is a meaningful shift. Bringing mobile and aerial datasets together has often required moving data between separate platforms, sometimes with consistency compromises at each handoff. A unified pipeline would reduce those friction points and keep the reconstruction logic aligned throughout. Better point clouds, not just better pictures. One of the more underappreciated consequences of Gaussian splatting in PIX4Dmatic is what it does to the dense point cloud. In traditional photogrammetry workflows, the point cloud comes out of a densification process that can struggle with low-texture areas, reflective surfaces, and complex geometry. Gaps, noise, and uneven density in these areas are common, and they require cleanup time before the data can move forward. Gaussian splatting builds the reconstruction on a more continuous representation of the scene. The resulting point cloud is more evenly distributed and less affected by the noise and inconsistencies that typically accumulate in those challenging zones. According to PIX4D, this is not simply a visual improvement. It translates into a dataset that requires less correction and carries more reliability into downstream stages of the workflow. This is true for orthomosaicas as well, which are the most sensitive outputs in any photogrammetry pipeline because they will show every weakness in the reconstruction, particularly along water edges, reflective materials, and low-texture ground. Gaussian splatting, as implemented here, can provide a more continuous underlying scene representation, helping the orthomosaic stitching to come out sharper and more complete. Surfaces that behave, meshes that hold up. The improvements from the addition of Gaussian splats can propagate upward through the pipeline. On its face, a 3D mesh is only as reliable as the data it is built from, and considerable time in many workflows is used to stabilize mesh surfaces, resolve irregular triangulation, smooth transitions, and correct artifacts introduced during reconstruction. With a more coherent dataset at the base, that cleanup effort is significantly reduced. PIX4D describes the mesh in PIX4Dmatic as behaving more predictably as a result: triangulation becomes more consistent, transitions are smoother, and edge conditions are better defined. Subtle inconsistencies that might not be visually obvious but can affect measurements are less likely to propagate through. Volume comparisons and the cost of noise. PIX4Dmatic's new surface comparison tool adds another dimension to what these improvements make possible. The tool allows users to compute volumes by directly comparing two surfaces to quantify change between surveys, track progress against design, or compare as-built conditions to what was planned. This kind of analysis has always been highly sensitive to data quality. Small inconsistencies in reconstruction could show up as apparent (but false) changes or hotspots, which are particularly costly in workflows like earthworks monitoring, stockpile tracking, or infrastructure progress assessment. Gaussian splatting, as applied here, reduces those inconsistencies at the source so the surfaces being compared align more coherently. What this means in practice is that surveyors spend less time checking whether a detected difference is real or an artifact. Performance and the hardware equation. Generating Gaussian splats at this quality level requires computing power. PIX4Dmatic processes locally through GPU resources, and PIX4D recommends modern NVIDIA GPUs with sufficient VRAM to get the most out of the feature. For teams with constrained local hardware or projects that need to scale, processing can be offloaded to PIX4Dcloud without disrupting the workflow or compromising consistency. What it means for the industry. The broader significance here is about how reconstruction quality improvements travel through a pipeline. When the foundational representation of a scene gets better, it does not stay isolated at the visualization stage. It changes the point cloud, the orthomosaic, the mesh, and the volumetric analysis, and it does so in ways that reduce the manual intervention required at each step. PIX4D's integration of Gaussian splatting into PIX4Dmatic, available as part of PIX4Dmatic Pro, is designed for workflows that require the highest level of reconstruction quality. For survey professionals already working within the PIX4D ecosystem, it completes a connected pipeline that keeps fidelity and georeferenced accuracy intact from capture through to final output. May 8, 2026 May 6, 2026 Carla Lauter is a geospatial media professional covering surveying technology, earth observation, lidar, and AEC innovation for Geo Week News. She brings more than a decade of direct experience working on NASA and National Science Foundation-funded Earth science projects, including satellite missions measuring sea level, salinity, and hyperspectral ocean color. As Senior Content Manager, she develops editorial and conference programming for Geo Week, connecting practitioners across the geospatial industry with the tools, research, and people shaping the field. Get in touch: Join the Discussion

Esri and Pix4D launch real-time terrestrial mapping workflow. Alliance Brings Seamless Integration of Asset Data into ArcGIS for PIX4Dcatch Users * Pix4D joins with Esri to launch new terrestrial scanning workflow for infrastructure-focused organizations. * Esri and Pix4D workflow lets field teams capture accurate 3D models and augmented reality using PIX4Dcatch app data integrated directly into ArcGIS Online in real time. * New integration enables users to document subsurface assets with seamless precision. * Workflow brings high-fidelity, georeferenced documentation and instant verification to the project lifecycle for improved accuracy. * Explore the PIX4Dcatch app and publish 3D data to ArcGIS Online. REDLANDS, Calif.-BUSINESS WIRE- Esri, the global leader in geographic information system (GIS) technology, has launched a terrestrial mapping workflow with Pix4D. A Switzerland-based photogrammetry software company, Pix4D specializes in mobile reality capture and site-digitization. The combined workflow allows field teams to capture and process asset data and manage it within ArcGIS, Esri's geospatial platform. Field teams use the PIX4Dcatch app with real-time kinematics (RTK) devices to scan trenches and infrastructure. The generated high precision georeferenced records can now be published into Esri's ArcGIS Online. Scans become 3D models and point clouds, viewable as a Scene Layer. Existing data can also be visualized in augmented reality for instant "as-designed" versus "as-built" verification before closing trenches. "We are focused on turning hidden infrastructure into functional data," said Andrey Kleymenov, CEO of Pix4D. "By connecting PIX4Dcatch with a compatible RTK device to the Esri ecosystem, we are enabling a terrestrial-scanning workflow that allows teams to update a master geodatabase with professional accuracy." "We are happy to work with Pix4D to offer infrastructure-focused organizations a seamless, direct path from the field into a geodatabase," said Thomas Fair, Director, Esri Partner Network. "This new integration with ArcGIS will provide users with high-fidelity documentation of subsurface assets." About Esri Esri, the global market leader in geographic information system (GIS) software, location intelligence, and mapping, helps customers unlock the full potential of data to improve operational and business results. Founded in 1969 in Redlands, California, USA, Esri software is deployed in hundreds of thousands of organizations globally, including Fortune 500 companies, government agencies, nonprofit institutions, and universities. Esri has regional offices, international distributors, and partners providing local support in over 100 countries on six continents. With its pioneering commitment to geospatial technology and analytics, Esri engineers the most innovative solutions that leverage a geographic approach to solving some of the world's most complex problems by placing them in the crucial context of location. Visit Connectweb at esri.com. Contact details: Jo Ann Pruchniewski Public Relations, Esri

Esri and Pix4D launch reality capture workflow. Integration between PIX4Dcatch app and ArcGIS Online enables users to document subsurface assets with 'seamless precision'. Esri has developed a terrestrial mapping workflow with Pix4D, which allows field teams to capture and process infrastructure asset data using the PIX4Dcatch smartphone app then manage it within ArcGIS, Esri's geospatial platform The PIX4Dcatch app, coupled with a real-time kinematics (RTK) device, is used to generate 'high precision' geoferenced 3D scans of trenches and infrastructure models, which are then viewable as a Scene Layer in Esri ArcGIS Online. Existing data can also be visualized in augmented reality for instant "as-designed" versus "as-built" verification before closing trenches. "We are focused on turning hidden infrastructure into functional data," said Andrey Kleymenov, CEO of Pix4D. "By connecting PIX4Dcatch with a compatible RTK device to the Esri ecosystem, we are enabling a terrestrial-scanning workflow that allows teams to update a master geodatabase with professional accuracy." "We are happy to work with Pix4D to offer infrastructure-focused organizations a seamless, direct path from the field into a geodatabase," said Thomas Fair, Director, Esri Partner Network. "This new integration with ArcGIS will provide users with high-fidelity documentation of subsurface assets."