SAIMOS swaps 180 stereo cameras for 12 Hesai XT32 lidar sensors in retail people-flow project
A retail people-flow analytics system planned around 180 ceiling-mounted stereo cameras went live with just 12 Hesai XT32 lidar sensors, cutting hardware, installation and network requirements.
- 12
- 180
- €200,000
- better than 10 cm
What Happened
SAIMOS runs a 3D lidar analytics platform built around its SAIMOS Control Center, a geo-database and geospatial-analytics capabilities. For one large retail environment, the original design relied on a large network of ceiling-mounted stereo cameras to cover corridors, open entrances, architectural elements and changing layouts. As the project developed, planned ceiling elements would have obstructed some of the intended camera views, and the overall number of cameras pushed hardware, installation and infrastructure requirements beyond the original budget. Following consultation with SAIMOS, the sensing architecture was redesigned around 3D lidar rather than simply swapping one sensor type for another.
- 180 stereo cameras planned; 12 Hesai XT32 lidar sensors installed
- Four XT32 sensors providing seamless coverage on each floor of approx. 110 m x 20 m
- Minimum €200,000 saving versus the originally planned camera system
€200,000
Compared with the originally planned stereo-camera system, including sensing devices, software licences and servers; further savings came from installation, cabling and network infrastructure.
- People counting across the retail environment
- Dwell-times within defined areas
- Heatmaps showing where activity is concentrated
- Statistical analysis of visitor movement over time
- Real-time and historical dashboards plus a map-based interface with detail for individual areas
- Geotemporal queries on how different areas have been used as layouts change
According to Hesai, 3D lidar enables positional accuracy of better than 10 cm when detecting people, maintaining consistent performance across relevant sensing distances, which supports accurate tracking of movement paths and reliable dwell-time measurement. Because four XT32 sensors cover each floor seamlessly, temporary promotional booths and other installations can be moved or replaced without repositioning the sensors. The SAIMOS platform also requires no changes to its heatmapping, counting or retention-time configuration when the layout changes, since spatial data is continuously stored in the geo-database and analysed across both time and location.
SAIMOS said the XT32 offered a strong balance between value and point-cloud density, providing the level of 3D data the analytics application requires while keeping network traffic within practical limits for the overall architecture. The design was built around the spatial data needed for reliable people-flow analytics rather than maximising point-cloud density. Reducing the number of sensing positions also lowered requirements for network infrastructure and cut installation effort related to cabling, switches and other supporting hardware. The system has been live since Q1 2024 and continues to generate real-time and historical spatial data as the retail environment changes.
Previously from Hesai Technology
Earlier Motorclaw coverage described a Visimind IDL device built around Hesai's XT32 lidar and SLAM that mapped forklift routes across 12 factory units, replacing a process where a person followed forklifts with pen and paper. Hesai has also explained how lidar helps warehouse robots, farm machines, drones and humanoids read the physical space around them rather than simply follow a route. Both are indoor, non-road applications of Hesai's sensing technology.
- Hesai lidar and SLAM power Visimind's IDL device to map forklift routes in 12 factory units
- Hesai explains how lidar helps warehouse robots, farm machines, drones and humanoids read physical space
Background drawn from MotorClaw's earlier coverage of Hesai Technology's official releases.
Why this matters
The project shows that the number of sensing devices drives more than hardware cost: every extra installation point also needs mounting, cabling, network connectivity, configuration and ongoing support. SAIMOS reports at least €200,000 in savings versus its originally planned camera system, and says the lidar layout keeps working as temporary retail spaces are moved.
Terms in This Story
- lidar
- A sensing method that measures distances by bouncing laser light off objects to build a three-dimensional picture of a space.
- point cloud
- A set of 3D data points, each marking a position in space, that sensors such as lidar produce to represent the shape of an environment.
- dwell-time
- The length of time a person stays in a defined area before moving on.
- heatmapping
- A visualisation that uses colour intensity on a map to show where activity is concentrated.
Summarised from the linked release; details can be imperfect — always verify against the original source.