LiDAR 3D Volume Measurement: Turning Guessed Stockpiles Into Numbers
2026-10-01 · data as of 2026-10-12
LiDAR takes the ground as a reference plane, collects the stockpile surface as thousands of 3D points, and integrates the volume between that surface and the ground. When conditions are right, errors of around 1% are achievable.
Have you ever stood in front of a pile of raw material in the yard and wondered, "how many tonnes is that, really?" I get asked this a lot, and the answer is usually "about this much, from experience."
The trouble is that experience changes with the weather and the day. The pile looks different every morning, yet the number in the spreadsheet is still last month-s survey.
An estimated stockpile always costs you somewhere
The old method is simple. Someone climbs the pile with a tape measure, or approximates the shape with a few geometric figures. Because each person picks a different datum, the same pile can come out 10% apart.
Those numbers feed straight into inventory value and purchasing. Under-count and production stops; over-count and cash sits idle in the yard. That is why "roughly" is not good enough here.
LiDAR turns the pile into coordinates
LiDAR fires laser pulses, times their return to measure distance, and stitches those points into a surface. No one climbs the pile, and no work has to stop.
The key is the reference plane. The system first records the empty floor height, then integrates the volume between that plane and the measured surface. In dusty indoor silos where parts of the pile are hidden, several viewpoints are combined to fill the gaps.
How accurate does it really get?
A Purdue University team working with the Indiana Department of Transportation reported in 2022 that they could estimate bulk stockpile volume with under 1% error and in less than 10 minutes per facility. A highway pilot even set a target of staying within ±2% of the LiDAR value.
Accuracy is not a constant, though. In a drone LiDAR mapping study, error reached 11% when material was piled against walls and partly hidden; when the scan angle was changed to reduce that occlusion, error fell to 0.14%. Same sensor, same pile — only the setup changed.
A common myth: a better sensor means better accuracy
Many teams assume a more expensive LiDAR instantly buys 1% accuracy. In practice, error is driven by point density, occlusion, and how the ground reference is established. A sensor spec is a starting point, not a guarantee.
Volume accuracy comes from installation design, not from a datasheet. Deciding how many sensors, at what height and angle, matters more than the integration math underneath.
If you plan to trust the number enough to place orders against it, sketch out how you would mount the sensors first. Once that picture is clear, the rest is arithmetic.
Source page
https://www.melissasoft.com/en/volume