How to Automatically Calculate Embankment Volume from 3D Point Clouds｜Comparison of Accuracy and Time Compared with Conventional Methods

Volume calculation is an indispensable task in the management of embankment construction, but traditional hand calculations or simple CAD operations have always faced limits in accuracy and placed a burden on time. 3D point cloud technology offers an innovative solution to these challenges. A 3D point cloud is the countless set of point data in space acquired by laser scanning or photogrammetry, and by properly processing this data, highly accurate and efficient volume calculations that were not achievable with conventional methods become possible. This article explains in detail the specific process of automated calculation using 3D point clouds, a comparison of accuracy with traditional methods, and practical applications in the field. We provide all the information practitioners need when considering the adoption of this technology.

Automated calculations using 3D point clouds are not merely an advancement in measurement technology; they bring a fundamental change to the entire business process of embankment volume management. Whereas it was common for the process from measurement to final reporting to take several weeks, leveraging 3D point clouds can significantly shorten that timeframe. In addition, because measurement data is stored in digital form, subsequent verification and additional analysis become easier, and management transparency is greatly improved. Multiple stakeholders can access the same data and enable real-time information sharing. This directly leads to a qualitative improvement in overall project management.

Limitations and Challenges of Conventional Embankment Volume Calculation Methods

Traditional embankment volume calculations were mainly divided into two methods. One is a mathematical calculation method based on on-site actual measurements, and the other is a method of geometric processing using CAD software. In the former, measurement-based method, multiple cross sections are established on site, the cross-sectional area of the soil at each section is determined by surveying, and those areas are used to calculate the volume. Specifically, the average cross-section method is common: the volume for a segment is calculated by multiplying the average of two adjacent cross-sectional areas by the distance between the sections. This method is a traditional and highly reliable technique that has been used for decades.

The biggest challenge of this method is the limitation on the number of measurement points. Measurements taken in the field require manpower and time, so the feasible number of measurement points is limited to several dozen to several hundred. As a result, fine topographic variations between measurement points can be overlooked, creating a risk of systematic errors in the final volume calculation. In addition, differences in the skill of the surveyors can cause variability in measurement accuracy. Particularly in complex terrain, measurement results can vary greatly depending on where cross-sections are set, and this subjective judgment affects accuracy. For example, accurately capturing minor changes in valley topography requires many measurement points, which is not realistic under time constraints. In large-scale embankment projects, this limitation becomes a serious obstacle. Even when contractors measure the same area, the results can vary, which becomes a quality-control issue.

In the method using CAD, measurement data is entered into the CAD system, cross-sectional drawings are created, and then volumes are calculated. This method offers high reproducibility because the measurement data are clearly recorded, but it does not resolve issues that occur during the measurement stage. In addition, geometric processing in CAD requires operator skill, and creating appropriate geometry for complex terrain takes a considerable amount of time. Furthermore, if measurement data are added or corrected afterward, all drawings must be recreated, which presents a serious inefficiency. When new measurement data are added during construction, all calculations must be redone, and this is a very time-consuming process. In practice, this lack of flexibility becomes a major burden on construction management.

Characteristics and Overwhelming Advantages of 3D Point Cloud Data

3D point cloud data has the potential to fundamentally solve the challenges of conventional cross-section methods. A 3D point cloud is digital data that records the coordinate information of countless points in space, and each point contains three-dimensional X, Y, and Z (elevation) coordinates. Such dense point cloud data can be efficiently acquired by laser scanners or by drones equipped with cameras, IMUs, and GNSS receivers. Importantly, this point cloud data captures even the subtle undulations of the ground surface. Differences in height on the order of millimeters (mm (in)) are also recorded.

The greatest advantage of 3D point cloud data is its overwhelmingly high density of measurement points. In conventional surveying, it was common to have tens to hundreds of measurement points per hectare. In contrast, 3D point clouds yield tens of thousands to hundreds of thousands of points over the same area. This overwhelming difference in density means that all subtle changes in terrain are captured, allowing volumes to be calculated with greater accuracy. When using a drone, measurements from an altitude of 100 meters (328.1 ft) can even detect height differences of about 1 centimeter (0.4 in) on the ground. Also, once 3D point cloud data has been acquired, it can be analyzed later from various angles and viewpoints. High reproducibility—where different surveyors analyzing the same data obtain the same results—is also extremely important. Because the objectivity of the data is ensured, if questions arise later about calculated results, the data can be reused to recompute them.

Complete workflow from acquisition to computation of 3D point cloud data

Automated calculations using 3D point clouds are carried out through several stages. The first stage is the acquisition of 3D point cloud data. Laser scanners and cameras mounted on drones are used to capture the three-dimensional geometry of the construction area. In this acquisition process, the measurement equipment is positioned on site and flown along a designated route, or measurements are taken sequentially from multiple positions. A single scan can collect millions of points, which are recorded as data with three-dimensional coordinate information. The flight plan is designed in detail in advance, with flight altitude, flight speed, image overlap rate, and other parameters determined. Weather and the angle of sunlight during the day are also taken into account, and the operation is planned to be carried out under optimal measurement conditions.

The next step for the acquired data is noise removal and preprocessing. Real-world laser scans and camera images inevitably contain noise. This stems from mechanical vibration during measurement, atmospheric effects, or reflections from objects outside the measurement target. A process is needed to automatically detect and remove such noise. At the same time, data obtained from multiple measurement positions are transformed into a unified coordinate system. This registration process is critically important to ensure multiple scans overlap precisely. Through careful processing, high accuracy is achieved.

Once noise has been removed, the next step is the extraction of embankment areas. From all point cloud data, only the parts corresponding to embankments are automatically classified and extracted. This is often carried out using machine learning algorithms, and classification is performed based on features such as surface slope and elevation. However, automated classification can contain errors, so verification and manual correction may sometimes be necessary. For example, it is necessary to check whether parts of existing structures have not been mistakenly included in the embankment area, or whether land that is not embankment has not been mistakenly included in the embankment area. Human verification work significantly improves quality.

When the fill area is confirmed, a final volume calculation is performed. Pre-construction and post-construction point cloud data are overlaid, and elevation differences at each point are calculated. By aggregating these elevation differences into a grid, the total volume of the fill area is computed. The calculation results provide not only the overall total volume but also the volume distribution for each sub-area, allowing a detailed understanding of construction progress. By setting a finer grid size, a more detailed spatial distribution can be obtained. This enables visualization of the distribution of construction quality.

Accuracy comparison results between conventional methods and 3D point clouds

Accuracy comparisons in real projects have made the superiority of 3D point clouds clear. It has been reported that the traditional average cross-section method based on conventional measurements can produce errors of up to about 5 percent in complex terrain because measurement points are limited. In particular, this error tends to be larger in areas where the terrain varies in a wave-like manner. For example, comparing the average cross-section method calculated with 200 measurement points per hectare to calculations using a 3D point cloud (300,000 points over the same area) can result in discrepancies of up to about 500 cubic meters. By contrast, automated calculations using 3D point clouds typically keep errors below 1 percent.

The primary reason for the difference in accuracy is the difference in the density of measurement points. In conventional methods, the spacing between measurement points is about 5 m (16.4 ft) to 20 m (65.6 ft), whereas 3D point clouds provide tens of thousands to hundreds of thousands of points over the same area. This overwhelming difference in density allows all subtle terrain changes to be captured, greatly improving calculation accuracy.

Moreover, the impact of differences in skill and subjective judgment among measurers is minimized in 3D point clouds. Because automatic calculation algorithms always output the same result for the same input data, reproducibility is complete. The high level of reliability—processing results that agree even for data measured by multiple measurers at different times—is a major advantage that conventional manual calculations could not achieve.

Furthermore, 3D point clouds make time-series comparisons very easy. By capturing point cloud data at each stage of construction and overlaying them, you can accurately track construction progress. With traditional methods, differences in measurement techniques and the like could affect comparisons when reviewing the results from each stage afterward, but with 3D point clouds, data comparisons can be performed using a completely unified method.

Dramatic Reduction and Streamlining of Processing Time

Automatic calculations using 3D point clouds dramatically reduce processing time. With conventional methods, from on-site measurement to final report preparation, a series of processes were carried out sequentially: pre-measurement preparation, on-site measurement work (several days to several weeks), conversion of measurement data into drawings (several days to one week), volume calculations (several days), and preparation of the final report (several days). Overall, it was common for the entire process to take at least about two weeks to one month. For large-scale projects, it could take more than two months.

In contrast, using 3D point clouds, data acquisition by drone can be completed in a matter of hours, and the subsequent automated processing can be finished in a few hours to about one day. In other words, from the start of measurement to report preparation, the process can be carried out in just one to two weeks. Furthermore, because many processes are automated, human resource allocation can be optimized. Traditionally, multiple technicians had to be assigned to drafting and calculations, but with 3D point clouds their work can be limited to verifying data quality and preparing the final report.

Especially when regular measurements and reporting are required during the construction period, the time-saving benefits are immeasurable. With conventional methods, volume management was limited to monthly measurements, but with 3D point clouds, carrying them out on a weekly basis or even more frequently becomes realistic. As a result, construction progress can be monitored more frequently, and responses to problems can be expedited.

Selecting and Using 3D Point Cloud Processing Software

Specialized software is required to process 3D point cloud data. There are multiple software products on the market, each offering different features and characteristics. Point cloud processing software includes functions such as point cloud visualization, noise removal, filtering, classification, volume calculation, and report generation. The extent to which these functions are implemented varies by product, so choosing the right product for the intended use is important.

For software specialized in volume calculation, multiple calculation algorithms are often incorporated, and the optimal method is automatically selected according to the characteristics of the terrain. In addition, functions for evaluating the reliability of the calculation results — for example, the number of points used in the calculation, density, and estimated accuracy — are output simultaneously. This information enables a qualitative assessment of the calculation results.

When selecting software, processing efficiency, ease of use, and support infrastructure are also important considerations. In complex point cloud processing, technical support may be required, and choosing a vendor with a comprehensive support system makes post-deployment troubleshooting easier.

Quality Control and Reliability Assurance

To ensure the reliability of automated calculations using 3D point clouds, strict quality control is required across all processes from measurement through computation. During the measurement stage, drone calibration, verification of flight parameters, and checks of weather conditions are important. In the data processing stage, it is necessary to verify the appropriateness of noise removal, the consistency of coordinate systems, and the accuracy of classification algorithms. In the validation stage of calculation results, compare the outcomes of multiple calculation methods and carry out simple, sample on-site measurements to verify them.

Such a multi-stage quality assurance process ensures the reliability of the final calculation results. Especially during the initial implementation of the system, particularly rigorous verification is required. By conducting repeated validations across multiple projects and determining the parameters best suited to the company's conditions, the reliability of subsequent calculations is improved.

Utilization of the Latest High-Precision Measurement Devices

Mobile measurement technologies, such as iPhone-mounted high-precision GNSS positioning devices that have emerged in recent years, provide accuracy comparable to 3D point clouds while enabling more flexible and lower-cost measurements. By deploying multiple devices on site, high-density position data can be acquired in real time. Cloud processing allows on-site verification of computation results, enabling immediate judgement on the need for additional measurements. This mobile-centric new measurement system offers convenience and efficiency not found in traditional measurement methods, and is likely to become the mainstream for embankment volume management going forward.

Considerations During the Implementation Phase

When considering the introduction of a 3D point cloud system, there are several points to keep in mind. First, comparing initial investment and operating costs is important. Purchasing high-performance drones and processing software requires a corresponding investment. However, if you comprehensively evaluate reductions in labor costs due to time savings, quality improvements from increased accuracy, and avoidance of losses from unexpected errors, a sufficient return on investment can be expected in the medium term.

Next, staff training and proficiency are required. Even if you introduce new technology, if there are no staff who can use it effectively, the return on investment will be limited. Multiple skills are necessary, such as drone operation, data processing, and interpretation of results. It is important to conduct sufficient technical training before implementation.

It is also necessary to consider alignment with existing business processes. When introducing a new system, it may become necessary to modify existing processes to accommodate it. If these adjustments are not carried out smoothly, the effectiveness of the system implementation will be diminished.

Future Developments and Further Evolution

3D point cloud technology is expected to continue evolving rapidly. Developments such as the creation of higher-precision, lower-cost sensors, improvements in processing speed, and increased levels of automation are anticipated. Furthermore, the introduction of AI and deep learning technologies will further enhance functions such as automatic classification and anomaly detection.

In addition, by leveraging IoT technologies, systems could evolve to enable real-time monitoring and management. By utilizing these more advanced technologies, the level of management of construction projects is expected to improve significantly.

Case Studies and Keys to Success

Companies that have actually implemented 3D point cloud systems report various benefits. A major construction company reported that by introducing 3D point clouds they were able to reduce measurement time by 30 percent compared with the previous approach and improve accuracy to more than twice the previous level. Another company said it was able to increase the frequency of regular measurements from once a month to once a week, enabling more detailed tracking of construction progress.

What these success stories have in common is that they did not simply introduce technology, but also improved their business processes to match it at the same time. To leverage new technologies, it is essential to establish an organizational structure and operating methods suited to them.

Potential Applications in Other Industries

Use of 3D point cloud technology is not limited to embankment volume calculations. It can be applied to many civil engineering projects, such as managing extraction volumes at quarries, measuring sedimentation in dam construction, and quantifying earth and sediment volumes in disaster recovery works. Furthermore, it has broad applicability in agriculture for measuring cultivated area, in forestry for measuring timber harvest volumes, and in mining for measuring mineral quantities. Such cross-sector applications are expected to drive further development of 3D point cloud technology.

Technology Integration and Data Strategy

To maximize the utilization of 3D point cloud data, integration with other construction technologies is important. By integrating with BIM (Building Information Modeling) systems, design data and as-built construction data can be managed in a unified manner. Additionally, integrating with GIS (Geographic Information Systems) makes it possible to understand wide-area topographic changes. These integrations enable more advanced construction management.

From a data strategy perspective, systematic management that anticipates the long-term preservation and reuse of 3D point cloud data is necessary. By utilizing cloud storage, past project data can be accessed at any time and leveraged as benchmark information for new projects.