1.Software

SLAM GO is a mobile APP that is used together with SLAM100.  It can be connected to SLAM100 through mobile phones to do operations like project management, real-time point cloud puzzle display, image preview, firmware upgrade and so on.

SLAM GO POST Pro, is a kind of PC software that is used together with SLAM100, embedded in the professional edition pf drone butler. It can be applied to make post-processing on the data collected by SLAM100 to produce color point clouds and partial panoramas with high precision and fineness, allowing browsing through and optimizing the point clouds.

Smart point cloud is not only equipped with basic functions like browsing, rendering, denoising, resampling, removing redundancy and coloring, and function to automatically quickly filter out noise and make classification, but also has diverse data interactive editing tools. 

CloudCompare, a free commercial software, uses the grid method to calculate earthwork quantity based on point cloud data. The size of the grid can be set by yourself. It can be used to calculate the earthwork variation of two stage of point clouds and the earthwork quantity of one stage of point cloud to specified elevation surface.

Figure Earthwork Calculation in CloudCompare

2.Job in the Field

Field operation flow chart

① Site Survey
Check the survey area on site to know about the distribution of routes and ground features in the survey area, and the connectivity of roads, exits and channels, so as to facilitate the follow-up scanning route planning. 

② Layout of Control Points
For some projects, point cloud data is required to be an absolute coordinate system. On this occasion, coordinates of control points can be measured by using GNSS-RTK or total station for subsequent automatic extraction of control points and coordinate transformation. This step can be passed if there are no requirements on point cloud data.

③ Route Planning
Scanning route can be planned reasonably according to the survey situation and distribution you have. The most ideal scanning route is "Closed-loop Scanning". When "Closed-loop Scanning" is performed, the software will automatically identify the overlapping areas simply by returning back to a point scanned previously. This step is used to calculate or follow up with and rectify any offset errors occurred, similar to the "Closed Loop" in leveling. When a loop can't be formed because of the condition limitation, SLAM GO POST Pro, the post-processing software used to work together with SLAM100, can still ensure the quality of point cloud data.

Figure Scanning Route Planning

④ Data Collection
The collection personnel hold the SLAM100 to collect point cloud data according to the planned route. When he walks near the control point, he overlaps the cross center of the SLAM100 base and the control point and stands for 10 seconds. Effect of splicing for real-time point cloud can be browsed through SLAM GO during scanning. If there is anything lost, it is allowed to scan again, to make sure the completeness of point cloud data.

Figure Real-time Display in SLAM GO

3.Technical Route in Office Work

① Point Cloud Calculation

Figure Data Graph Import(Left), One-button Calculation(Right)

SLAM GO POST Pro can be applied to make post-processing on original data by one button, generate point-cloud data with high precision, color point cloud, partial panoramas and others.

Figure Partial Point Cloud Effect

Figure Partial Panorama Effect

② Coordinate Transformation
In SLAM GO POST Pro, coordinate transformation is performed by extracting the position of control points from point cloud based on algorithm through automatic extraction of control points, free from error resulting from manually adding control points.

③ Precision Checking and Verifying
Import point cloud into the smart point cloud, and then check the precision of point cloud after transformation by using tool "Precision Checking and Verifying". If the error between point cloud and point checked meets the requirements on precision, the subsequent calculation can be carried out.

Figure Checking and Verifying the Precision of Point Cloud

④ Point Cloud Classification
The purpose of point cloud classification is to accurately extract terrain points. This step is very critical because it determines the accuracy of precision for earthwork quantity. The collection personnel holds SLAM100 to get data at close range. Under this situation, the penetration effect of point cloud is good. Combined with the automatic classification calculation of terrain points in smart point cloud to calculate the terrain points separated quickly and accurately.

Figure Terrain Points

⑤ Earthwork Calculation
Earthwork calculation can be performed in software CloudCompare. Parameter settings like size of grid, reference surface, grid height based on the terrain points after classification can be determined and then the result of earthwork calculation will be output.

Figure Earthwork Calculation Results

4.Scheme Characteristics

SLAM100 is successfully applied to earthwork calculation in this scheme. Point cloud data in survey area can be obtained thoroughly by using SLAM100. Terrain points are acquired after coordinate transformation and point cloud classification. Combining with the third software, earthwork quantity can be calculated based on the precision of site point cloud, giving data support for the subsequent modification and construction. The features of the scheme are as follows:

① It innovates a new way to acquire measurement data, forming incremental maps by autonomous navigation and localization without GNSS.

②It breaks through the constraints and shortcomings of traditional measurement method of single-point position, free from the trouble of frequent station change and point cloud splicing of static scanners, improving the office work and field work efficiency.

③Handheld acquisition method applied ensures penetration of data. Matched software SLAM GO achieves thorough field scanning, keeping the integrity of the point cloud.

④The relative accuracy and absolute accuracy of the post-processing algorithm of industrial SLAM are better than ±2cm and ±5cm respectively, ensuring the accuracy of the earthwork calculation results. 

⑤Automatic extraction of terrain points in smart cloud is of high accuracy, reducing heavy workload of manual classification, to improve the efficiency of office work.

⑥In CloudCompare, one-stage earthwork quantity and multi-stage earthwork quantity are calculated by using grid method based on site point clouds. The software has wide application range, strong data bearing capacity, simple operation processes, high precision and fast operation speed. 

⑦It is a integrated software and hardware solution, achieving seamless connection between data, suitable for earthwork calculation in various scenarios. 

5. Application Cases

①Park Renovation and Expansion
To increase the coverage area of green plants in a park, it is needed to do planning and transformation on its existing sites. In this case, accurate earthwork calculation of the reconstruction site is required to provide data support for construction. The transformation area is in the center of the park, covering an area of about 1500m². The terrain relief is not excessive. With large vegetation coverage, GNSS signals are seriously blocked, and sighting situation is obstructed.

Figure Location and General Situation of the Transformation Area

The collection personnel collect data at the site. In this way, it can not only deal with the problems of unallowable measurement because of force majeure factors (vegetation occlusion, etc.) when using conventional measurement method, but also it can be free from the trouble of frequent station change and point cloud splicing of static scanners, improving work efficiency. Effect of splicing for real-time point cloud can be browsed through SLAM GO during scanning. If there is anything lost, it is allowed to scan again in time, to make sure the integrity of point cloud data. Field scanning takes 10 minutes.

Figure Field Data Collection(Left), Raw Point Cloud Data(Right)

The raw data is imported into SLAM GO POST Pro for post-processing to create point cloud map with high precision. After classification of point clouds, it can go ahead to make earthwork calculation by using CloudCompare. The processing takes 1.5 hours. According to the actual situation of the site, the filling volume is 85m³, which is negligible. The excavation volume is 2898.107m³. 

Figure Terrain Points and Triangulation Mesh Rendering

Figure Earthwork Calculation Result

② Inventory of Granary Reserves

For a granary, it is required to do regular inventory of the surplus grain. In the granary, airborne Lidar can't be applied to obtain data because it is unavailable to receive GNSS signals. Besides, static scanners can't acquire the data of the top of grain pile whose height is about 4-5m. Held with SLAM100, the collection personnel walk around the grain pile and can collect complete 3D point cloud results by working with an extension rod. Owing to the different sizes of grain depot, it spends about 5-7 minutes to scan a grain depot.

The point cloud of the grain pile can be separated out individually after denoising and cutting. If there is no disorder point cloud on the grain pile surface, grain reserves can be calculated directly. The storage volume of a grain depot is 3400m³. According to different sizes and internal space layout for grain depots, it takes about 25-30 minutes to make inventory of grain reserves for a grain depot.

Figure Site of SLAM100 Scanning Grain Piles

Figure Point Clouds of Grain Pile

Figure Earthwork Calculation of Grain Pile

③ Earthwork Calculation of Construction and Demolition Wastes
There is a large amount of construction and demolition wastes generated during the construction in a construction site, and now it is required to make a cleanup plan by confirming the volume of the construction and demolition wastes. The construction and demolition wastes covers an area of about 1400㎡, and its shape is extremely irregular. If static scanners are applied , it is needed to set station for 12-14 times to finish scanning. However, held with SLAM100 in hand, the technicians can obtain continuous heap data after scanning for once around the construction and demolition wastes and its top, free from post splicing, taking 9 minutes in field measurement.   After office personnel spends 1.2 hours to do operations on point cloud data like denoising, cutting, classification and so on. The volume of heap formed by construction and demolition wastes and ground is 2704.506m³.

Figure Construction and Demolition Wastes

Figure Point Cloud Map of Construction and Demolition Wastes

Figure Volume Calculation Results

________________________________________________________________________________________________