Volume Estimation of Raw Materials inside the Silo using Convex Hull Method

The uneven surface of the raw materials stored inside a big voluminous container, such as a silo, makes it hard for the conventional technique (sounding method) to accurately estimate the volume of the raw materials inside the silo. Thus, this study presented an alternate approach that reduces effort and improves accuracy by developing a three-dimensional raw material mapping with point cloud processing for more accurate volume estimation. A prototype silo was used as the testing site of the test subject and was scanned using the 2D light detection and ranging (LiDAR) scanner. The system used ROS to connect to the scanner and handle the acquired data. Then, the researchers implemented an algorithm to convert the scanned two-dimensional data to three-dimensional data and processed it using statistical outlier removal to remove the outliers. The Convex Hull method calculates the volume of the processed point cloud data. Compared to the 14.1598% average Mean Absolute Percentage Errors (MAPE) of the sounding approach, the experimental results demonstrated that the 1.239% average MAPE of the suggested method increases the accuracy percentage of volume estimation by approximately 12%.