Accurate measurement of granary stockpile volume based on fast registration of multi-station scans. Issue 6 (3rd June 2018)

• Record Type: Journal Article
• Authors: Zhu, Jianjun; Yang, Jian; Fan, Jingfan; Ai, Danni; Jiang, Yurong; Song, Hong; Wang, Yongtian
• Is Part Of: Remote sensing letters. Volume 9:Issue 6(2018)
• Page Start: 569
• Page End: 577
• Publication Date: 2018-06-03
• Subjects: Remote sensing -- Periodicals; Remote sensing; Periodicals; 621.3678
• DOI: 10.1080/2150704X.2018.1452060
• Languages: English
• ISSNs: 2150-704X

Abstract

ABSTRACT: A high-quality model with sufficient accuracy and efficiency is crucial in volume measurement of granary stockpile. Single scanning data usually cannot satisfy the high-accuracy requirement of application. Hence, multi-station scanning is usually performed to obtain a complete and dense point cloud granary model. In this study, the convex hull indexed Gaussian mixture model is introduced for accurate point cloud registration of the granary. The granary stockpile volume is calculated by subtracting the volume of full granary model from empty granary model. Hence, a novel strategy for full granary scan calculation is proposed to extend the application of granary stockpile measurement. This strategy is achieved by reconstructing a virtual empty granary model using grain line baseline and vertical principal axis. This study presents the registration and fused results of multi-station scans of different granaries. Experiments are designed to compare the stockpile volume measurements of single scanning and fused data, thereby demonstrating that the proposed method is very effective and robust for granary stockpile measurement. The proposed method has the potential to be utilized in intelligent granary management in the future because it is fully automatic.