Research on Intelligent detection of concrete aggregate level based on monocular imaging. (15th May 2022)

• Record Type: Journal Article
• Authors: Liu, Mingtang; Yue, Shuang; Li, Shu; Du, Yuying; Li, Bin
• Is Part Of: Measurement. Volume 194(2022)
• Publication Date: 2022-05-15
• Subjects: Concrete batching plant -- YOLOv5 -- Aggregate level -- Object detection -- Monocular imaging; Weights and measures -- Periodicals; Measurement -- Periodicals; Measurement; Weights and measures; Periodicals; 530.8
• DOI: 10.1016/j.measurement.2022.111036
• Languages: English
• ISSNs: 0263-2241

Abstract

Highlights: Construct a specific projection model of camera and aggregate storage bin. Establish a mapping relationship between the image coordinates and the actual imaging angle. Obtain the real-time image information on the aggregate level of the aggregate storage bin. Abstract: For the current problems of inaccurate, non-real-time and costly of the measurement of the aggregate level in concrete batching plant, this paper proposes an intelligent detection method of concrete aggregate level based on monocular imaging. This method uses a monocular camera, installed at 45-degree angle, construct a specific projection model of camera and storage bin to establish a mapping relationship between the image coordinates and actual imaging angle, and then combine peak and valley positioning information derived from YOLOv5 to find the height of aggregate level. The experimental results show that the average accuracy of the concrete batching plant aggregate level detection method based on monocular imaging is 95.3%, and the real-time monitoring speed is 111f/s, which is higher than the traditional corner detection method 21.6% and 99f/s, and higher than the YOLOv4 monocular measurement method 14.5% and 51f/s, respectively. The effects can meet the demand of aggregate level monitoring.