HCHODetector: Formaldehyde concentration detection based on deep learning. Issue 1 (April 2021)

• Record Type: Journal Article
• Authors: Cao, Zhihao; Shao, Mingfeng; Shi, Aiju; Qu, Hongchun
• Is Part Of: Journal of physics. Volume 1848:Issue 1(2021)
• Publication Date: 2021-04
• Subjects: Physics -- Congresses; 530.5
• DOI: 10.1088/1742-6596/1848/1/012047
• Languages: English
• ISSNs: 1742-6588

Abstract

Abstract: Currently, deep learning technology is developing rapidly. Deep learning is mainly used in the fields of vision and hearing for human beings, but less in the field of olfactory. Formaldehyde is a common gas harmful to human health. However, the traditional methods of Formaldehyde concentration detection are inefficient in some cases. As for this problem, this paper proposes a novel formaldehyde detector namely HCHODetector. Specifically, this detector is based on deep learning and HSV colour space augmentation. Moreover, we propose a novel Mask-guided module and a novel pre-training network to enhance the colour discrimination ability of HCHODetector. As a consequence, the experimental results show that the detection error is within 0.08 mg/m 3 in the actual environment, which provides a new idea for Formaldehyde concentration detection.