Correlation Analysis Between Environmental Pressure and Vision-Based Flames from Monitoring Camera During Depressurization Process.

Li, Cong,Yao, Yina,Yang, Rui,Li, Chengwu,Zhang, Hui

DOI: https://doi.org/10.1007/s11760-021-02089-9

IF: 1.583

2022-01-01

Signal Image and Video Processing

Abstract:The depressurization process usually occurs in a cargo compartment when an aircraft is taking off. The fire detection based on the vision-based flame has unique advantages in the confined and inaccessible environment. In this study, the n-heptane pool fires with a diameter of 30 cm were performed in a pressure cabin under decreasing pressure from 101 to 30 kPa. The vision-based flames were recorded by a high-resolution monitoring camera with a fixed aperture and exposure time. The shape features and texture features including flame area, perimeter, convexity, derived ratio, contrast, and correlation of gray-level co-occurrence matrix (GLCM) were extracted, which form a multidimensional vector. Then, the dimensionality reduction was conducted using principal component analysis (PCA) method, and a principal component was obtained to characterize the flame image, whose value had an excellent linear relationship with the instantaneous decreasing pressure. Further, the relationship was verified by a series of 20-cm-diameter pool fires. The study proves the feasibility of the quantitative fire detection method based on visual flame and provides the theoretical basics in applying to the confined and inaccessible space.