Extraction and analysis of spatial heterodyne potassium signals based on principal component analysis and non-local means method

The use of flame suppressants in jet-propelled aircraft significantly reduces the infrared radiation of their exhaust plumes, thereby increasing the difficulty of target detection based on infrared radiation. Potassium salts, as a component of flame suppressants, produce characteristic signals when burned. To probe into new methods for detecting flying targets, a spatial heterodyne spectrometer is utilized to detect the weak signals from potassium salt combustion against a sky background, combined with data processing techniques. In the experiment, a potassium lamp is employed to simulate the potassium combustion signals and placed in a sky background for data acquisition. Preliminary processing results revealed that the signals were submerged within the atmospheric background. Principal Component Analysis (PCA) is then applied to separate the atmospheric background from the weak potassium lamp signals in the mixed signals, followed by the introduction of the Non-Local Means (NLM) denoising algorithm to suppress noise. Finally, Principal Component Regression (PCR) is used to restore the potassium lamp signals. Quantitative analysis demonstrated that the potassium lamp signals could be effectively extracted at a signal-to-noise ratio (SNR) of 0.1310, and the signal intensity exhibited a linear relationship with the current, with a correlation coefficient of 0.9823. Thus, the combination of spatial heterodyne detection technology with PCA and NLM methods is feasible for detecting potassium combustion signals against an atmospheric background to identify jet-propelled flying targets.