Deep Factor Analysis for Weather Varied Sense-through-foliage Target Detection
Abstract In this paper, the influence of seasonal variation on target detection accuracy and the effectiveness of deep factor analysis(DFA) in signal denoising are studied. To extensively verify the universality of the proposed DFA approach, a variety of target objects, including no target, human, wood board and iron cabinet targets, are measured in foliage environment under four different weather conditions. Then, after removing background noise from the collected data, deep factor analysis is carried out to further reduce the impact of noise. The experimental results show that the influence of weather variation on target detection can be effectively eliminated by the presented DFA, which can improve the average classification accuracy in all seasons. At the end of the paper, it is verified by cross validation that the DFA method can be stabilized at around 93% even in hazy day and snowy day which has stability and universality in any weather conditions, even in snowy and haze days.