A novel method for quantifying target tracking difficulty of the infrared image sequence

Abstract

Metrics to measure the difficulty of target tracking on the infrared image sequence were seldom studied. This paper proposes an effective method for quantifying target tracking difficulty of the infrared image sequence. We first analyze the factors interfering with infrared target tracking, and conclude that the intra-frame image quality and the inter-frame variation will jointly make great effects on infrared target tracking. Furthermore, five metrics concerning target tracking difficulty of the infrared image sequence are proposed: intra-frame degree of occlusion (IFDO), intra-frame degree of confusion (IFDC), inter-frame variation degree of target texture (IFVDTT), inter-frame variation degree of target size (IFVDTS) and inter-frame variation degree of target location (IFVDTL). These metrics complement each other and measure the infrared image sequence concerning target tracking difficulty together. An integrated indicator named sequence difficulty degree (SDD) is proposed by combining these five metrics to intuitively represent the sequence-level tracking difficulty of the infrared image sequence. Related experiments are designed to illustrate the effectiveness of the proposed metrics. Experimental results prove that the proposed metrics are valid to measure infrared image sequences concerning target tracking difficulty, and the performance of the proposed metrics is superior to that of the traditional infrared image sequence metric.