Infrared Small Target Detection Based on Hierarchical Terrace Contrast Measure

The detection of infrared small targets is a topic of significant theoretical research value and practical application potential. However, the complex and dynamic detection environment introduces challenges such as intricate backgrounds, which can greatly impact the accuracy of the detection task. The existing method has limitations including the inability to detect extremely small targets, and also cannot competent to maintain exceptional detectability and good real-time performance when the small targets are under complex background. To address these issues, this paper presents an infrared small target detection method while offering superior real-time performance and exceptional detection capability. Firstly, a detection model for omnidirectional pixel differences is formulated in this paper by calculating hierarchical terrace contrast measure (HTCM) through pixel differences between the center pixel and its neighboring pixels at various levels and directions. Then, this paper exploits the radiative properties of small infrared targets by computing the minimum pixel difference between the center pixel and its neighboring pixels at various levels to simplify the model and enhance the real-time performance of the algorithm. Experimental testing on six publicly available datasets is conducted to compare our proposed method with seven existing algorithms. The results demonstrate that our approach achieves rapid detection for both large and small targets while exhibiting superior overall detection capability along with excellent real-time performance. HTCM can fastest achieve a single frame in only 0.0403 seconds and represents a 17.8% increase compared to the second place within an average single frame time.