In complex non-smooth backgrounds, infrared dim and small target targets generally have lower energy and occupy fewer pixels, and are easily swamped by clutter. To improve the detection capability of dim and small targets in non-smooth scenes, this paper proposes a new dim and small target detection method combining multidirectional gradient difference regularization principal component decomposition model. The method first establishes a new gradient difference regularization to constrain the low-rank subspaces of different image components, then construct a gradient difference regularization-based principal component decomposition model (GDR-PCD), and finally decomposes the model using the overlapping directional multiplier method to obtain the background impedance. The experimental results show that the method performs better in all six sequential image scenes than the traditional algorithm. Furthermore, the detection results verify the algorithm's effectiveness in this paper.INDEX TERMS infrared dim and small target detection, principal component decomposition model, gradient difference regularization principal component decomposition model, overlapping directional multiplier method

Infrared target detection is often disrupted by a complex background, resulting in a high false alarm and low target recognition. This paper proposes a robust principal component decomposition model with joint spatial and temporal filtering and L1 norm regularization to effectively suppress the complex backgrounds. The model establishes a new anisotropic Gaussian kernel diffusion function, which exploits the difference between the target and the background in the spatial domain to suppress the edge contours. Furthermore, in order to suppress the dynamically changing background, we construct an inversion model that combines temporal domain information and L1 norm regularization to globally constrain the low rank characteristics of the background, and characterize the target sparse component with L1 norm. Finally, the overlapping multiplier method is used for decomposition and reconstruction to complete the target detection.Through relevant experiments, the proposed background modeling method in this paper has a better background suppression effect in different scenes. The average values of the three evaluation indexes, SSIM, BSF and IC, are 0.986, 88.357 and 18.967, respectively. Meanwhile, the proposed detection method obtains a higher detection rate compared with other algorithms under the same false alarm rate.

In the face of complex scenes with strong edge contours and high levels of noise, suppressing edge contours and noise levels is challenging with infrared dim and small target detection algorithms. Many advanced algorithms suffer from high false alarm rates when facing this problem. To solve this, a new anisotropic background feature weight function based on the infrared patch tensor (IPT) model was developed in this study to characterize the background airspace difference features by effectively combining the local features with the global features to suppress the strong edge contours in the structural tensor. Secondly, to enhance the target energy in the a priori model, an improved high-order cumulative model was proposed to establish the local significance region of the target as a way to achieve energy enhancement of the significant target in the structural tensor. Finally, the energy-enhanced structural tensor was introduced into the partial sum of the sensor nuclear norm (PSTNN) model as a local feature information weight matrix; the detection results were obtained by solving the model with the help of ADMM. A series of experiments show that the algorithm in this paper achieves better detection results compared with other algorithms.