No abstract

The TSSIM clutter metrics correlate amazingly well to both the experimental detection probabilities and the mean detection time. Based on the analysis of both probabilities of the correct detections and of the total (correct and false) number of detections made by human observers, a mathematical formula for predicting the probability of false alarms as a function of clutter metrics is presented in this paper. Comparing real experimental data with the predicted products reveal very good agreement, which is very helpful in understanding human behavior mechanisms regarding target detection tasks. It is concluded that the human observer behaves as fixed threshold signal processor /Non-CFAR.

A new algorithm is presented which deals with the problem of detecting small moving targets in infrared image sequences that also contain drifting and evolving clutter. Through development of models of the temporal behavior of the static background, target and cloud edge on a single pixel basis, the new algorithm employing the connecting line of the stagnation points (CLSP) of the temporal profile as the baseline is created and tested. The deviation of the temporal profile and its CLSP is analyzed and it is determined that the distribution of the residual temporal profile obtained by subtracting the baseline from the temporal profile can be modeled by a Gaussian distribution. The occurrences of the targets have intensity values significantly different to the distribution of the residual temporal profile. Unlike the conventional 3-D method, this new algorithm operates on the temporal profile in 1-D space, not in 3-D space, thus having a higher computational efficiency. Experiments with real IR image sequences have proved the validity of the new approach.

Anomaly detection (AD), which aims to distinguish targets with significant spectral differences from the background, has become an important topic in hyperspectral imagery (HSI) processing. In this paper, a novel anomaly detection algorithm via dictionary construction-based low-rank representation (LRR) and adaptive weighting is proposed. This algorithm has three main advantages. First, based on the consistency with AD problem, the LRR is employed to mine the lowest-rank representation of hyperspectral data by imposing a low-rank constraint on the representation coefficients. Sparse component contains most of the anomaly information and can be used for anomaly detection. Second, to better separate the sparse anomalies from the background component, a background dictionary construction strategy based on the usage frequency of the dictionary atoms for HSI reconstruction is proposed. The constructed dictionary excludes possible anomalies and contains all background categories, thus spanning a more reasonable background space. Finally, to further enhance the response difference between the background pixels and anomalies, the response output obtained by LRR is multiplied by an adaptive weighting matrix. Therefore, the anomaly pixels are more easily distinguished from the background. Experiments on synthetic and real-world hyperspectral datasets demonstrate the superiority of our proposed method over other AD detectors.

No abstract