Tactical reconnaissance using small unmanned aerial vehicles has become a common military scenario. However, since their sensor systems are usually limited to rudimentary visual or thermal imaging, the detection of camouflaged objects can be a particularly hard challenge. With respect to SWaP-C criteria, multispectral sensors represent a promising solution to increase the spectral information that could lead to unveiling camouflage. Therefore, this paper investigates and evaluates the applicability of four well-known hyperspectral anomaly detection methods (RX, LRX, CRD, and AED) and a method developed by the authors called local point density (LPD) for near real-time camouflage detection in multispectral imagery based on a specially created dataset. Results show that all targets in the dataset could successfully be detected with an AUC greater than 0.9 by multiple methods, with some methods even reaching an AUC relatively close to 1.0 for certain targets. Yet, great variations in detection performance over all targets and methods were observed. The dataset was additionally enhanced by multiple vegetation indices (BNDVI, GNDVI, and NDRE), which resulted in generally higher detection performances of all methods. Overall, the results demonstrated the general applicability of the hyperspectral anomaly detection methods for camouflage detection in multispectral imagery.

UAVs are widely used for aerial reconnaissance with imaging sensors. For this, a high detection performance (accuracy of object detection) is desired in order to increase mission success. However, different environmental conditions (negatively) affect sensory data acquisition and automated object detection. For this reason, we present an innovative concept that maps the influence of selected environmental conditions on detection performance utilizing sensor performance models. These models are used in sensor-model-based trajectory optimization to generate optimized reference flight trajectories with aligned sensor control for a fixed-wing UAV in order to increase detection performance. These reference trajectories are calculated using nonlinear model predictive control as well as dynamic programming, both in combination with a newly developed sensor performance model, which is described in this work. To the best of our knowledge, this is the first sensor performance model to be used in unmanned aerial reconnaissance that maps the detection performance for a perception chain with a deep learning-based object detector with respect to selected environmental states. The reference trajectory determines the spatial and temporal positioning of the UAV and its imaging sensor with respect to the reconnaissance object on the ground. The trajectory optimization aims to influence sensor data acquisition by adjusting the sensor position, as part of the environmental states, in such a way that the subsequent automated object detection yields enhanced detection performance. Different constraints derived from perceptual, platform-specific, environmental, and mission-relevant requirements are incorporated into the optimization process. We evaluate the capabilities of the sensor performance model and our approach to sensor-model-based trajectory optimization by a series of simulated aerial reconnaissance tasks for ground vehicle detection. Compared to a variety of benchmark trajectories, our approach achieves an increase in detection performance of 4.48% on average for trajectory optimization with nonlinear model predictive control. With dynamic programming, we achieve even higher performance values that are equal to or close to the theoretical maximum detection performance values.

To improve the management of multispectral sensor systems on small reconnaissance drones, this paper proposes an approach to predict the performance of a sensor band with respect to its ability to expose camouflaged targets under a given environmental context. As a reference for sensor performance, a new metric is introduced that quantifies the visibility of camouflaged targets in a particular sensor band: the Target Visibility Index (TVI). For the sensor performance prediction, several machine learning models are trained to learn the relationship between the TVI for a specific sensor band and an environmental context state extracted from the visual band by multiple image descriptors. Using a predicted measure of performance, the sensor bands are ranked according to their significance. For the training and evaluation of the performance prediction approach, a dataset featuring 853 multispectral captures and numerous camouflaged targets in different environments was created and has been made publicly available for download. The results show that the proposed approach can successfully determine the most informative sensor bands in most cases. Therefore, this performance prediction approach has great potential to improve camouflage detection performance in real-world reconnaissance scenarios by increasing the utility of each sensor band and reducing the associated workload of complex multispectral sensor systems.

Target detection is a cornerstone task in hyperspectral image processing but faces significant challenges due to domain gaps. While statistical detectors like Constrained Energy Minimization (CEM) and Adaptive Cosine Estimator (ACE) are not prone to learned biases, in practice they still suffer from mismatches between the reference target spectrum and the spectral characteristics of the target in the test scene. We propose Test-time Adaptive Spectrum Refinement (TASR), a novel framework addressing this problem. TASR operates in an interpretable, lightweight, data-efficient manner, requiring only a single labeled source image of the target material. At test time, TASR dynamically refines the target spectrum to better align with the spectral properties of the test scene. This adaptive refinement enables detectors to effectively handle data with spectral variations, bridging the gap between the source and test spectra. To validate TASR, we conduct extensive experiments on established benchmarks and introduce a new dataset—ShadySunnyDiffuse (SSD)—which explicitly tests detector robustness to naturally occurring illumination changes. We further demonstrate the method’s versatility by applying it to camouflage detection and show compatibility with multiple statistical detectors. Our results establish TASR as a state-of-the-art approach in domain-adaptive hyperspectral target detection and target spectrum management.

The performance of deep learning based algorithms is significantly influenced by the quantity and quality of the available training and test datasets. Since data acquisition is complex and expensive, especially in the field of airborne sensor data evaluation, the use of virtual simulation environments for generating synthetic data are increasingly sought. In this article, the complete process chain is evaluated regarding the use of synthetic data based on vehicle detection. Among other things, content-equivalent real and synthetic aerial images are used in the process. This includes, in the first step, the learning of models with different training data configurations and the evaluation of the resulting detection performance. Subsequently, a statistical evaluation procedure based on a classification chain with image descriptors as features is used to identify important influencing factors in this respect. The resulting findings are finally incorporated into the synthetic training data generation and in the last step, it is investigated to what extent an increase of the detection performance is possible. The overall objective of the experiments is to derive design guidelines for the generation and use of synthetic data.