The camouflage assessment tool LOAT was introduced to the Austrian Air Reconnaissance Center at the end of 1999. This segmentation based and texture oriented method was already presented at AeroSense Conferences. For the new task of assessing video sequences of partly moving objects this single image method seemed to be too time consuming - a different approach relying on former experiences seemed to be feasible. In June 2002 a three nation (Austria, Germany, Switzerland) camouflage-field-campaign "MUSTAFA" was conducted on an Austrian proving-ground; data was recorded for the following expert analysis of the participants. Motion imagery was recorded in visual and thermal infrared bands. The scenario was the approach of an attack-helicopter towards typical military targets which were camouflaged with multi spectral measures. Besides a validation of the LOAT-System by selecting a small sample of the collected video frames a so called "videobase analysis method" is being developed and is scheduled for examining the videos for motion-camouflage-assessment.
The "videobase analysis method" (it will become LOMAT-system = low observables motion assessment tool if successful) is based on a very fast video-processing software. The assessment algorithm to be implemented is based on a MDL clustering algorithm in conjunction with a database for numerical inter frame analysis. Features are the results of the clustering process like information content of the data, number of classes found and the like. We consider it a very promising approach for the projected real-time task mainly because there is no calibration process involved. The research done by D2K Solutions Ruppert&Partner OEG, an Austrian spin-off company of Joanneum Research is scheduled to provide the basic tool with end of 2002; the examination of the MUSTAFA data is due until May 2003.
The approach of the "videobase-analysis method", the introduction of the MDL based assessment method will be presented together with some preliminary visualizations e.g. screenshots of the work done so far. Samples of the collected motion-image will also be presented.
A robust computer based camouflage assessment approach was presented at the AeroSense 2000 conference. Based on experiments with human observers a separability measure was developed. The method was classifier based and best results could be obtained using the C4.5 classifier as a separability measure. Using this method makes camouflage assessment transparent and deterministic presuming correctly specified regions of interest. This paper describes our effort to overcome the drawback of the need of user input at such a critical step within the method. We used unsupervised learning along with an optimizing method to derive information about the number of clusters and other performance measurements. All these measurements coming from the optimization step were adopted to camouflage assessment.
A human-in-the-loop computer based camouflage assessment approach was already presented at the AeroSense 1998 conference.3 The same image sets were used for human photosimulation as well as for the computer assessment method. The human photosimulation results suggested four camouflage classes which were used to develop and verify the separability measure. Analyzing camouflage effectiveness using separability measures induces a very complex feature space. Best results were obtained using the C4.5 classifier as a separability measure. The size of the objects presented duing the photosimulation sessions and tactical knowledge of the observers had significant influence on the detectoin/recognition performance of humans. The most important advantage of our method is to make camouflage assessment more transparent and deterministic. Results of a selected experiment during a field test are shown in this paper.
A key point for good camouflage results in the thermal infrared domain lies in the ability of the camouflage system to adapt to the thermal emission behavior of the surrounding background. In order to obtain reliable assessments of the camouflage effectiveness, evaluation has to take place under various environment condition. The combination of the different results leads to a assessment measure with the demanded reliability. The object quantization of the individual camouflage effectiveness and the following combination is very difficult to achieve by human operators. Therefore an Infrared Camouflage Effectiveness Assessment Tool (ICEAT) has been developed, which needs only minor human interaction and supports the automated combination of the results of various test scenes. In a first step hot spots of the object and the background are detected. In a second phase various features are calculated which are combined to a single assessment measure in the third phase by using fuzzy logic. The fuzzy logic approach has the advantage that the customization of the ICEAT can be achieved by simply modifying the used membership functions.
An accurate method to detect and classify military vehicles based on the recognition of shapes is presented in this work. FFT-Descriptors are used to generate a scale, translation and rotation invariant characterization of the shape of such an object. By interpreting the boundary pixels of an object as complex numbers it is possible to calculate an FFT-Descriptor based on the spectrum of a Fast Fourier Transform of these numbers. It is shown that by using this characterization it is possible to match such representations with models in a database of known vehicles and thereby gaining a highly robust and fault tolerant object classification. By selecting a specific number of components of a FFT-Descriptor the classification process can by tailored to different needs of recognition accuracy, allowed shape deviation and classification speed.
Shape and shape disruption have significant influence to the human target acquisition mechanism. A special testing method (the so called `photo-simulation') was developed in the eighties to present a set of image slides of camouflaged and not camouflaged objects in preferably natural backgrounds to military personnel to quantify differences in object camouflage effectiveness. Statistically significant results were achieved, however, the high test requirements limited its practical use. The project is motivated by an urgent need for a camouflage evaluation system based on computer vision with a fast response so that the user in a field test can be supported to further improve his camouflage skills. Hence, the photo simulation method cannot be regarded as obsolete, it can be used to compare the results of the camouflage evaluation system with the results of human perception. With an human-in-the-loop computer based camouflage assessment system, processing should be sped up by some orders of magnitude, could be automated for field tests and would yield several additional features. To overcome the problem of quantifying e.g. texture similarity of different camouflage nets to blend into the natural background, an image processing/visualization method was pursued by the Austrian Ministry of Defense. Now the same image-sets can be used for the human photo-simulation as well as for segmentation/classification by the camouflage assessment tool. Today a modified Euclid-distance measurement for visual images is being used while similarity of shapes (gestalt) to a selected region can be visualized. Feature selection is being done by training a neural network with the results of the human perception data. A cost effective prototype of a camouflage assessment tool based on standard hardware can be presented. Its promising performance gives hope to get beyond subjective camouflage experts stimuli. In the next project phase also thermal images shall be handled with the camouflage assessment tool.