This paper describes ATR/I algorithms that were developed by Thales Optronique for the real-time automatic target recognition and identification of ground vehicles in a Air-to-Ground non cooperative context. The main principles of the algorithm based on an exhaustive comparison between the input image and the elements of a 〈 Model Data Set 〉 are:
• To avoid the variability on the gray-levels of the target, the comparison is not performed directly on the input gray-level image but on an edge image which is obtained with a segmentation algorithm derived from the classical Canny-Deriche edge detector.
• The selected architecture is chosen to repeat many times a single instruction rather than to execute only one time a lot of different instructions. Therefore, the comparisons between the input image and the elements of the 〈 Model Data Set 〉 are performed in 2D with a correlative technique.
• The computation time is achieved thanks to a coarse to fine analysis with different levels of comparison : in a first stage, a simple comparison measure is used which enables quick selection of a preliminary list of potential hypotheses. This measure is discriminating enough to select a small number of hypotheses and robust enough to select the true hypothesis associated with the target. These selected hypotheses are then analyzed during a second stage of processing using a more refined measure, and thus more time consuming than the previous one, but which is applied on a significantly reduced number of hypotheses.
Segmentation of coherent active images, which schematically consists in determining the shape of objects present in a scene, is a challenging problem due to the large fluctuations inherent to speckle noise. Snake algorithms have made it possible to improve the segmentation performance of a single object in video images, but they are not well suited to speckled images. Furthermore, they require regularization techniques to obtain smooth contours, which introduces free parameters in the algorithm that must be adjusted by supervised learning. We recently introduced a new technique based on a polygonal description of the contour to be segmented and on the optimization of a statistical criterion. This approach leads to good performance for the segmentation of a target with homogeneous random gray levels but still requires a regularization term. Here, we show that a new technique based on the Minimum Description Length principle makes it possible to efficiently segment an object in a speckled image with a fast algorithm which has no free parameters. This method is thus fully automatic and well suited to speckled images, and we propose to illustrate its capabilities on classical and polarimetric active speckled images.
We have recently proposed an original approach for the statistical segmentation of an object, based on active contours. In this paper, we propose a comparison of several Hausdorff distances performances in dissimilarity measurements for silhouette discrimination. For this purpose, we apply on the silhouettes of a dataset six kinds of perturbations that can occur with an active contour technique and we compute the good discrimination rate versus the reject rate for each distance. We also propose a simple method to accelerate the calculus of the Hausdorff distances.
Target recognition is an important task for many automatic systems based on imagery. The recent technique of active contours (snakes) is well adapted to the segmentation step when the recognition is made from the shape of the target. Classical segmentation strategies are generally edge-based in the sense that the segmentation is driven from an edge map of the scene. Consequently, these methods which are efficient with a certain class of problem could fail in presence of strong noise. We have recently proposed an original approach for the statistical segmentation of an object (statistical snake) for which the image is assumed to be made of two regions (the object and the background) composed of homogeneous intensity random fields. In this article, we characterize the quality of the segmentation as a function of the target resolution and noise level with two similarity measurements based on Hausdorff distance between the exact contour and the result of the segmentation.
It is well known that using different versions of a scene perturbed with different blurs improves the quality of the restored image compared to using a single blurred image. In this paper, we consider images perturbed with large defocus blurs. In the case where two different blurs are used, we characterize the influence of the relative diameter of both blurs on the restoration quality. We show that using two different blurs significantly increases the robustness of the restoration to a mismatch between the kernels used in designing the restoration algorithm and the actual blurs that have perturbed the image. We finally show that using three different kernels may not improve the restoration performance compared to the two-kernel approach, but still improves the robustness to kernel estimation.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.