Methods for automatic target recognition by use of electro-optic sensors: introduction to the feature issue

摘要

The world of image processing and pattern recognition is vast. This Feature issue is intended to record current trends in the field and to encompass areas that are traditionally outside the domain of Applied Optics. It is the editors' hope that in doing so, increased interactions with the larger image processing and pattern recognition community are facilitated and that the readership will find the enhanced scope of the journal to be of interest. With this goal in mind, we are pleased to bring forward a collection of papers by experts in academia, industry, and government, from both the U.S. and many other countries around the world. The focus of the special issue is on methods for recognizing spatial, temporal, and spectral patterns in multidimensional images, with application to object detection and recognition. As illustrated by the diversity of the papers, the work in this area spans the management and indexing of large databases of images; extraction of object information from aerial imagery; a multitude of techniques for object detection, including the exploitation of hyperspectral imagery, the use of video for detection and tracking, color image processing, and last but not least the exploitation of three-dimensional (3D) data. The collection of papers in this issue reflects the many avenues of research in this rich and diverse field. The special issue is organized into four subsections: (i) management and processing of surveillance images and databases, (ii) techniques for target detection, (iii) target tracking, and (iv) multidimensional image recognition (including the use of color, video and 3D processing). We now provide a brief overview of each of these sections.