Showing papers on "Object detection published in 1987"

Segmentation and Classification of Range Images

Abstract: The recognition of objects in three-dimensional space is a desirable capability of a computer vision system. Range images, which directly measure 3-D surface coordinates of a scene, are well suited for this task. In this paper we report a procedure to detect connected planar, convex, and concave surfaces of 3-D objects. This is accomplished in three stages. The first stage segments the range image into ``surface patches'' by a square error criterion clustering algorithm using surface points and associated surface normals. The second stage classifies these patches as planar, convex, or concave based on a non-parametric statistical test for trend, curvature values, and eigenvalue analysis. In the final stage, boundaries between adjacent surface patches are classified as crease or noncrease edges, and this information is used to merge compatible patches to produce reasonable faces of the object(s). This procedure has been successfully applied to a large number of real and synthetic images, four of which we present in this paper.

Three-dimensional model matching from an unconstrained viewpoint

Abstract: It is demonstrated that the affine viewing transformation is a reasonable approximation to perspective. A group of image vertices and edges, called the vertex-pair, which fully determines the affine transformation between a three-dimensional model and a two-dimensional image is defined. A clustering approach, which produces a set of consistent assignments between vertex-pairs in the model and in the image is described. A number of experimental results on outdoor images are presented.

Knowledge-Based Aerial Image Understanding Systems and Expert Systems for Image Processing

Abstract: This paper discusses roles of artificial intelligence in the automatic interpretation of remotely sensed imagery. We first discuss several image understanding systems for analyzing complex aerial photographs. The discussion is mainly concerned with knowledge representation and control structure in the aerial image understanding systems: a blackboard model for integrating diverse object detection modules, a symbolic model representation for three-dimensional object recognition, and integration of bottom-up and top-down analyses. Then, a model of expert systems for image processing is introduced that discusses which and what combinations of image processing operators are effective to analyze an image. Various information about image processing techniques is used to find efficient and reliable image analysis processes. In general, two kinds of knowledge, that is, knowledge about objects and about analysis tools (i. e., image processing techniques) are required to realize versatile photointerpretation systems.

Multiresolution object detection and delineation

Abstract: This paper describes methods of detecting and extracting (“delineating,” i.e., segmenting) compact objects from an image. The methods are designed for implementation on an exponentially tapering “pyramid” of processors, and require only O (log n ) time for an n by n image. Objects are detected using an “interest measure” derived from local comparisons between fathers and sons in the pyramid. They are extracted by a top-down tree growing process in which the leaves of the tree are the pixels belonging to the detected object.

Fractal Image Models And Object Detection

Abstract: This paper discusses a new technique for object detection that uses fractals to model the natural background in a visible image. Our technique is based on the fact that fractal-based models have been found to be good models for natural objects as well as images of natural objects. On the other hand, man-made objects are decidedly not self-similar and therefore fractal-based models are not good models for man-made objects and their images. The technique adaptively fits a fractal-based model and a 2-D autoregressive model over the image and the fractal dimension and model-fit errors are used to identify regions of anomalous dimension and high error. Thus the technique uses a dual approach to object detection by modeling and deemphasizing the natural background instead of explicitly modeling and identifying the man-made object. Results are shown for a real image.

Pulse radar method and apparatus for detecting an object

Abstract: An object detection method and apparatus in which, in order to distinguish only an echo wave which is returned back from an object of interest, such as an underground buried object, from those echo waves returned back from another object of objects, the observation signal is divided (an electromagnetic wave as an echo wave) into portions, and the signal portion is converted into a corresponding frequency region to evaluate that spectral distribution and computes frequency parameter values from the spectrum distribution. The object of interest is detected by comparing the reference data of various fields with the parameters of the object of interest, extracting only an echo wave returned back from the object of interest and displaying a corresponding image.

Polyhedra Recognition by Hypothesis Accumulation

Abstract: A new method is presented for the recognition of polyhedra in range data. The method is based on a hypothesis accumulation scheme which allows parallel implementations. The different objects to be recognized are modeled by a set of local geometrical patterns. Local patterns of the same nature are extracted from the scene. For the recognition of an object, local scene and model patterns having the same geometrical characteristics are matched. For each of the possible matches, the geometric transformations (i.e., rotations and translations) are computed, which allows the overlapping of the model elements with those from the scene. This transformation permits the establishment of a hypothesis on the location of the object in the scene and the determination of a point in the transformation space. The presence of an object similar to a model involves the generation of several compatible hypotheses and creates a compact cluster in the transformation space. The recognition of the object is based on the detection of this cluster. The cluster coordinates give the values of the rotations and the translations to be applied to the model such that it corresponds to the object in the scene. The exact location of this object is given by the transformed model.

The Acquisition Probability for a Minimum Distance One-Class Classifier

Abstract: An approximation for the acquisition probability for a minimumdistance one-class classifier is derived. An exact expression for theacquisition probability is dependent upon the operatingcharacteristics in the distance space, the number of targets detected,and the number of other objects detected. An approximateexpression replaces the operating characteristics curve by a singlepoint. Experimental results are presented to demonstrate thevalidity of the approximation. Combinatorial techniques can be usedwhen only the total number of objects detected is known. All ofthese results can be extended to include the multitarget, multipleshotcase.

Model-based knowledge-based optical processors

Abstract: An efficient 3-D object-centered knowledge base is described. The ability to on-line generate a 2-D image projection or range image for any object/viewer orientation from this knowledge base is addressed. Applications of this knowledge base in associative processors and symbolic correlators are then discussed. Initial test results are presented for a multiple degree of freedom object recognition problem. These include new techniques to achieve object orientation information and two new associative memory matrix formulations.

Detection of target trajectories using the Hough transform

Abstract: Recently developed Hough transform techniques for curved object detection are applied to the detection of target trajectories. Multitarget missile trajectories are considered with noise present and missing data.

Target State Estimation in a Multitarget Environment Using Multiple Sensors

Abstract: An approach for target detection/state estimation that views theproblem as one of a finite state search over the target parameterspace is presented. This approach allows for a natural way toassociate different types of measurements, such as frequency andcoherence from multiple sensors, and also for dealing with multipletargets, dropouts, and clutter. We describe our model and present acomputationally efficient search algorithm for target detection andtarget state estimation in a multitarget environment based on thismodel. The results of a two-sensor, multitarget computer simulationare discussed.

Object detection and experimental designs

Abstract: A new object detection classifier is developed by combining the supervised learning model, hypothesis testing techniques, and the robust F -statistic in conjunction with the analysis of variance (ANOVA) process. The visual equivalence of two similarly dimensioned images is interpreted in terms of the F -statistic of two standard patterns resulting from transformations by a set of randomly generated rules. The object registration is related to the target and background portions of the reference template. Using the F -statistic as the test statistic, the object detection procedure is reduced to a simple ANOVA process. The multiple-hits problem is also resolved using either the distance between the test statistic and the threshold values or a set of floating threshold values associated with various confidence levels. The techniques introduced here are particularly effective if the background noise is non-gaussian, contaminated gaussian, or gaussian with an unknown variance. In these situations the optimum linear processors, which are best object extractors in a known gaussian environment, perform poorly. Computer simulations are used to verify the theory developed in the paper and to evaluate the computational simplicity of the procedure and demonstrate the deterioration in performance of linear processors in contaminated gaussian noise while the new procedure, introduced in the paper, remains efficient.

Passive infrared burglar sensor equipped with visual field interruption monitor mechanism.

Abstract: In a passive sensor of the type in which infrared energy emitted by an object inside a predetermined monitor field is detected by an infrared sensing element (6) and an alarm is sounded in response to the detection signal, the present invention discloses a passive type infrared burglar sensor characterized in that a man or an object coming close to a sensor main body is detected by an infrared sensor (7) which exhibits its detection function independently of the object detection function inside the monitor field described above, and any vision interference with the object detection function inside the monitor field is monitored.

Object Detection Using Their Multispectral Properties

Abstract: Detection of objects is an important task of computer vision systems. In this paper we present the development of an object detection system to be useful while analyzing multispectral images. In this formulation general knowledge about spectral characteristics of the objects to be detected is utilized in the search for their location in an image. Efficiency of the system is derived by using a hierarchical framework with pyramid data structure to store multiresolution, multispectral copies of an image. At every level of processing a fuzzy cluster analysis algorithm is utilized to uncover the membership of individual picture elements. These membership values are used with the general knoweldge of spectral properties of objects to guide the search for their locations. The methodology is tested using several experiments involving multispectral satellite and aerial images. The system is shown to be successful in efficient detection of objects such as rivers, roads, and various types of buildings.

Efficient CFAR detection of line segments in a 2-D image

Abstract: A computationally efficient algorithm is proposed for detecting line segments in an image of additive, i.i.d. (independent, identically distributed) Gaussian noise. Meteors, satellites, or other moving objects may be optically detected using the algorithm. A CFAR (Constant False Alarm Rate) characteristic is designed into the algorithm to give equal probabilities of false alarm for all streak lengths. Compared to the 2-D optimum matched filter approach, the algorithm loses 2 dB in signal-to-noise ratio, but requires hundreds of times less computation.

Knowledge Based Sar Images Exploitation

Abstract: One of the basic functions of SAR images exploitation system is the detection of man-made objects. The perfor-mance of object detection is strongly limited by perfor-mance of sementation modules. This paper presents a detection paradigm composed of an adaptive segmentation algorithm based on a priori knowledge of objects followed by a top-down hierarchical detection process that generates and evaluates object hypotheses. Shadow information and inter-object relationships can be added to the knowledge base to improve performance over that of a statistical detector based only on the attributes of individual objects.

Knowledge based object detection using SAR images

Abstract: One of the basic research problems in analyzing synthetic aperture radar (SAR) imagery is the detection of man-made objects. This paper presents a detection paradigm composed of an adaptive segmentation algorithm based on a priori knowledge of objects followed by a top-down hierarchical detection process that generates and evaluates object hypotheses. At the end of this process the most likely hypothesis for the object is selected. The hierarchical structure allows the use of shadow information and inter-object relationships to improve performance over that of a statistical detector based only on the properties of individual objects.

3 - Identification et codage d'objets en un passage par analyse d'image

Abstract: This paper describes the coding phase that precedes a pattern recognition in a system of automatical object detection . The shapes to be analysed are modelled by simple geometrical figures (rectangle, ellipse) ; this identification is made by a one pass calculation of the smaller number of geometrical characteristics (moment order 0, 1 and 2) . This method gives automatically a classification of all the objects present in the field and defines the zero level of coding. When the objects contain points of ramification, their shape are modelled with several primary figures, a higher level of coding is then defined. This procedure leads to a more or less precise representation of the object to be recognized ; the choice of the level representation depends on the current application . All the steps, computation of geometrical properties, identification of parameters and construction of an arborescent pattern of structure are obtained in one pass mode. This method is particularly well suited when the image analysis is sequential (Une by line) and avoids a storage of the total image.

Incorporation Of Map Information In Object Detection

Abstract: This paper elucidates the viability of utilizing map information effectively in the interpretation of aerial images. Motivation for using ancillary information is the desire to harness knowledge intrinsic to a study area. A knowledge-based scheme is presented that resembles post-classification sorting in principle, but exceeds it in sophistication. The prototype is data-driven, with the knowledge configured in a production system format. Information derived from maps can be utilized to support or refute the presence of an object detected by analyzing image data. Several experiments were conducted to evaluate various belief maintenance schemes applied to refining object detection results produced by the analysis of image-only data.

Multi-Sensor Processing: Object Detection And Identification

Abstract: Following a brief review of multi-sensor image processing techniques for obstacle detection, we consider a new method to employ range data to extract object regions of interest from an outdoor natural scene. Our emphasis and intent is scene analysis and object identification in the face of 3-D distortions using range sensor data. The range image is segmented into background/nonbackground pixels based on line-by-line processing. Non-background pixels are clustered together by a projection-based technique to determine possible regions of interest. Following extraction of object regions of interest, we can merge the range and other sensor image data to obtain multi-sensor images of the regions in the field of view. Emphasis in this paper is given to the aforementioned tasks; however, for completeness, the remainder of the full processor is briefly described. The use of a hybrid optical/digital processor for such tasks is noted.

Quelques Cas Critiques De Detection Des Images Astronomiques.

Abstract: The faint object detection plays always an important part in astronomical image analysis. Classical adapted filtering must be carefully used. We have to take into account the background variations, some defects and the bright object neighbourhoods. We describe in this paper the methods used in two cases of image processing for electronographic plates taken at the Cassegrain focus of the CFH Telescope with the Large Field Electronic Camera.

Automatic recognition of isolated or occluded planar objects by a two steps processing

Abstract: As an application of robot vision to automatic recognition of industriel parts, we present a method which is capable of recognizing isolated or partially occluded two-dimensional objects Starting from contour following, global features are measured and compared to stored data in order to make a preselection of the models In the case of isolated object, the preselection leads to an immediate identification of objects Otherwise, overlapping boundaries are detected and used to accomplish the recognition of occluded objects By the combination of overlapping boundaries and the object's contour, a sequential segmentation algorithm is applied to find the complete contour of no hidden objects As soon as we have obtained the complete contour of an object, we can use once again the global features to perform identification processing This method has the advantage of being simple,rapide and efficient Il works well even when the number of objects to be recognized is large