Showing papers on "Object detection published in 1982"
TL;DR: Results provide converging evidence that semantic relations can be accessed from the results of a single fixation and were available sufficiently early during the time course of scene perception to affect the perception of the objects in the scene.
1,027 citations
TL;DR: A new technique for matching image features to maps or models which forms all possible pairs of image features and model features which match on the basis of local evidence alone and which is robust with respect to changes of image orientation and content.
Abstract: A new technique is presented for matching image features to maps or models. The technique forms all possible pairs of image features and model features which match on the basis of local evidence alone. For each possible pair of matching features the parameters of an RST (rotation, scaling, and translation) transformation are derived. Clustering in the space of all possible RST parameter sets reveals a good global transformation which matches many image features to many model features. Results with a variety of data sets are presented which demonstrate that the technique does not require sophisticated feature detection and is robust with respect to changes of image orientation and content. Examples in both cartography and object detection are given.
304 citations
Patent•
26 Jul 1982
TL;DR: In this paper, an object detector which incorporates a photoelectric detector in the vacuum flow path of a vacuum pickup system to serve as a device to determine whether an object has been successfully engaged, retained, and transported by a vacuum orifice, so that transport cycles may be modified or terminated by control circuitry in the event that the object has not been successfully retained and transported, thereby saving time and reducing damage to object, transport means or surface to which the object is transported.
Abstract: This disclosure relates to an object detector which incorporates a photoelectric detector in the vacuum flow path of a vacuum pickup system to serve as a device to determine whether an object has been successfully engaged, retained and transported by a vacuum orifice, so that transport cycles may be modified or terminated by control circuitry in the event that the object has not been successfully engaged, retained and transported, thereby saving time and reducing damage to object, transport means or the surface to which the object is transported.
47 citations
4 citations
Patent•
25 Mar 1982
TL;DR: In this paper, the authors proposed to detect moving objects with good accuracy by digitally processing the detection output of a magnetic detector, and removing DC components then detecting the output larger than the stored noise quantity in detecting the degree of proximity between relatively moving two objects.
Abstract: PURPOSE:To detect moving objects with good accuracy by digitally processing the detection output of a magnetic detector, and removing DC components then detecting the output larger than the stored noise quantity in detecting the degree of proximity between relatively moving two objects. CONSTITUTION:External magnetic fields are detected by using a magnetic detector 1 in order to detect the change in the physical quantity by moving objects. The output of the detector is inputted to an A/D converter 2. A central processing device 5 operates according to the program of a read only memory 6. The device 5 sets an initial period at the time Ta and preheats the ceramic detector at the time Tb. It removes background magnetic fields from the output of the detector 1 at Tc after the times Ta, Tb, and stores the max. value and max. change quantity of noise components in a temporary memory 7 at Te. When the detection output larger than said max. value and max. change quantity during stanbly of the time Tf is detected, it signifies the detection of a moving object. Thence, the extremal value point (time) and amplitude of the detection signal are detected and when the preset value is obtained, an object detection signal is outputted from the device 9.
2 citations
01 May 1982
TL;DR: A new type of array processor is presented which, coupled to an image store, works synchronously with the scan at a pixel rate of 8 MHz, and allows access to 16 points within a local neighbourhood.
Abstract: A new type of array processor is presented which, coupled to an image store, works synchronously with the scan at a pixel rate of 8 MHz. It allows access to 16 points within a local neighbourhood. A new value for the center point is evaluated in a cascade of look-up tables with two entries. A vast range of local transforms may therefore be programmed which may include nonlinear or logical operations. The system performs up to 10 transforms/sec yielding roughly 11O Moperations/sec on 6 bit pixel data. Each operation (= one table look-up) may however consist of several instructions. Examples of biological applications include detection of objects by texture and parallel tracking.
1 citations
01 Jan 1982
TL;DR: This paper summarizes the major research activities since the last Image Understanding Workshop in April 1981, which focused on symbolic matching of an image to a map or another image and included work on segmentation, texture analysis, object detection and description.
Abstract: : This paper summarizes our major research activities since the last Image Understanding Workshop in April 1981. Our research goals have been to develop a set of general techniques that can be applied to a number of application tasks. Our focus at the high level has been on symbolic matching of an image to a map or another image. This task is central to many applications, e.g., image based navigation, map-updating, and change detection. Our other research has been in support of this high level goal and has included work on segmentation, texture analysis, object detection and description. We have also worked with Hughes Research Laboratories in defining and developing special purpose hardware for IU tasks.
1 citations
01 May 1982
TL;DR: A high-speed algorithm has been devised to determine the range from a video camera to objects in its field-of-view, by simulating a variable parameter camera which views objects whose shapes, S/N ratios, and ranges may be varied.
Abstract: A high-speed algorithm has been devised to determine the range from a video camera to objects in its field-of-view. As the camera focus control is adjusted, the algorithm will detect each object feature as it comes into sharpest focus. The range from the image plane to the object can then be computed from the corresponding lens-to-image distance, which is known. The level-of-focus (LOF) algorithm is computationally simple, and may be performed in hardware in real-time for reasonable frame rates and pixel densities. The spatial resolution of the ranging information is a function of the lens F-stop, the sensitivity of the video sensors, the S/N ratio of the observed scene, and the range to the object. The LOF algorithm has been validated by simulating a variable parameter camera which views objects whose shapes, S/N ratios, and ranges may be varied.