We develop a real time monitoring system for evaluating both frequency of posture change and respiration rate without any contact to the subject. This system detects the optical flow (apparent velocity) using a temporal-spatial local optimization method to evaluate the subject motion on a bed. The system consists of a CCD video camera, a parallel image processing board and a PC. The parallel image processing board includes 256 CPUs that calculates the optical flow of the whole image (256 × 240 pixels) within 150ms. Healthy young volunteers were participated to evaluate the system. We propose a function that includes information of respiration and body movement from the detected optical flows on the blanket area of breast. The periodic fluctuation synchronizes with the volunteer's respiration. The large peaks occur during volunteer's posture change.
This paper proposes a robust method for tracking many pedestrians viewed from an upper oblique angle. Persons are extracted while they are isolated, and they are tracked forward and backward. In tracking, if the region of a tracked pedestrian is ambiguous because of overlapping with other pedestrians, plural candidates for a pedestrian path are generated, and false candidates are eliminated in other frames. We make an experiment using real image sequences to show the effectiveness of the method.
In this paper, we propose an inspection system of subject with three dimensional surface and volume by image processing using the distribution of brightness and section curves. The distribution of brightness on image depends on the inclination of subject's surface from an optical axis of the camera unit. Therefore, it is possible to inspect the three dimensional surface based on it. The sample is classified into some patterns by the visual inspection of worker. Next, the brightness pattern of a sample is measured, and its standard index is calculated. The weighting coefficient is determined based on the standard deviation of width of bright area on the sample. The volume of inspection subject is measured using a line laser device, and a control chart method is applied to maintain the quality of shape and volume of subjects through the production line. The proposed method is applied to soldering on electric device, and weight of lead in soldering is estimated from the measured volume.
We have developed an inspection system that reads a ring bar code stamped on data recording discs, such as CDs and DVDs, with a high degree of accuracy and at high speed by using image processing technology. The system uses the following two methods to achieve high reading speed : a method of bar code position detection which uses a reduced number of reference pixels, and a method of brightness addition data production which determines the bar code thickness based on a polar projection table. The system allows us to inspect ring bar codes with a minimum element width of 130 μ m and 22 characters in approximately 250 ms (less than 1/3 the time required by the conventional method).
While techniques of 3D-Graphics to visualize 3D-shape has increasingly come into use in ordinary PC-environment, emerging issue is how to effectively acquire the information of 3D-shape as the input. In this paper we propose a new PC application that computes 3D-shape of an object or a scene. The algorithm we have employed for the shape recovery is based on the stereo vision technique and the developed system is kept rather simple without requiring a special tool. We describe our prototype system and its performance examined through experiments. Particularly, we show that the system works on a smooth surface shape that is usually difficult to input for conventional CAD systems.
Described here is a visual navigation method for navigating a mobile robot along a man-made route such as a corridor or a street. We propose an image sensor, named HyperOmniVision, with a, hyperboloidal mirror for vision based navigation of the mobile robot. This sensing system can acquire an omnidirectional view around the robot in real time. In the case of the man-made route, the road boundary between the ground plane and wall appears as a close-looped curve in the image. By making use of this optical characteristic, the robot can avoid obstacles and move along the corridor by tracking the close-looped curve with an active contour model. Experiments that have been done in a real environment are described.
A method which automatically generates the descriptions of both static objects and moving elastic objects in dynamic scenes is newly developed. The descriptions consists of parameters of mixture Gaussian distributions, sequentially estimated from the intensity histogram information by means of the EM (Expectation-Maximization) algorithm. Experimental results using indoor and outdoor scenes show the effectiveness of the method. This method is expected to be in use as the first stage of a variety of the object recognition systems applied for various industrial applications, such as visual surveillance, Intelligent Transport System (ITS), intelligent robots and welfare support systems.
We present a practical bin-picking system that uses stereo vision from multiple viewpoints and the three-dimensional recognition using salient features and uncertainty evaluation. First, the salient features of objects in images are extracted. Under the guidance of the features, the two-dimensional object candidates are detected. And the feature correspondence among stereo images is made based on the global features of an object model. Second, the three-dimensional features are reconstructed using the stereo vision results. Next, with the object features and the three-dimensional measurements, the uncertainty is evaluated. Then the results from the different viewpoints are integrated to determine the optimal results for each object by using the uncertainty and the three-dimensional position of object candidates. Based on this practical three-dimensional vision, an intelligent robot system is realized to pick up objects in a bin.