Moving Object Detection And Recognition Algorithm Research

Image object detection is a task valuable for both theoretical research and practical usage, and is served to many fields of defense as well as civil economic construction. It is also image processing and image understanding hot direction. With application fields expanding, people pay more attention to its practical value. Moving object detection and recognition algorithm research sequence images or video images. We process and detect sequence images in order to find interesting object and extract it. The task of image recognition is to recognize object. From the object we can obtain its property which can be used to recognize object information.Mathematical morphology is a nonlinear theory for image and signal analysis and processing, which can estimate many information of the geometrical structure in the signal and concurs with our instinctive perceptual system. For this reason, it has been paid more and more attentions and developed rapidly. In this paper, main works needed by an object detection system are deeply researched based on a thorough investigation of the basic theories of mathematical morphology and its current usages in object detection. We emphasize research one of important techniques reference with moving object detection: image feature extraction. In this paper, we present a method edge detection based on morphology. We propose edge detection method of pseudo top-hat transformation combining with multi-scale structure elements. Compared with other traditional methods, the method can obtain abundant edges and complete object edges.When object detection, one common thing is that vidicon is stillness and focus is root. Here, the background is fastness. In this situation, background subtraction is often used for object detection. For simple background subtraction, when background movement, for example, moving object high frequency in background or illumination change, we can't extract true background which is looked as reference image. Thereby, we can't exactly segment moving object. In order to get over this problem, we have to update background in time.However, as a result of illumination influence, most images have shadows. The exit of shadow will disturb object detection. The difference between detect object and true object. It has great influence to after process, for instance, object recognize or behavior estimate. Gaussian model is unable to eliminate shadow, so it is important to find a method which can eliminate shadow. In recent years, many methods are put forward. In this paper, normalized cross correlation is applied to eliminate shadow. The speed is greatly improved. Moreover, we present chroma difference to eliminate shadow. Chroma difference makes full use of color information. The result of shadow elimination is better than single brightness. In this paper, we combine normalized cross correlation with brightness and chroma difference to eliminate shadow. The examination illuminate that the result is good.