Object Search System in the Image

The topic of the article relates to the direction of digital image processing. The main topic of the article is the method between the frame difference. The task is to get acquainted with the basic principles of constructing monitoring systems that are used to detect objects of various kinds by analyzing images. Under the analysis of images in this paper, one should understand the pixel subtraction of two images. In this paper, describe the typical problems that concern a system of this kind. Also, features of the use of other methods that can be used for analysis in such systems are described.

The purpose of this method is to analyze the methods of processing images that need to be taken into account in order to implement one of the methods of such a system in practice. To implement it in practice, the method of segmentation is selected, later on. Terms of work with hardware solutions, the data they work with and the data that are used by the user after processing can vary greatly. For example, if you are dealing with a binary black and white image, you need to historically simplify it compared to the random, when the input across all-round raster with paid cross-colors from one to the other. The image can be enlarged with a static camera location, so all the frames will have each other background with possible differences in lighting level. On the other hand, the camera located on the moving object may be different with respect to it. The noise level can vary greatly. Natural phenomena, such as rain, snow, fog, wind, can make a noticeable element in the primary real estate. Please describe what manipulations can be made with one or more frames that should be in this case.

Most single-dimensional signal processing methods (for example, a median filter) are applied to two-dimensional signals that are an image. Some of these one-dimensional methods are considerably complicated with the transition to a two-dimensional signal. Signal processing is widely used for Fourier transforms, as well as wavelet transform and Gabor filter. Image processing is divided into processing in the spatial region (brightness transformation, gamma correction, etc.) and frequency (Fourier transform, etc.)

This article is an overview on the topic of image processing, which deals with the topic of searching for moving and static objects through image analysis. It mentions typical tasks for image processing, editing and retouching, compensation for loss of sharpness, text recognition, computer vision and methods for processing them: gradation, spatial, arithmetic logic. One of the arithmetic-logical methods, the method of segmentation was described more widely. Those signs that must be taken into account when constructing a processing system based on it: texture, shape, motion, brightness.

The next stage in the study of this topic will be the construction of a mathematical model in the high-level MATLAB mathematical calculations system.

Ref. 4, fig. 1.