July 02 Reviewed: August 2022
Keywords: adaptation of the brightness, mathematical model, adaptive filters. INTRODUCTION
Download 0.5 Mb. Pdf ko'rish
|
50 13 34 1-6
Keywords: adaptation of the brightness, mathematical model, adaptive filters.
INTRODUCTION As part of determining the degree of effectiveness of the image recovery algorithm, a number of experiments were conducted on two types of graphical information – static and dynamic. All the plots contain four frames, the first of which is a reference and represents a static image, and the next three are a dynamic video. A certain number of defects (from 1 to 6) were applied to each frame to simulate real damage to the images. The standard deviation (RMS) was used as a measure of the success of the recovery. The following are examples of test plots in the form of reference frames, with the original images on the left and the defects on the right. [2-4]. General testing methodology: 1. Open the video sequence using the "Upload" button. 2. On the reference frame, we select all areas of the image with defects. 3. For each defect, we set private processing settings in accordance with its surroundings. 4. We process the reference frame and search for control coefficients using the "Control" button. 5. We process the remaining frames of the video sequence using the "Apply" button. 6. Save the processed images using the "Save" button. 7. We analyze the original (without defects) and processed image to evaluate the COEX. 8. Repeat steps 1-7 for the other two video sequences. Since the experiment is carried out in two phases, its results will be grouped by the type of graphic information – first processing the reference frames of all three video sequences as static graphics, then the results of processing all frames in the video sequence in the form of a histogram of COEX systems reflecting the processing error for each defect as dynamic graphics. The adaptive median filtering algorithm is designed to attenuate more intense bipolar pulse interference, the probability of occurrence of pulses of which exceeds 0, 2 n p [5; 23-24]. In addition, this algorithm has the advantage that it distorts image details to a lesser extent that are not damaged by pulse noise. A feature of the adaptive algorithm is that, unlike a conventional median filter, it increases the size of the window under certain conditions, covering an odd number of pixels that the filtered image is scanned with. When implementing the algorithm, the following values of the intensities of pixels located within the window are measured, which, as before, can have any shape (rectangular, cross-shaped, etc.): the maximum intensity value; the minimum intensity value (brightness); the intensity value of the pixel occupying the central position in the window; the median of the sequence of pixels located in the the maximum allowable size of the filter window , which is set in the dialog by the number of pixels. The adaptive median filtering algorithm includes two branches: I and II. The task performed by the first branch is to determine whether the median is the result of an interference pulse (positive or negative) on the image, or not. If the condition < < is met, then it is assumed that the found value is not the result of the interference pulse acting on the image, and then the transition to the execution of the second branch of the algorithm is made. When executing the second branch of the algorithm, it is checked whether the intensity value of the pixel occupying the central position in the window is the result of an interference pulse (positive or negative) on the image, or not. If the condition < < is met, then it is assumed that the value is not the result of the interference pulse acting on the image, and the value, not the median value, is taken as the filtering result. Download 0.5 Mb. Do'stlaringiz bilan baham: |
Ma'lumotlar bazasi mualliflik huquqi bilan himoyalangan ©fayllar.org 2024
ma'muriyatiga murojaat qiling
ma'muriyatiga murojaat qiling