Image Processing When remote sensing data to digital format (Digital) are possible by using computer processing and digital analysis done. The process for increasing the quality of data and visual interpretation is performed. Also, post information or especially from the acquired image that all is done automatically by computer.
Analogue Images: Images such as aerial photographs by the photographed systems (camera) to come. Since the images of photographic film is used, then does not require any processing.
Digital images: are sensor images, which are made of a large number of small squares (pixels. Each pixel contains a digit number which is representative of its brightness. These types of images are called image raster. Raster images have row and column. A picture can be defined by two-dimensional function f (x, y) where X and Y are coordinates of the location and the value of “f” is intensity of brightness at every point. The term gray level light intensity also refers Monochrome images. Color images as well as a number of two-dimensional image will be formed. When X and Y values and the value f (x, y) values are discrete and finite expression, the image is called a digital image. Digital X and Y values to the Digital Sampling and the amount of f (x, y) to say Quantization. To display an image of an M * N two-dimensional array (matrix), which has M rows and N columns use. Amount of each element of the array represents the brightness intensity point. All the functions that will be implemented, each element of an array of 8-bit value that can be a value between 0 and 255 is. Zero indicates dark color (black) and 255 indicates the amount of light colors (white) is. For example, against the image that its size is 288 * 265 with a matrix that has 288 rows and 265 columns for display uses. Each pixel of the image has a value between 0 and 255. Values close to 255 points of light and dark areas have values close to 0. All image processing functions using these values and the necessary actions on the files do. Pixel values: Magnetic energy amount which a digital image obtain, forms Digit binary or Bits that is valuate from zero to twoexponents. The maximum number of light depends on the number of bits. Therefore, 8-bit, i.e., 256, depends on digital numbers that range from 0 to 255. That is why when you enter raster files from special sensor likes TM to software, it shows changes in light levels between 0 and 255. Resolution: Resolution depends on the number of pixels. With a 2-bit image, the maximum range of brightness is 22 = 4 which its range is varied from 0 to 3. In this case, the image has not accuracy (required resolution). 8-bit image has a maximum range of 256 and its variation is between 0 and 255 which has higher accuracy.
دامنه تغييرات رنگ
28 = 256
216 = 65536
224 = 16777216
Image processing methods: Gray-Level Slicing Suppose you want to determine total forest area of the province. Assuming that the aerial photos we have from the province, we can use segmentation of gray level. So that we can calculated the forest area to appear the gray levels, representative of forest and declining of intensity of other parts of image. Follow chart shows type of mapping the input image pixel values. Also, the following chart for mapping the input image pixel values can be used.
The chart above shows that the output files, generated by this diagram is a binary image.
Image restoration Most of the images which are recorded by satellite or radar, result in disturbances in the image that is due to scratches. Two major disturbances is multi-band images, banding and the lines lines which are ignored. 1 - Banding: the mistake which is occurred by the sensor in the record and transferring of data or change in the pixels, between rows can cause this wrong. 2 - Missed Lines (error in picture): the mistake which is occurred in the record and transferring of data and consequently, a row of pixels is lost in the image.
Increasing the accuracy Photos One of the important work that is done in image processing, is increasing of accuracy to see photos and visual interpretation. Many ways are existed to achieve this goal there, but most important is increasing of contrast and image filtering operation. Histogram of images In each digital image, pixel values show the image characteristics (such as brightness and clarity of it). Histogram of imagesis, in fact, graphical expression of brightness level. Brightness values (for example, 0-255) is expressed during the X axis and the frequency of each value is expressed in the Y axis.
Increasing divergence through extended numbers (DN) pixel Usually values range of image pixels with each bit (eg 8 bits here), is not between 0-255 and for example between 48 to 153. To increase contrast, values of pixels are extended to place 48 instead of 0 and 153 instead of 256. So divergence and also image quality goes up. This action is said linear stretch. Original pixel values (at the top) and stretched image in the bottom.