5.17.2. Image enhancement

Image enhancement is mainly carried out to improve the quality of images. Often, the images do not have the optimum contrast so objects are not clearly identified visually. Enhancement is made to make the visual appearance of image better. It is the modification of images to improve the interpretability through human vision. The main purpose of image enhancement is to increase the contrast in a low contrast image. It does not add any information to the original image but it enhances the visual appearances of already captured features.

Image enhancement improves how an image looks, making it easier for people to identify objects and features within it. It focuses primarily on adjusting contrast, which refers to the difference between the lightest and darkest parts of an image. By enhancing the contrast, images become clearer and more distinct, allowing viewers to better discern various elements. Importantly, this process does not change the actual data captured in the image; rather, it helps highlight the existing features.

Think of image enhancement like adjusting the brightness and contrast on your TV or computer screen. When a movie appears too dark or washed out, you can tweak these settings to make details more visible. Similarly, image enhancement processes make sure all the necessary visual details in an image pop out better, just like adjusting the settings for a clearer view.

Before image enhancement is done it is necessary to understand the image characteristics through its histogram. A histogram is a graphical representation of the DN values (i.e., 0-255) in an image that are displayed along x-axis, while the frequency of occurrence of these values is plotted on y-axis. Image histogram is a way to portray the information present on an image. In raw imagery, the useful data often occupies only a small portion of the available range of DN values (256 levels in an 8 bit image).

An image histogram is a bar graph that represents the distribution of pixel intensity values in an image. The x-axis shows the possible pixel values (ranging from 0 to 255 for an 8-bit image), while the y-axis represents how many times each pixel value appears in the image. Understanding the histogram helps identify how much of the image data is effectively being used and whether adjustments are needed to improve contrast and other features. For example, if many pixel values cluster at one end (either dark or light), it indicates poor contrast that could benefit from enhancement.

Imagine the histogram as a scoreboard showing how different scores are distributed in a basketball game. If most players score low points (like clustering at dark values), the game seems dull or uninteresting, much like an image with low contrast. Enhancing the image is similar to encouraging more players to score higher points, creating a more exciting game. By adjusting the histogram, you can make the image engaging and informative.

Contrast enhancement involves changing the original DN values so that more available range is utilised, thereby increasing the contrast between the objects and their background. It basically improves the interpretability for human viewing, and provides enhanced input to be used for image processing. There are many different techniques and methods of enhancing the contrast, and details can be found in Garg (2022). The linear contrast enhancement is the most popular technique used for image enhancement.

Contrast enhancement adjusts the pixel values in an image so that the full range of intensity levels is utilized. When this range is expanded, the difference between light and dark areas becomes more pronounced, making objects easier to distinguish. One widely used method is linear contrast enhancement, which involves identifying the minimum and maximum pixel values and stretching the current range of values to fit the full spectrum available (0-255). This enhancement makes light areas lighter and dark areas darker, improving image clarity.

Consider how a simple cooking recipe can change based on ingredient adjustments. If you add more spices, it can transform a bland dish into a flavorful one. Similarly, enhancing the image contrast is like adding ‘spices' to the image—it brings out the hidden features and ensures everything is visible and appealing to the viewer, just like a well-seasoned meal is delightful to eat.

The image transformation is the creation of new image by using some mathematical function on the original images. The image transformation will normally yield synthetic images which are very useful for specific applications, as they enhance certain features of interest. Some examples of transformations include; simple arithmetic operations, Vegetation Indices (VI), Normalised Difference Vegetation Index (NDVI), Principal Component Analysis (PCA) and Tasselled Cap Transformations (TCT).

Image transformation involves applying mathematical formulas to an original image, resulting in new images that emphasize specific features. For example, transformations like NDVI can highlight vegetation health by comparing the reflectance of near-infrared and red light. By analyzing these transformed images, users can better study and monitor environmental conditions. Such transformations often provide clearer insights into the image data, revealing patterns or changes that may not be apparent from the original image.

Think of image transformations as filtering your Instagram photos. When you apply a filter to a picture, certain colors or features are enhanced, making the image more visually appealing or informative. For instance, a filter might make the greenery in a landscape richer and more vibrant. In the same way, image transformations can help researchers and scientists highlight specific data within their images, making it easier to analyze trends and changes.

--