1. Introduction
Humans need the vision to form perception and understanding of their environment. The goal of computer vision is to adapt human vision by giving computers the ability to electronically understand and perceive an image [2]. Computer vision provides an output in the form of image understanding when given a digital image as input. It combines various ideas, techniques, and concepts from image processing, artificial intelligence, and computer graphics to obtain information from an input image [2] [3]. Computer vision aims to obtain data and information from digital images. However, the process used to manipulate the properties of images using a digital computer so that computer vision can produce efficient and accurate image understanding is called image processing [1]. If we consider image processing at one end of a continuum and computer vision at the other, it would be very difficult to draw a clear line between the two. However, we can divide the continuum into low, medium, and high levels. Image sharpening, pre-processing, contrast enhancement, and noise reduction, etc. are some primitive operations that can be considered as low-level processing. It produces a simple output in the form of an image. Image processing at the medium level includes image segmentation and individual object detection operations. Medium level processing extracts the image’s attributes (e.g., contours, objects, and edges, etc.) as output. High-level processing produces cognitive outputs after analysing the detected objects in the image [1].
1.1. History of Image Processing
After John Neumann developed the first modern digital computer in 1940 which has two basic components Memory and Conditional branching, it took almost 20 more years to finally build a digital computer powerful enough to handle the tasks of digital image processing. After a lot of successive advancements and developments in the computer from 1940, we finally had a Digital computer in 1960 which can solve the image processing problems of practical significance. This can be referred to as the birth year of digital image processing and it all started at Jet Propulsion Laboratories (California) in 1964. Engineers and scientists received pictures of the moon from Ranger 7 on-board Camera and these pictures were digitally processed to improve the inbuilt distortion by utilizing computer techniques. Another area where image processing was successfully applied was to find solutions for Machine perception-related problems. Extraction of the information from an image for the further processing best suitable for the computer was almost similar to the human vision and interpretation of information in an image [1]. At the beginning of the nineties, Artificial intelligence was efficiently combined with image processing and the techniques of pattern recognition. In today’s world powerful computers combined with innovative digital equipment like cameras, sensors, and tablets, we have been successful in utilizing the benefits of methodologies such as Hidden Markov models, Fuzzy set reasoning, Natural language processing, and neural networks [7].
1.2. Literature Review
In the past few years, many scholars are drawn toward Computer vision and image processing because these two-term can be implemented in a wide range of disciplines. The latest developments in the field of Image processing and computer vision are covered in this literature. Scholars have tried to establish a relationship between how a human eye perceives and analyzes data in an image. computer vision, however, cannot produce exact similar interpretations from the image because of many limitations [3]. The biggest limitation of computer vision is information loss from 3-D world to 2-D images. Image capturing devices such as cameras used for computer vision records information in the form of a 2-D image [2]. This image is then processed with the help of various image processing tools. However, the accuracy and efficiency highly depend on how strong the algorithm is, control, and monitoring of the system performance and strength.
Another literature explains and reviews image processing using MATLAB. Improvement in the inherited properties of a digital image with the help of a digital computer is called digital image processing [8]. The motivation behind this improvement or quality enhancement can be categorized mainly into three applications. Enhancement of qualities such as color, contrast, and noise reduction can be used for better perception for the human eye. We have many industries and military applications such as Autonomous cars and machines which require better quality images in order to make autonomous decisions. The third application would be efficiently storing and transmission of a digital image. The author has explained many latest developments in methodologies such as image restoration, image enhancement, image segmentation, Gabor filter, etc. in the mentioned literature.
1.3. Image processing
The process of modifying and altering the pictorial information with a digital computer so that the information may appear better to human eyes, extraction of information from an image so it can be efficiently used in autonomous machines or robots is known as image processing [8]. The input before the process is always an image however the output may vary according to the level of processing and the application it is used for. On a low level of image processing, the basic inbuild flaws are removed and the quality of visuals is increased so that it may appear modified and improved to the human eye. Image processing operations are not limited to the images which are in the visibility range of humans, in fact, image processing also applies to the electromagnetic energy spectrum below and above the human vision range. Sometimes we do not process images just to enhance the visual appearance but to extract the characteristics or features available such as the type of objects present in the image, size of the objects, etc. This level of image processing is called mid-level processing. The third and final level of image processing deals with the cognitive function almost like human vision [1].
2022-1-24-1643060058