Machine Vision as a Materials Conservation Technology
- PDF / 453,046 Bytes
- 6 Pages / 404.64 x 643.86 pts Page_size
- 17 Downloads / 200 Views
Although machine vision is often only seen in an industrial context, it can be interpreted for archaeological and architectural contexts as equivalent to real-time, digital close range photogrammetry, encompassing all other applications for which close range photogrammetry is relevant. In practice, every machine vision system is built around the principles of digital image processing (DIP) [4,5]. Implementing DIP requires the use of a suitable Camera System, Digitizer, Computer Hardware and Software. The medium of information is a digital image of the subject to be evaluated, say that of a masonry wall. In computer vision paradigm, the image of the subject is viewed as a collection of "objects", where an "object" is an area on the image that differs from others by its "gray scale". In the digital form, the image is broken down into a large number of tiny cells ("pixels"), each of which is assigned an appropriate gray scale value chosen from a wide range (typically 256). Subsequently, this digital information is stored and manipulated, as desired, in the computer. Indeed, this is the overriding principle by which object measurement is made possible. The medium may also be a polychrome image, instead of monochrome; however, not all systems have polychromatic capabilities. DIP progresses in a sequence of four main procedures, namely, image acquisition, image enhancement, image measurement, and image analysis. Image acquisition involves capturing the subject in a digital image. The image can be acquired through conventional light photography and then fed to the computer through a Digital Scanner. Image enhancement reduces noise and heightens visual distinction among different objects on the image, thus facilitating measurement. This is accomplished by passing the digitized image information through a set of appropriate Digital Filters. Selection of the set and sequence of filters is a critical exercise. Image measurement involves calibrating the spatial scale and performing metrological tasks, such as, distance measurement, area measurement, angle measurement, etc. Finally, image analysis utilizes the measurement data as input, and produces statistical results, such as, density and location distribution of objects. Material degradation reveals itself to visual recording because of the localized changes in color and contrast on the surface of the structure. Once recorded, such an image can be effectively processed and analyzed digitally using a computer software yielding results that are faster, more informative and reliable than the conventional eye estimation. This paper will discuss results of an exploratory investigation and a case study of employing machine vision as a preservation technology. EXPERIMENTAL EXPLORATION Scope The purpose of this study was to inquire into a novel strategy for objective and quantitative recording and characterization of material conditions in historic monuments. In an exploratory effort, images of historically significant artifacts have been subjected to digital image processing with
Data Loading...
