Level Measurement And Control Using Single Digital Camera Based On Image Processing In Matlab

Abstract

This study examines the effectiveness of image-based level measurement and control using a single digital camera by comparing various segmentation models for liquid level detection. The main goal of this study was to evaluate some common approaches used in segmenting images from single cameras such as simple thresholding, adaptive thresholding, Otsu's thresholding and global thresholding with manual adjustment that will identify liquid levels using blob inside it correctly. All segmentation models consistently show black colour for liquid and white for objects/blobs. In addition, there is a significant correlation between the size of detected objects/blobs and liquid depth suggesting potential use of diameter variations as an indirect measure of changes in liquid depths. Thus, this article presents a technique to compute depth of liquid based on object diameter measured from detected object or blob using height of camera above tank so that one can improve industrial applications involving development of improved level measurement and control systems.