A Similarity Method of Color Frequency Histogram With Dynamic Reference Frame for Image Visual Titrations

Abstract

Visual titration is an important method for detecting the main and minor components in the samples. In visual titration, the image features are easily affected by the dichroism of the indicator, ambient light, and bubbles in the solution, which adversely affect the precision of titration endpoint recognition. This article proposes a similarity method of image color frequency histograms with dynamic reference frames (SHDRFs) for addressing the problems of color frequency histograms of time-varying and time-shifting characteristics in image visual titrations. With this method, the similarity was redefined as the difference between the color frequency histograms of the two images with a simple and efficient calculation. The titration curve was established based on the similarity values, and the titration endpoint was located by the threshold line of the similarity. To prove feasibility and effectiveness of this method, six zinc electrolyte samples were configured for titrations. Compared with the concentrations of the samples, there was no significant systematic deviation of the concentrations of sulfuric acid and zinc ions between the results detected through this method and their corresponding true concentrations through a one-sample $t$ -test at the 95% confidence level.