Real-time analysis in the chemical industry is becoming increasingly important as science and technology develops. The reality that real-time analysis is an essential tool for process automation and remote control is encouraging the introduction of new analytical systems more convenient and more accurate.
The establishment of analytical systems based on chemical image analysis has attracted great attention in the fields of process automation, agriculture, biology and medicine as well as analytical chemistry worldwide for its great prospect of application in process analysis.
The amount of nickel in plating solution is analyzed by photoelectric colorimetry, complex titration, gravimetry, polarography, etc. However, the chemical analysis of these nickel plating solutions is only possible in laboratories and limited in large-scale processes due to the manual work throughout the process.
Kim Yong Ok, a researcher at the Faculty of Chemical Engineering, has proposed a methodology for applying AI technology to chemical image analysis of nickel plating solutions.
First, she performed the preprocessing of solution images using dynamic contrast enhancement technique and constructed an extensible private function for capturing representative pixels from the preprocessed image.
Next, she developed a mobile application to establish a nickel analysis system in nickel plating solution by using deep learning. In particular, she built a model that underwent mechanical learning into Android applications so that all processes from photography to analysis could be carried out in a single program.
© 2021 KumChaek University of Technology