A customizable automated container-free multi-strip detection and line recognition system for colorimetric analysis with lateral flow immunoassay for lean meat powder based on machine vision and smartphone

• Published in: Talanta (Oxford) 2023-02, Vol.253, p.123925-123925, Article 123925
• Main Authors: Zhao, Guanao, Liu, Sijie, Li, Guo, Fang, Wentai, Liao, Yangjun, Li, Rui, Fu, Longsheng, Wang, Jianlong
• Format: Article
• Language: English
• ISSN: 0039-9140; 1873-3573
• DOI: 10.1016/j.talanta.2022.123925

Ractopamine (RAC) and clenbuterol (CLE) are feed additives with adverse effects of consuming too much to food safety. It is necessary to develop an efficient and accurate colorimetric analysis method for immune-based detection of RAC and CLE. Traditional human-vision-based colorimetric analysis for lateral flow immunoassay (LFIA) is non-quantifiable and low-in-automation, while container-based and analysis-instrument-based methods are unrepeatable and high-cost. Therefore, a container-free colorimetric analysis method was developed with LFIAs image captured in dark background under smartphone flash. A multi-strip detection algorithm based on contours extraction, as well as line recognition algorithm based on grayscale projection of LFIA was developed. Finally, relative grayscale (RGS) of lines were calculated and then input into editable fitting curves to estimate concentrations. Results showed the multi-strip detection algorithm reached 98.85% and 93.70% of Recall and intersection over union (IoU), while the line recognition algorithm reached 95.07% and 97.95% of Recall and color similarity, respectively. As a result, an App was fabricated through employing LFIA of RAC and CLE, with colorimetric analysis accuracy of 98.25% and 94.50%, respectively. This study provides a container-free multi-strip colorimetric analysis method with low-cost and illumination robustness, which is a substitution for container-based and single-strip colorimetric analysis methods.