A water quality analysis method based on a visible-light multispectral sensor

Purpose Traditional spectrophotometry is limited by complex operational procedures, high laboratory dependency and challenges in rapid on-site detection. To address these issues, this study introduces a water quality analysis method that combines a visible-light multispectral sensor with color reactions to determine the concentrations of six substances: ammonia nitrogen, nitrite nitrogen, sulfide, fluoride, formaldehyde, and total residual chlorine. Design/methodology/approach A method combining a visible-light multispectral sensor with color reactions was proposed. A constant-current-driven circuit for the light source, a multispectral data acquisition device and corresponding software were developed. The performance was evaluated using single- and multiwavelength regression analyses, and the results were compared with those obtained using traditional spectrophotometry. Findings The coefficients of determination (R2) were 0.9786–0.9947 for single-wavelength regression analyses and 0.9910–0.9999 for multiwavelength regression analyses for the six substances. Traditional spectrophotometry yielded R2 values of 0.9770–0.9970. Multiwavelength linear regression therefore offers a viable alternative, featuring streamlined instrument design, reduced costs and enhanced suitability for portable, on-site detection. Originality/value This study explores the potential of combining a visible-light multispectral sensor with color reactions for water quality analysis. Multiwavelength linear regression is identified as an effective alternative to traditional spectrophotometry, with significant potential for portable, on-site water quality assessments and substance quantification.