A Novel Method for Chemical Oxygen Demand (Cod) Estimation Using Alexnet

The Chemical Oxygen Demand (COD) analysis is a cumulative parameter that indicates the extent of organic pollution in water. In the COD analysis, the reagents reacted with the organic matter in the water, producing distinct colour intensities, with the degree of reaction varying according to the organic load. The COD concentration was determined by measuring the absorbance of the resulting colour using the spectrophotometric method. This analysis, which is traditionally performed using a spectrophotometer, is conducted for the first time in this study without the use of a spectrophotometer, using mobile device support and AlexNet. Chemicals necessary for COD analysis are produced, and the reaction procedures are executed. The resulting mixtures were photographed with the camera of the mobile device and input into the developed software. The mobile application predicts the image class using an AlexNet architecture with Leaky Rectified Linear Unit (LeakyReLU) activations in the background. The network, trained on a dataset of 504 images across 21 classes, each containing 24 images per class, displays the estimated results with 96.19% accuracy, 97.85% precision, 96.19% recall, 97.05% F1-score, and an R² value of 0.9892. These results are consistent with those obtained using a spectrophotometer. The proposed method accurately estimates the COD concentration without using the values measured by the spectrophotometer at 420 and 600 nm. The proposed method successfully estimates the COD concentration using only images. The proposed method is advantageous because it eliminates the need for a spectrophotometer in COD analysis.