Research on Underwater Image Classification and Enhancement Based on Traditional Image Processing Algorithms

The degradation of underwater images is typically caused by absorption and scattering of light during its propagation in water, leading to color shifts and a decrease in contrast. To effectively address the underwater image enhancement problem, this paper systematically constructs a model. First, based on the average RGB color values of the image, the average brightness channel values, and clarity, corresponding thresholds are set to identify color deviations, low light, and blurring phenomena, thereby establishing an image classification model. Inspired by the Jaffe-McGlamery model, an improved degradation model is proposed, incorporating the light transmission model, ambient light, and blur effects. Subsequently, a convolutional image enhancement model is introduced, combining clarity features, Gaussian blur kernels, and sharpening filters, integrated with advanced algorithms such as color correction and histogram equalization. The model primarily relies on traditional image processing methods and physical models, making it easy to implement and highly suitable for real-time applications. Experimental results show that the enhanced images exhibit significant improvements in color, contrast, naturalness, and object prominence.