<p>Polymetallic nodules are commonly found on deep-sea floors. To enhance the accuracy of resource assessment for these nodules, a new evaluation method that utilises image enhancement techniques to estimate their abundance based on key parameters, such as quantity, coverage rate and size, was proposed. The deep-sea mining environment presents challenges, as the image of polymetallic nodules often exhibits low-contrast and blue–green colour, making identification difficult. To address this, a method that combines the Sea-thru algorithm with YOLOv8 was proposed. This approach estimates scene depth based on varying light attenuation rates in water, improving the recognition accuracy of polymetallic nodules in deep-sea imagery. Additionally, to enhance the estimation accuracy of nodule mineral coverage affected by sediment occlusion, employed a binary morphological dilation method. By modelling polymetallic nodules as regular circles and assessing their height above the seabed, a quantitative expression for the actual grain size of buried nodules was provided, further enhancing the accuracy of mineral resource assessments. This method provides a solid foundation for decision-making in mining vehicle operations and route planning, thereby enhancing mining efficiency.</p>