Automated mapping of the mean particle diameter characteristics from UAV-imagery using the CNN-based GRAINet model

Abstract

This study uses the GRAINet CNN approach on UAV optical aerial imagery to analyze and predict grain size characteristics, specifically mean diameter (dm), along a gravel river point bar in Šumava National Park (Šumava NP), Czechia. By employing a digital line sampling technique and manual annotations as ground truth, GRAINet offers an innovative solution for particle size analysis. Eight UAV overflights were conducted between 2014 and 2022 to monitor changes in grain size dm across the river point bar. The resulting dm prediction maps showed reasonably accurate results, with Mean Absolute Error (MAE) values ranging from 1.9 cm to 4.4 cm in tenfold cross-validations. Mean Squared Error (MSE) and Root Mean Square Error (RMSE) values varied from 7.13 cm to 27.24 cm and 2.49 cm to 4.07 cm, respectively. Most models underestimated grain size, with around 68.5% falling within 1σ and 90.75% falling within 2σ of the predicted GRAINet mean dm. However, deviations from actual grain sizes were observed, particularly for grains smaller than 5 cm. The study highlights the importance of a large manually labeled training dataset for the GRAINet approach, eliminating the need for user-parameter tuning and improving its suitability for large-scale applications. HIGHLIGHTS Assessing the effectiveness of a cutting-edge deep learning algorithm in predicting mean diameter (dm) from an individual UAV-based orthophoto.; Creating robust maps to predict spatial and temporal variations in mean dm across one entire point bare over time.; The method enhances efficient decision-making capabilities by reducing reliance on laborious and resource-intensive field probing techniques.; The method demonstrates robustness against light condition impacts.;

Keywords

• convolutional neural networks
• deep learning
• fluvial environments
• grainet algorithm
• particle size distribution
• uavs