An Enhanced Randomly Initialized Convolutional Neural Network for Columnar Cactus Recognition in Unmanned Aerial Vehicle imagery

A new deep learning model, the Enhanced Randomly Initialized Convolutional Neural Network (ERI-CNN), achieves high accuracy in recognizing columnar cacti from unmanned aerial vehicle (UAV) imagery. With 98% accuracy, 97% precision and recall, and an F1-score of 97.5%, the model outperforms existing architectures like InceptionV3 and modified LeNet-5 when tested on a dataset of over 20,000 remote sensing images from Mexico's Tehuacán-Cuicatlán Valley. This advancement in plant recognition technology supports environmental conservation efforts by improving surveillance and monitoring of protected ecosystems.

Atitallah, S. B., Driss, M., Boulila, W., Koubaa, A., Atitallah, N., & Ghézala, H. B. (2021). An Enhanced Randomly Initialized Convolutional Neural Network for Columnar Cactus Recognition in Unmanned Aerial Vehicle imagery. Procedia Computer Science, 192, 573–581. https://doi.org/10.1016/j.procs.2021.08.059