SMYOLO: Lightweight Pedestrian Target Detection Algorithm in Low-Altitude Scenarios

Abstract

Due to the arbitrariness of the drone’s shooting angle of view and camera movement and the limited computing power of the drone platform, pedestrian detection in the drone scene poses a greater challenge. This paper proposes a new convolutional neural network structure, SMYOLO, which achieves the balance of accuracy and speed from three aspects: (1) By combining deep separable convolution and point convolution and replacing the activation function, the calculation amount and parameters of the original network are reduced; (2) by adding a batch normalization (BN) layer, SMYOLO accelerates the convergence and improves the generalization ability; and (3) through scale matching, reduces the feature loss of the original network. Compared with the original network model, SMYOLO reduces the accuracy of the model by only 4.36%, the model size is reduced by 76.90%, the inference speed is increased by 43.29%, and the detection target is accelerated by 33.33%, achieving minimization of the network model volume while ensuring the detection accuracy of the model.

Space-Air-Ground Integrated 6G Wireless Communication Networks: A Review of Antenna Technologies and Application Scenarios

“…LAPs can fly at an altitude of tens of meters up to a few kilometers (km), and their greatest strengths are essentially fast movements as well as their extreme flexibility [102,103]. Therefore, they can easily recharge or be replaced if needed.…”

YOLO-SS-Large: A Lightweight and High-Performance Model for Defect Detection in Substations

“…Therefore, many researchers have conducted research on lightweight methods in recent years. Zhang W et al combined depth-separable convolution with point convolution and added batch normalization layers [8], which accelerated model convergence. H Xu et al weighted each channel using the coordinate attention (CA) mechanism [9] to remove redundant features and effectively reduce the number of parameters.…”