Fig. 1. Details of YOLOv5 network

Fig. 2. Overall structure of the proposed model

Fig. 3. ResCatPAN structure

Fig. 4. ResCat structure

Fig. 5. Overall structure of the cross-layer attention. (a) Complete flow of the cross-layer attention; (b) catt module of the cross-layer attention; (c) satt module of the cross-layer attention

Fig. 6. Distribution statistics of label boxes in the data set. (a) Distribution of the size of the label box; (b) distribution of the center points of the label box

Fig. 7. Effect analysis on hyperparameter $\gamma$. (a) Influence of hyperparameter $\gamma$ on detection performance for small object; (b) influence of hyperparameter $\gamma$ on detection performance for medium object; (c) influence of hyperparameter $\gamma$ on detection performance for large object; (d) influence of hyperparameter $\gamma$ on detection performance for object

Fig. 8. Heat map generated by the proposed CLAT module by using the Grad-CAM method

Fig. 9. Effect contrast of the proposed model and baseline on the test set. (a) baseline; (b) proposed model

Table 1. Contrast experiment

Table 2. Ablation experiment

Table 3. Sample distribution of the DIOR dataset

Table 4. Time-consuming comparison