MTD-YOLOv5: Enhancing marine target detection with multi-scale feature fusion in YOLOv5 model.

Underwater light attenuation leads to decreased image contrast. This reduction in contrast subsequently decreases target visibility. Additionally, marine target detection is challenging due to multi-scale problems from varying target-to-device distances, complex target clustering, and noise from waterborne particulates.To address these issues, we propose MTD-YOLOv5.Initially, we enhance image contrast with grayscale equalization and mitigate color shift issues through color space transformation.We then introduce a novel feature extraction module, PCBR, combining max pooling and convolution layers for more effective target feature extraction from the background.Furthermore, we present the Multi-Scale Perceptual Hybrid Pooling (MHP) module.This module integrates horizontal and vertical receptive fields to establish long-range dependencies, thereby capturing hidden target information in deep network feature maps. In the Labeled Fishes in the Wild test datasets, MTD-YOLOv5 achieves a precision of 88.1% and a mean Average Precision (mAP[0.5:.95]) of 49.6%.These results represent improvements of 2.6% in precision and 0.4% in mAP over the original YOLOv5.

A topology control algorithm for fusion networks based on link quality.

To extend the network life, an ordinal potential game is introduced into network topology control, and a network topology control algorithm based on the Theil entropy measure is designed by improving the revenue function. The revenue function is based on the Theil entropy measure, whose factors are in the residual energy of the nodes and their neighbors. By using a primary payoff function that considers all initial network factors and a secondary one that addresses connectivity in case of reduced residual energy, a segmentation function can calculate the model's payoff. Simulation experiments show that compared to the existing game algorithms of 3DK-RNG, DEBA, and EFPC, the proposed algorithm can effectively eliminate redundant links, reduce node degree and network link length, balance node energy consumption, enhance network load, and extend the network life cycle.•A Network Topology Control Algorithm Based on the Fusion of Thiel Entropy and Ordinal Potential Games.•A network topology control method for extending the network lifecycle is successfully established.•It proposes that the performance of the fusion topology control algorithm is significantly superior to other single algorithms in the paper.