Research on underwater object detection of improved YOLOv7 model based on attention mechanism
Open Access
- 16 December 2022
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 302-307
Abstract
Due to the complex content and blurry image of the images obtained in the underwater environment, which affect the detection results, resulting in the problems of false detection of objects and difficulty in feature extraction in conventional object detection models, this paper proposes an improved YOLOv7 model to improve the accuracy and real-time performance of underwater object detection models. Based on the YOLOv7 model, this model integrates the CBAM to enhance the weight of the feature information of the detection object in the spatial dimension and channel dimension, so that the feature information of the model object can be captured more easily and the detection accuracy is improved. The SPPFCSPC module is used to reduce the calculation amount of the model and improve the operation speed of the model while maintaining the receptive field. Through comparative ablation experiments, it is proved that the proposed method has higher detection accuracy in object detection in complex underwater environment than several detection models and YOLOv7 models.Keywords
This publication has 4 references indexed in Scilit:
- Underwater Object Detection based on YOLO-v3 network2021
- Lightweight Deep Neural Network for Joint Learning of Underwater Object Detection and Color ConversionIEEE Transactions on Neural Networks and Learning Systems, 2021
- Performance Prediction and Estimation for Underwater Target Detection Using Multichannel SonarIEEE Journal of Oceanic Engineering, 2019
- Crack detection in photovoltaic cells by interferometric analysis of electronic speckle patternsSolar Energy Materials and Solar Cells, 2012