Complex scene-oriented ship black smoke detection method and system

By constructing a dataset of ship black smoke in complex scenarios, adaptive preprocessing, and improving the YOLOv5s model, the problem of ship black smoke detection in complex scenarios was solved, and efficient identification and accurate detection were achieved under adverse weather conditions.

CN121921745APending Publication Date: 2026-04-24SHANGHAI JIUHENG TECHNOLOGY CO LTD
0 Cites 0 Cited by

Patent Information

Application Number
CN202610067319.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-01-19
Publication Date
2026-04-24

AI Technical Summary

Technical Problem

Existing technologies are insufficient for efficiently identifying ship smoke in complex scenarios, especially under adverse weather conditions, and cannot meet actual monitoring needs.

Method used

A dataset of ship black smoke containing different resolutions and weather conditions was constructed. Adaptive histogram equalization preprocessing was adopted, the detection head of the YOLOv5s model was replaced with a dynamic detection head, and a cross-dimensional feature fusion module was added to the Neck network. Scale-aware, spatial-aware, and task-aware attention modules were used, and the loss function was improved to the minimum point distance intersection-union loss function.

Benefits of technology

The model's robustness and recognition ability against ship smoke were improved, environmental interference was reduced, and the detection accuracy and small target detection ability for multi-scale smoke targets were enhanced.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention discloses a complex scene-oriented ship black smoke detection method and system, and the method comprises the steps: constructing a ship black smoke data set containing different weather conditions, and carrying out the image enhancement preprocessing through an adaptive histogram equalization method; on the basis of a YOLOv5s model, a detection head of the YOLOv5s model is replaced by a dynamic detection head comprising a scale perception attention module, a space perception attention module and a task perception attention module, a cross-dimension feature fusion module is added in a Neck network, and a loss function is replaced by an MPDIOU loss function based on the minimum point pitch intersection-to-union ratio; and training and evaluating the improved model by using the preprocessed data set. According to the method, the detection precision and robustness of the model on the ship black smoke in complex weather and multi-scale scenes are effectively improved, the recognition capability on small targets and diffusion form black smoke is enhanced, and reliable technical support is provided for ship emission monitoring.
Need to check novelty before this filing date? Find Prior Art