Intelligent water level detection method based on deep learning

• Main Authors: CHEN JIAXIN, MA TIEHUA, WU ZHIBO, WANG WENBO, WANG YU, PEI DONGXING, SUN CHUANMENG, YUAN YUE, KONG XIANGNIAN
• Format: Patent
• Language: Chinese; English
• Published: 03.10.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; PHYSICS

The invention relates to the technical field of water level detection, and particularly discloses an intelligent water level detection method based on deep learning, which comprises the following steps: firstly, providing a multi-level feature fusion method for strengthening small-scale features to improve a deep learning YOLOv5 algorithm, and strengthening the ability of the algorithm to capture small targets such as water gauge scales; secondly, fusing an improved compression activation network (RankSENet module) and a Bottleneck Transform module to further improve the sensing capability of the scale of the water gauge; and finally, a brand new water level elevation resolving scheme is provided, accurate water level elevation information can be obtained only by utilizing part of water gauge scale anchor frame information, and the robustness of the detection method is greatly improved. 本发明涉及水位检测技术领域，具体公开了一种基于深度学习的水位智能检测方法，首先，提出强化小尺度特征的多层级特征融合方法来改进深度学习YOLOv5算法，强化算法对水尺刻度此类小目标的捕捉能力；然后，融合改进压缩激活网络（RankSENet模块）与Bo