Scholarlink

<p>Crucian carp is a commonly consumed fish, and evaluating its freshness is essential for ensuring food safety. This study aims to develop an adaptive, cross-modal method for non-invasive freshness detection in crucian carp, utilizing flexible haptics and multi-angle visual perception. An inspection system was designed incorporating a multi-angle industrial camera and a flexible pressure sensor, enabling non-destructive freshness detection through the acquisition, processing, and fusion of multi-angle images and tactile signals from the fish’s meat. Experimental results showed that, as freshness declined, visible changes occurred in the pupil, iris, gill, and belly color, while the tactile elasticity of the meat decreased. Based on these observations, a multitasking shared model was developed, using flexible tactile and multi-angle image signals as inputs. The method achieved 99.54% accuracy, demonstrating the potential of adaptive cross-modal fusion technology for freshness detection in crucian carp, offering an innovative approach to quality monitoring.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Adaptive Cross-Modal Non-Invasive Detection of Crucian Carp Freshness Based on Flexible Haptics and Multi-Angle Visual Perception

  • 期刊论文

  • 发表日期：

    2025年9月20日

查看全文
  • Bin Chen,
  • Jiahao Yu,
  • Huanhuan Feng,
  • Vilai Rungsardthong,
  • Xiaoshuan Zhang

摘要

Crucian carp is a commonly consumed fish, and evaluating its freshness is essential for ensuring food safety. This study aims to develop an adaptive, cross-modal method for non-invasive freshness detection in crucian carp, utilizing flexible haptics and multi-angle visual perception. An inspection system was designed incorporating a multi-angle industrial camera and a flexible pressure sensor, enabling non-destructive freshness detection through the acquisition, processing, and fusion of multi-angle images and tactile signals from the fish’s meat. Experimental results showed that, as freshness declined, visible changes occurred in the pupil, iris, gill, and belly color, while the tactile elasticity of the meat decreased. Based on these observations, a multitasking shared model was developed, using flexible tactile and multi-angle image signals as inputs. The method achieved 99.54% accuracy, demonstrating the potential of adaptive cross-modal fusion technology for freshness detection in crucian carp, offering an innovative approach to quality monitoring.