基于机器视觉多维特征融合的禽肉品质自动检测及分级方法
By constructing a multi-task parallel convolutional neural network model MQ-MultiNet, and combining a multi-task loss function and a fuzzy logic decision module, the problems of subjectivity and poor robustness of traditional poultry grading methods are solved, and efficient multi-dimensional detection of poultry quality is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-03-30
- Publication Date
- 2026-07-17
AI Technical Summary
Traditional poultry grading methods rely on human sensory evaluation, which is highly subjective, inefficient, and unable to meet the demands of the high-end market. Furthermore, existing machine vision methods are not robust enough to handle complex lighting environments and cannot achieve precise detection of multi-dimensional quality indicators.
A multi-task parallel convolutional neural network model, MQ-MultiNet, was constructed. By combining a multi-task loss function and a fuzzy logic decision module, automated grading of poultry meat quality was achieved through image acquisition, preprocessing, and multi-dimensional feature fusion.
It improves the accuracy and efficiency of poultry meat quality testing, reduces the number of parameters, lowers computational complexity, and enables simultaneous testing of meat color, fat distribution, and tenderness, meeting the demands of the high-end market.
Smart Images
Figure CN122416433A_ABST