畜禽养殖疾病诊断智能传感技术研究进展

李奇峰,李嘉位,马为红,高荣华,余礼根,丁露雨,于沁杨

Research Progress of Intelligent Sensing Technology for Diagnosis of Livestock and Poultry Diseases

LI QiFeng,LI JiaWei,MA WeiHong,GAO RongHua,YU LiGen,DING LuYu,YU QinYang

表4 疾病自动诊疗主流代表方法

Table 4 Mainstream representative methods of automatic disease diagnosis

诊断依据 Diagnostic basis 技术手段 Technical means 主要用途 The main purpose 识别精度 identification accuracy 代表文献 The literature 形态特征 Morphological characteristics SVM分类 SVM classification 深度学习分类 Deep learning classify 疾病诊病/敏感特征提取 Disease diagnosis/ feature extraction 97.5% -97.8% 91.7% [7] [20] 动作姿态 Action & posture 三轴加速传感器 Acceleration sensor SVM分类 SVM classification 视频特征识别 Video feature recognition 深度学习分类 Deep learning classify 动物姿态与健康状态评价 Posture and health evaluation 80% 94% 90% 99% [60] [9] [22][46] [47] 位置特征 Location feature 视频光流法 Video optical flow UWB定位模块法 Positioning module 深度学习目标检测 Deep learning Target Detection 运动量检测、采食量、饮水量估计等 Exercise detection, estimated feed intake and water consumption 较低 Lower 最高 Highest 较高 Higher [52] [101] [49][50] 声音特征 Voice characteristics 发声图谱分析 Voice atlas analysis 偏度聚类分析 Skewness cluster analysis 疾病诊断及应激评价 Disease diagnosis & stress evaluation 88% 95% [71][75] [70] 呼吸特征 Respiratory characteristics 稀疏光流法 Sparse optical flow WIFI感知法 WIFI perception method 深度图像分析 Depth image analysis 穿戴式传感器 Wearable sensor 代谢评价及舒适度评价 Metabolic & Comfort evaluation 98.58% 约98% 85.3% — [63] [92] [48] [96] 体温特征 Body temperature 植入传感器 Implanted sensor 红外测温法 Infrared Thermometry 疾病诊断及应激评价 Disease diagnosis & stress evaluation ±0.05℃ ±1.5℃ [80][87] [86][88]