基于深度学习方法的N5/NiCrAlY涂层图像识别的研究
王明好, 王欢, 刘叡, 孟凡帝, 刘莉, 王福会

Research on Image Recognition for NiCrAlY Coating/N5 High-temperature Alloy System Based on Deep Learning Method
WANG Minghao, WANG Huan, LIU Rui, MENG Fandi, LIU Li, WANG Fuhui
表1 网络的结构参数
Table 1 Parameters of the nework
LayerConnection typeOutput shapeParam
Input layer(64, 64, 3)64×64×3
Conv layer_1ReLU(62, 62, 32)(3×3, 32)
Pooling layer_1(31, 31, 32)

MaxPooling,

stride=2

Conv layer_2ReLU(29, 29, 64)(3×3, 32)
Pooling layer_2(14, 14, 64)

MaxPooling,

stride=2

Conv layer_3ReLU(12, 12, 64)(3×3, 64)
Pooling layer_3(6, 6, 64)

MaxPooling,

stride=2

Full connected layer2304Full
Output layer3SoftMax