Waste steel fine classification method based on inception-v3 model and transfer learning

• Main Authors: MIAO QUANMING, ZHOU ZITIAN, LOU BAO, WU YUE, SI YUNHENG, LI YONG, ZHANG SHIWEI, XIE PENG, GU YI
• Format: Patent
• Language: Chinese; English
• Published: 21.03.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS

The invention discloses a scrap steel fine classification method based on an inception-v3 model and transfer learning, and belongs to the technical field of scrap steel. The method comprises the following steps: S1, acquiring different types of scrap steel images according to the types of required scrap steel, and dividing the scrap steel images into a training set, a verification set and a test set; s2, a convolutional neural network Inception-v3 model is adopted as a pre-training model, and image features are obtained by using a feature extraction model of the convolutional neural network Inception-v3 model; s3, establishing a Softmax regression model; s4, training a scrap steel image automatic identification and classification model; and S5, testing the scrap steel image automatic identification and classification model. According to the method, the scrap steel fine classification model is trained by adopting a method of combining the inception-v3 pre-training model and transfer learning, so that the requi