Based on Hybrid Densenet-121 with Support Vector Machine Algorithm for Lettuce and Chili

In this paper, to improve the recognition accuracy of Densenet-121, we combine Densenet-121 with SVM to generate a new network Densenet-121-SVM. In this paper, the two datasets are combined to generate a new dataset named Mixed-dataset, which has four plants, namely: Lettuce, Chili, Chinee apple, and Lantana, and this dataset is used to train the network. Finally, the recognition accuracies of SVM, Densenet121 and Densenet-121-SVM are 84.144%, 96.211% and 98.181%, respectively, and it can be seen that the accuracy of Densenet-121-SVM is improved by 14% and 1.97% compared to SVM and Densenet-121, respectively.