An Effective Approach to Classify Terrain and Geographical Images Using Alexnet Algorithm Over VGG-16 Algorithm
V. Dhanushkodi, N. Bharatha Devi, G. Vinuja · 2023
This study will compare the AlexNet algorithm against the VGG-16 method in the classification of topography and geographical imagery. The dataset consists of eight different types of terrain which are represented by the 2400 total images in this collection: “asphalt,” “dirt,” “grass,” “floor,” “gravel,” “rock,” “sand,” and “wood chips”. The number of iterations for each group was determined to be 20 using CliniCal software. The novel Alexnet Algorithm having a higher accuracy of 91.55% with an accuracy of 88.38% and a significance level of p=0.011 (Independent group T-test p0.05), there is a statistically significant distinction between the two groups. Regarding the accurate classification of terrain and geographic images, Novel Alexnet Algorithm accuracy of 91.55% is superior to VGG- 16 algorithm accuracy of 88.38%.