Road Plant Image Classification Algorithm Based on Segmentation Prior

As an important part of urban ecology, it is of great significance to grasp and analyze the detailed distribution of road plants for both scientific research and landscape planning. However, due to the advancement of urbanization, the types and numbers of road plants are increasing, and it is difficult to obtain accurate and timely plant distribution information by traditional field research and remote sensing techniques. In order to solve this problem, we proposes a deep learning-based road plant image classification algorithm, which identifies plants in streetscape data by multi- label image classification to obtain the distribution of road plants. In order to solve the problem of feature information loss in the network, we increases the channel capacity of the bottleneck structure in the network, and introduces the group convolution structure to optimize the network performance. Due to the complexity of the image background, we proposes a road plant image classification algorithm based on a priori segmentation by training the binary classification segmentation model of plants and background through transfer learning. This makes the classification model pay more attention to the area of road plants in the image and improves the classification accuracy by 5.16%.