Sensing Urban Land-Use Patterns By Integrating Google Tensorflow And Scene-Classification Models

Yao Yao,Haolin Liang,Xia Li,Jinbao Zhang,Jialv He

DOI: https://doi.org/10.48550/arXiv.1708.01580

2017-08-05

Abstract:With the rapid progress of China's urbanization, research on the automatic detection of land-use patterns in Chinese cities is of substantial importance. Deep learning is an effective method to extract image features. To take advantage of the deep-learning method in detecting urban land-use patterns, we applied a transfer-learning-based remote-sensing image approach to extract and classify features. Using the Google Tensorflow framework, a powerful convolution neural network (CNN) library was created. First, the transferred model was previously trained on ImageNet, one of the largest object-image data sets, to fully develop the model's ability to generate feature vectors of standard remote-sensing land-cover data sets (UC Merced and WHU-SIRI). Then, a random-forest-based classifier was constructed and trained on these generated vectors to classify the actual urban land-use pattern on the scale of traffic analysis zones (TAZs). To avoid the multi-scale effect of remote-sensing imagery, a large random patch (LRP) method was used. The proposed method could efficiently obtain acceptable accuracy (OA = 0.794, Kappa = 0.737) for the study area. In addition, the results show that the proposed method can effectively overcome the multi-scale effect that occurs in urban land-use classification at the irregular land-parcel level. The proposed method can help planners monitor dynamic urban land use and evaluate the impact of urban-planning schemes.

Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

The paper attempts to address the issue of how to quickly and accurately automatically detect urban land use patterns in the context of accelerated urbanization in China. Traditional methods are relatively expensive in terms of time and cost, so there is a need to develop an efficient method to perceive urban land use patterns through remote sensing images. Specifically, this study aims to solve the following problems:

1. **Multi-scale effect**: The spatial structure heterogeneity exhibited by remote sensing images at different scales leads to classification difficulties.
2. **Irregular plots**: The basic land units in cities are usually irregular plots or Traffic Analysis Zones (TAZ), which pose challenges for feature extraction and classification.
3. **Semantic gap**: Traditional object recognition methods find it difficult to capture high-level semantic information, thus affecting classification accuracy.

To address these issues, the authors propose a method that combines deep learning and scene classification, utilizing the Google TensorFlow framework and Convolutional Neural Networks (CNN) for feature extraction, and classifying urban land use types through a Random Forest classifier. This method can effectively overcome the multi-scale effect, improve classification accuracy, and perform exceptionally well in handling irregular plots.