Rainfall data have a profound significance for meteorology, climatology, hydrology, and environmental sciences. However, existing rainfall observation methods (including ground-based rain gauges and radar-/satellite-based remote sensing) are not efficient in terms of spatiotemporal resolution and cannot meet the needs of high-resolution application scenarios (urban waterlogging, emergency rescue, etc.). Widespread surveillance cameras have been regarded as alternative rain gauges in existing studies. Surveillance audio, through exploiting their nonstop use to record rainfall acoustic signals, should be considered a type of data source to obtain high-resolution and all-weather data. In this study, a method named parallel neural network based on attention mechanisms and multi-scale fusion (PNNAMMS) is proposed for automatically classifying rainfall levels by surveillance audio. The proposed model employs a parallel dual-channel network with spatial channel extracting the frequency domain correlation, and temporal channel capturing the time-domain continuity of the rainfall sound. Additionally, attention mechanisms are used on the two channels to obtain significant spatiotemporal elements. A multi-scale fusion method was adopted to fuse different scale features in the spatial channel for more robust performance in complex surveillance scenarios. In experiments showed that our method achieved an estimation accuracy of 84.64% for rainfall levels and outperformed previously proposed methods.

The Forest Resources Assets Departure Audit (FRADA) is an important measure for the Chinese government to protect and develop forest resources, as well as to promote carbon peaking and carbon neutrality. Explicit audit content and scientific audit method are the basic premises for conducting an accurate audit. In this study, landscape ecology methods were employed to learn the spatiotemporal properties of forest resources, with the support of spatial information technology such as Geography Information System (GIS) and Remote Sensing (RS). A FRADA system from the perspective of landscape pattern index and ecological vulnerability was constructed. A case study was conducted for the evaluation of the proposed FRADA System. Lichuan, a county-level city in Hubei Province, China that is rich in forest resources was selected as the audit sample area, and the tenure period from 2015 to 2019 was regarded as the tenure cycle. The results showed that: (1) Before and after the tenure cycle, the number of patches (NP), patch density (PD) and largest patch index (LPI) of forest landscape showed a downward trend. Meanwhile, the decreased quantity and quality of forest landscapes showed inadequate protection of large forest patches with significant ecological value. In general, it was not beneficial for the sustainable development of forest resources in the view of ecology. (2) The landscape shape index (LSI) and connectivity index (CONNECT) showed an upward trend, which indicates that the shape of the forest landscape is changing in a complex way, and the connectivity has been improved. While, fractal dimension (FD) and fragmentation index (FN) showed a downward trend, indicating that the fragmentation trend of the forest landscape has been alleviated and the forest resources have a continuous and concentrated distribution. (3) The vulnerability of the forest landscape was significantly aggravated. Regarding administrative areas under the jurisdictions of Nanping Township, Jiannan Town and Fobaoshan Development Zone were particularly disturbed, and required special attention. The research results enrich and develop the theories and methods of FRADA and provide suggestions for the sustainable development of forest resources in Lichuan.