SWDet: Anchor-Based Object Detector for Solid Waste Detection in Aerial Images

Abstract

As we all know, waste pollution is one of the most serious environmental issues in the world. Efficient detection of Solid Waste (SW) in aerial images can improve subsequent waste classification and automatic sorting on the ground. However, traditional methods have some problems, such as poor generalization and limited detection performance. This paper presents an Anchor-based Object Detector for Solid Waste in Aerial Images (SWDet). Specifically, we construct Asymmetric Deep Aggregation (ADA) network with structurally re-parameterized asymmetric blocks to extract waste features with inconspicuous appearance. Besides, considering the waste with blurred boundaries caused by the resolution of aerial images, this paper constructs Efficient Attention Fusion Pyramid Network (EAFPN) to obtain contextual information and multi-scale geospatial information via attention fusion. And the model can capture the scattering features of irregular shape waste. In addition, we construct the dataset for Solid Waste Aerial Detection (SWAD) by collecting aerial images of solid waste in Henan Province, China, to validate the effectiveness of our method. Experimental results show that SWDet outperforms most of existing methods for solid waste detection in aerial images. Code is available at https://github.com/shenhaibb/SWDet.

Image Processing System for the Detection of Recyclable Solid Waste

“…In addition, it should be noted that this system applies a unique work strategy to detect recyclable solid waste and help avoid pollution or accumulation of garbage. Likewise, this proposed system differs from other systems carried out by its methodology, for example, the research work developed by [15], where the researchers proposed to develop an object detector system based on the detection of solid waste by means of aerial images. Reaching as a result an efficiency of 89.14%, but this system does not run the tests in real situations but has only tested it with images that have been acquired online, limiting its use and losing confidence in its use by people.…”

“…Today society takes of great importance the correct handling of recyclable solid waste by the great waste caused by humans that does not allow the preservation of the terrestrial ecosystem, therefore the function of recycling plays a fundamental role throughout the world and has been proposed since before, that is why tool systems have been developed, such as: In [15], The researchers determined that pollution by accumulation of solid waste garbage is a problem that covers different social aspects with respect to environmental measures, over the years this problem of solid waste becomes more serious in the world and it is necessary to efficiently detect recyclable solid waste by aerial images to give it a second chance with respect to its reuse.…”

A construction waste landfill dataset of two districts in Beijing, China from high resolution satellite images

“… 21 and the SWAD dataset provided by Liming Zhou et al . 22 AerialWaste is an illegal landfill detection dataset containing manually labeled airborne, WorldView-3, and Google Earth imagery. The AerialWaste dataset focuses on illegal landfill detection and includes manually labeled airborne, WorldView-3, and Google Earth imagery.…”

The Cost of Urban Renewal: Annual Construction Waste Estimation via Multi-Scale Target Information Extraction and Attention-Enhanced Networks in Changping District, Beijing

“…There is a severe shortage of publicly available datasets designed explicitly for construction waste identification, and the existing datasets adhere to different standards. The commonly used datasets for construction waste extraction include the Aer-ialWaste dataset [47] and the SWAD dataset [48]. Still, they exhibit several shortcomings in practical applications: (1) The AerialWaste dataset lacks the annotated information required for semantic segmentation, as its classification is based on the presence of solid waste rather than on each pixel in the image.…”