دانلود رایگان مقاله شخیص خودکار آرایه های خورشیدی فتوولتائیک در تصاویر هوایی

Abstract

The quantity of small scale solar photovoltaic (PV) arrays in the United States has grown rapidly in recent years. As a result, there is substantial interest in high quality information about the quantity, power capacity, and energy generated by such arrays, including at a high spatial resolution (e.g., cities, counties, or other small regions). Unfortunately, existing methods for obtaining this information, such as surveys and utility interconnection filings, are limited in their completeness and spatial resolution. This work presents a computer algorithm that automatically detects PV panels using very high resolution color satellite imagery. The approach potentially offers a fast, scalable method for obtaining accurate information on PV array location and size, and at much higher spatial resolutions than are currently available. The method is validated using a very large (135 km2) collection of publicly available (Bradbury et al., 2016) aerial imagery, with over 2700 human annotated PV array locations. The results demonstrate the algorithm is highly effective on a per-pixel basis. It is likewise effective at object-level PV array detection, but with significant potential for improvement in estimating the precise shape/size of the PV arrays. These results are the first of their kind for the detection of solar PV in aerial imagery, demonstrating the feasibility of the approach and establishing a baseline performance for future investigations.

7. Conclusions and future work

We investigated a new approach for the problem of collecting information for small-scale solar PV arrays over large areas. The proposed approach employs a computer algorithm that automatically detects solar PV arrays in high resolution (60.3 m) color (RGB) imagery data. A detection algorithm was developed and validated on a very large collection of aerial imagery (P135 km2 ) collected over the city of Fresno, CA. Human annotators manually scanned and annotated solar PV locations to provide ground truth for evaluating the performance of the proposed algorithm. Performance was measured in a pixel-based and object-based manner, respectively, using PR curves. In the case of object-based scoring, the algorithm was also scored based on how well it can identify the shape and size of the true panel object. The results demonstrate that the algorithm is highly effective on a per-pixel basis. The PR measures indicate it can detect most of the true PV pixels while removing the vast majority of the non-PV pixels. The object-based PR curves indicated that the algorithm was likewise effective at object detection, however, it was far less effective at estimating the precise shape/size of the PV arrays. The results presented here are the first of their kind for distributed PV detection in aerial imagery, and demonstrate the feasibility of collecting distributed PV information over large areas using aerial or satellite imagery. This may ultimately yield a faster, cheaper, and more scalable solution for the large scale collection of distributed PV information, and potentially information for other aspects of energy production and consumption as well. While the results here demonstrate the promise of this approach to information collection, several challenges remain as opportunities for future work.