Solar roads are roads embedded with solar panels which can converting solar energy radiated on the road into storable electricity. Over the recent years, pioneering solar road prototypes were tested in different regions around the world. Driven by demand, road photovoltaic production calculation, based on street view images (SVI) has been proposed (Liu et al., 2019). However, in addition to the high runtime overhead, this method cannot be applied to cities in which recent SVI are unavailable. At the meantime, researches on the estimation of road photovoltaic production of cities are rare, especially ones with spatially explicit inferences. This study proposes an innovative predictive model that can estimate road photovoltaic capacity of cities with urban features obtained from remote sensing images and other multi-source GIS data. As a scaffolding step, accurate estimation of potential road PV in 27 cities were calculated using SVI. Compared with the SVI approach, our predictive model is fast, robust and yet accurate as well. As a result, the spatial distribution of the potential energy production of solar roads for the 27 cities are mapped, which provides insights into which area should be prioritized for building solar roads. By analysing and comparing the estimated results and current vehicle energy demand, we propose different suggestions for the construction of photovoltaic roads for different types of cities. These suggestions may provide support for urban solar road planning in the course of adapting to cleaner energy sources. Additionally, data required by this predictive model is easy to access, which contributes to the universal applicability of this method.

太阳能道路是嵌入太阳能电池板的道路，可以将道路上辐射的太阳能转化为可储存的电能。近年来，开创性的太阳能道路原型在世界不同地区进行了测试。在需求的驱动下，已经提出了基于街景图像（SVI）的道路光伏产量计算（Liu et al., 2019）。然而，除了高运行时开销外，这种方法不能应用于最近 SVI 不可用的城市。同时，关于城市道路光伏发电量估算的研究很少，尤其是具有空间显性推断的研究。本研究提出了一种创新的预测模型，该模型可以利用遥感图像和其他多源 GIS 数据获得的城市特征来估计城市的道路光伏容量。作为脚手架步骤，使用 SVI 计算了 27 个城市的潜在道路 PV 的准确估计。与 SVI 方法相比，我们的预测模型快速、稳健且准确。因此，绘制了 27 个城市太阳能道路潜在能源生产的空间分布图，这有助于了解哪些区域应该优先建设太阳能道路。通过对估算结果和当前车辆能源需求的分析比较，针对不同类型城市的光伏道路建设提出了不同的建议。这些建议可为城市太阳能道路规划在适应清洁能源的过程中提供支持。此外，该预测模型所需的数据易于访问，这有助于该方法的普遍适用性。使用 SVI 计算了 27 个城市的潜在道路 PV 的准确估计。与 SVI 方法相比，我们的预测模型快速、稳健且准确。因此，绘制了 27 个城市太阳能道路潜在能源生产的空间分布图，这有助于了解哪些区域应该优先建设太阳能道路。通过对估算结果和当前车辆能源需求的分析比较，针对不同类型城市的光伏道路建设提出了不同的建议。这些建议可为城市太阳能道路规划在适应清洁能源的过程中提供支持。此外，该预测模型所需的数据易于访问，这有助于该方法的普遍适用性。使用 SVI 计算了 27 个城市的潜在道路 PV 的准确估计。与 SVI 方法相比，我们的预测模型快速、稳健且准确。因此，绘制了 27 个城市太阳能道路潜在能源生产的空间分布图，这有助于了解哪些区域应该优先建设太阳能道路。通过对估算结果和当前车辆能源需求的分析比较，针对不同类型城市的光伏道路建设提出了不同的建议。这些建议可为城市太阳能道路规划在适应清洁能源的过程中提供支持。此外，该预测模型所需的数据易于访问，这有助于该方法的普遍适用性。