【摘要】 为实现绿色低碳环保的公路工程勘测，研究提出了基于改进增强深度残差超分网络模型的红外多光谱公路工程污染监测平台。实验结果显示，改进增强深度残差超分网络模型能实现对红外多光谱图像的超分重建，其结构相似性区间及峰值信噪比区间分别为0.92～0.97和0.95和38.3～46.6,均优于其他算法。此外，以某地为例，研究通过上述勘测方法对环境污染情况进行了监测，结果显示该区域土壤Cr的含量基本均低于90 mg/kg,与实际检测结果一致。上述结果表明，研究提出的公路工程勘测平台能对工程区域进行高效勘测，并对水源保护区及污染区域准确区分。更多还原

【Abstract】 To achieve green, low-carbon and environmentally friendly highway engineering survey, a research proposes an infrared multispectral highway engineering pollution monitoring platform based on an improved enhanced deep residual super-resolution network model. The experimental results show that the improved enhanced depth residual super-resolution network model can achieve super-resolution reconstruction of infrared multispectral images. Its structural similarity interval and peak signal-to-noise ratio interval are 0.92-0.97 and 0.95 and 38.3-46.6, respectively, which are superior to other algorithms. Taking a certain area as an example, the environmental pollution situation was monitored through the above-mentioned survey method, and the results showed that the Cr content in the soil of the area was basically below 90 mg/kg, which is consistent with the actual detection results. The above results indicate that the proposed highway engineering survey platform can efficiently survey the engineering area and accurately distinguish between water source protection areas and polluted areas.更多还原