With the rapid growth of grids and other power supply facilities, transmission line monitoring system photography has become increasingly important in modern power systems. Transmission line monitoring systems can provide real-time equipment status information, helping to prevent faults and improve the safety of the power system. However, transmission line monitoring images captured in low-light environments are often subject to noise interference, leading to a decline in image quality, which in turn affects functions such as foreign object detection and target tracking. This paper compares and investigates some selected image denoising algorithms in transmission line monitoring images, with a particular focus on comparing the Non-Local Means Filtering (NLM) algorithm with traditional mean filtering and bilateral filtering algorithms. Since these algorithms do not require GPU usage, they can run more efficiently on power-constrained edge devices. To evaluate the denoising effects of these algorithms, we designed and conducted comparative experiments. The experimental results show that, compared to mean filtering and bilateral filtering, NLM effectively reduces noise while better preserving image details. The results indicate that participants recognized the denoising effect of NLM using Quality of Experience (QoE) as the evaluation metric. Thus, NLM is more suitable for transmission line monitoring system applications in low-light environments compared to mean filtering and bilateral filtering.