Researchers have been exploring ways to improve autonomous drone operations using onboard cameras for drones belonging to the nano/micro/small category. These drones may be mounted with only low-resolution camera(s) due to the maximum takeoff weight limitation of these drones. Decision-making using low-resolution images generated by the primary camera can be faulty unless the images are captured from a distance very close to the point of interest. Detection of the point(s) of interest as early as possible is required to ensure sufficient response time for safe maneuvering. Hence, the images are to be captured at greater heights from the point(s) of interest, and obtaining the high-resolution images from the captured low-resolution images is crucial. The presence of noise in the images adversely affects the onboard decision-making. This work highlights the limitations of the single-image super-resolution techniques to super-resolve noisy images and aims to remove noise from the images before super-resolution. We propose a deep learning-based novel architecture NICSR, i.e. Noise Identification and Compensation technique for Super-Resolution of images captured by a drone in the presence of noise. The performance analysis of the proposed technique on a publically available dataset reveals appreciable improvement in the quality of the super-resolved images, even if the input images are noisy. The proposed work will hence accelerate drone-based applications in civil and military sectors such as aerial photography, terrain mapping, surveillance, aerial inspection, drone-based delivery, and disaster management.