The solar corona is the outermost layer of the solar atmosphere, consisting of thin, highly ionized and hot plasma. In visible light, the solar corona is much dimmer than the photosphere, only one millionth as bright, so it is not visible in daylight. The only exception is during a total solar eclipse, when the moon blocks the light from the photosphere completely and reveals the corona. The coronagraph is a special telescope that can observe the solar corona even in the absence of a total solar eclipse. It creates an artificial eclipse by using an occulter to block the bright photosphere. Depending on the location of the occulter, there are two main types of coronagraphs: externally occulted and internally occulted. The externally occulted coronagraph blocks the direct sunlight with an occulter in front of the objective lens, but it also causes some vignetting in the field of view, which affects the quality of the observations. The internally occulted coronagraph has the occulter behind the objective lens, however, this means that some parts of the telescope, such as the objective lens and the tube, are exposed to direct sunlight and produce a lot of stray light, which needs to be strictly suppressed.We performed an experiment to explore how the amount of dust on the objective surface affects the intensity of the scattering background in the coronal image plane. We used the Lijiang 10 cm Lyot Coronagraph (Yunnan Observatories Green-line Imaging System, YOGIS), an internally occulted coronagraph that operates at a wavelength of 530.3 nm and has a main focal length of 1490 millimeters. YOGIS is a joint project between the Yunnan Observatories and the National Astronomical Observatory of Japan, and it is the only ground-based coronagraph in China that can perform regular observations, which has a total length of nearly 3 meters and a weight of 100 kilograms, and uses a CMOS camera with 2 048×2 048. The experiment was conducted at Lijiang Observatory (E100°01′4″, N26°41′42″), which is located at an altitude of 3 200 meters and has low atmospheric scattering, making it ideal for ground-based solar corona observation. Two experimenters and one observation assistant participated in the experiment on November 17, 2022, when the sky was clear and cloudless.The main goal of the experiment is to obtain dust information and measure its impact on the coronal image, called scattering background.To obtain dust data, we need the conjugate imaging of the objective lens. We achieve this by pointing the coronagraph at the sun and placing the camera at lens' focal plane. The resulting image has three parts. Firstly, light scattered by surface microroughness, which remains constant regardless of the cleanness level. Secondly, light scattered by dust particles on the objective surface, which varies depending on how clean it is. Finally, the light scattered by the earth's atmospheric is scattered by surface microroughness and dust particles, which is negligible compared to the second part. We use Otsu's method to process the image into binary pixels that show each particles' area, which tell us about its area and intensity. Assuming that the dust is distributed evenly on objective surface, we use total intensity (I) as a measure for dust information.To get the scattering background, we take two images of corona: one before and the other after cleaning the objective surface, and then align and subtract them. To account for the change in solar brightness, we normalize the two images by dividing them by their respective solar radiation intensities before subtraction.Based on the above two aspects, we conducted the experiments with the following steps. Firstly, we imaged lens and corona with significant dust on objective surface. Then we cleaned it and imaged again. This gave us scattering background and its related dust scattering points' intensity (I). Then, we converted this image into polar coordinates and took median intensities for each radius. This showed us that the scattering background decreased linearly with distance. We also saw that the slope and intercept of this linear equation depended linearly on dust intensity (I). Hence, we derived the function of the scattering background with respect to the distance and the total intensity of the scattering points (I).We can use these relationships to create different scattering background based on different intensities (I), and then subtract them from the corresponding coronal images to get the corrected ones. Our model is validated by comparing the intensity decay curves of the corrected coronal images with those of the clean images, which are similar. Our method not only effectively improves the quality of coronal data, but also makes faint coronal structure more prominent and clearer, which helps us to obtain more accurate results when analyzing the intensity decay trend of the inner corona and coronal structure changes. What's more, it will help us to further understand the effects of other stray light sources inside the coronagraph, helping the further technological development of large-aperture coronagraphs.The stray light of the internally occulted coronagraph can be divided into two parts: fixed and variable. The former does not change with time and environmental cleanliness, while the latter increases with decreasing environmental cleanliness. For the ground-based internally occulted coronagraph with regular observations, the scattering from dust on the objective surface is the variable stray light, as the dust continues to accumulate on the surface. Although the accumulation can be slowed down by frequent cleaning of the objective surface and the level of this stray light can be suppressed, the objective surface cannot be guaranteed to remain clean throughout the observation due to weather, atmospheric particle pollution and other factors. More importantly, this approach is somewhat subjective, as the observer cannot directly assess the current level of mirror cleanliness, resulting in an untimely cleaning and multiple disassembly of the objective lens, creating secondary contamination on the objective surface to some extent. So, the stray light levels are changing over time, leading to different amounts of scattered background in the coronal data, which brings great inconvenience in the analysis of the faint coronal structure and the coronal intensity calibration.