In a new study, scientists have presented an automatic method to improve photos and videos in difficult weather conditions. This will help you to use computer vision more useful.
Kyiv. Ukraine. Ukraine Gate – July 22, 2021 – Technology
The researchers explained that computer vision technology is increasingly being used in automated surveillance systems, face recognition, healthcare, and more. However, they may not perform well if they use poor-quality photos or videos. Usually, these materials are spoiled by lack of light, rain, or other weather conditions.
Scientists have determined that nighttime images are affected by low light and artificial lighting effects such as glare, glow, and spotlights. Rain also reduces the visibility of cameras by several times.
“Many computer vision systems, such as automatic surveillance or autonomous vehicles, rely on clear visibility of the input video to function. For example, cars cannot work reliably in heavy rain, and automatic video surveillance systems often fail at night, especially if the scenes are dark or there is significant glare, ”the scientists noted.
Scientists have applied deep learning algorithms to improve the quality of night photos and videos taken in the rain. In the first study, they increased brightness while simultaneously suppressing noise and light effects (glare, glow) to produce crisp night images. This technique is new and solves the problem of the clarity of night images and video when there are no normal weather conditions for a normal frame.
The researchers were also able to improve the three-dimensional assessment of the poses of those people who were in the frame. By combining these two approaches, the new method allows for a more reliable estimate of posture even within crowds, as well as accurately calculating the distance between people.
Read Also: Effective Ways to Improve Clay Soils
Source: Ukrgate
