Clear boundaries, exact shapes
All moving objects cast shadows, and while being under the source of light the shadow has small geometrical dimensions, with lateral lighting the size of the shadow may exceed the size of the object itself.
When the object moves, so does the shadow, thus increasing the motion area. If the video analytics module uses moving object identification, the shadow just increases the motion area and impedes object search in the motion zone (due to increased motion area), thus not affecting the accuracy of identification.
However, if the moving object is not identified by the video analytics module (for example, people count by the motion area size), it causes a great problem.
In such cases, a great solution is the shadow removal module, which allows unambiguously defining where the object’s boundaries end and its shadow begins.