Method for Illumination Target Recognition Based on Depth Information
The issue of energy waste in lighting has been increasingly receiving attention,and image sensing methods have been gradually applied in the field of lighting control.However,lighting sensing methods based on monocular images were susceptible to interference from factors such as lighting and texture.To address these issues,this paper proposed a method that combined depth information for background separation,extracted potential lighting targets in the scene,and utilized image moments as descriptors for target recognition.A binocular experimental platform was constructed for indoor scene testing.The results showed that this method could effectively separate,recognize,and lo-cate lighting targets in the scene.Compared to monocular background separation methods,the separation contours ob-tained were more complete and less sensitive to grayscale differences between the background and targets.Therefore,this method had certain application value for large-area and high-precision lighting control scenarios.
lighting controlbinocular visionbackground separationimage moment
NETL
NSTL
万方数据
