Stable Light Level Detection in the Light Pools Using Neural Network Algorithms

One of the requirements in machine vision applications is that the light source and the environment are selected correctly. If the analyzed object is not properly illuminated, it is inadequate for obtaining the desired result using the captured images. A correct lighting luminaire allows to distinguish the examined object from the environment it interacts with and ensures that its attributes are obtained correctly. In this study an intelligent dynamic lighting system for illumination of a closed area has been developed. The tested luminaire has three arms. Accelerometer readings are used for positioning and orientation. Using the developed system, the positions of the armature can be estimated by bipolar sigmoid technique, according to the point the user wants to illuminate. The system measures the light level with 72 LDR, calibrated with a luxmetre.