Low-Cost Implementation of Adaptive Luminescence Intensity in Headlamp Based on Scenario Extraction

In this work, we try to achieve a low-cost and easy implementation of an adaptive headlamp. We propose this simple system implementation to avoid accidents caused by improper lighting and exposure to glare from passing vehicles. In this work, we have used a classification network trained using a Google teachable machine to classify the parts of the day and the luminous intensity during driving time based on the live capturing of san image. Further using this classification, the required luminous intensity for comfortable driving is decided. The pattern required in the headlamp to provide the same is decided and controlled by the proposed system. The work has also been compared with the results using the averaging pixels method for each frame for classification. Both approaches are implemented on Raspberry Pi4 model B+, Raspberry Pi Cam, and Raspberry Pi sense hat to visualize the results and their effectiveness.