Radar-Based Dynamic Pixel Activation in Camera Sensors for Optimized Image Processing, Power Efficiency, and Thermal Management

This paper presents a radar-based dynamic imager pixel activation method designed to optimize image processing efficiency, power consumption, and thermal management in automotive In-cabin monitoring systems (DMS and OMS). By integrating radar sensing with high-resolution cameras, the proposed approach selectively activates image sensor pixels only within radar-identified regions of interest, such as moving occupants or key cabin areas. This selective pixel activation substantially reduces the computational workload and data bandwidth required, as the camera transmits high-resolution data solely for relevant zones while maintaining lower resolution or deactivating pixels elsewhere. The fusion of radar and camera data enables adaptive sensor control that caters to distinct scenarios, including driver gaze tracking and passenger monitoring, thereby enhancing DMS and OMS functionality. The architecture supports mounting configurations such as headliner or rearview mirror installations, leveraging the co-located radar-camera field of view for effective sensing. Furthermore, the technique addresses thermal challenges inherent in continuous full-frame high-resolution imaging by significantly reducing active pixel areas, thus contributing to lower heat generation and improved system reliability. Preliminary evaluations demonstrate that this dynamic activation strategy can decrease image processing demands by good percentage without compromising situational awareness. Additionally, the concept can be extended to front-facing exterior radar and camera combinations (ADAS system). Overall, this work advances sensor fusion approaches and intelligent camera designs, delivering a scalable solution that meets the stringent performance, power, and thermal demands of next-generation automotive safety and monitoring applications.