基于人脸区域亮度加权的图像调优策略研究  

Research of image optimization strategy based on face region brightness weighting

作　　者：张顺军 王红涛 Zhang Shunjun;Wang Hongtao(iTruckTec(Shanghai)Co,Ltd,Shenzhen 518000,China)

机构地区：[1]智博汽车科技(上海)有限公司,广东深圳518000

出　　处：《汽车知识》2023年第12期110-112,共3页Auto Know

摘　　要：人脸图像识别是驾驶员疲劳检测的重要方法,其相关技术的研究与应用成为智能驾驶领域的热点方向。为满足不同工况下人脸图像采集,采用红外LED补光的摄像头作为采集传感器,每顿图像采集都打开LED补光对图像曝光处理,曝光考虑LED控制,这两种情况都存在图像质量难以达到特征需求、LED发热严重、功耗增加的问题。文章提出一种基于人脸区域亮度加权的图像调优方法,将LED的控制和人脸曝光统一起来考虑,提供一种全新的补光控制策略和自适应增益调整的方法,提高了采集到的人脸图像质量,同时显著降低LED的功耗。Facial image recognition is an important of method for driver fatigue detection,and the research and application of its related technologies have become a hot topic in the field of intelligent driving.To meet the requirements of facial image acquisition under different working conditions,the camera with infrared LED fll light is used as the acquisition sensor.Each frame of image acquisition is processed by turning on LED fll light for image exposure,and regardless of how the exposure is controlled,only LED control is considered.Both of these situations have problems such as difficult image quality to meet the feature requirements,severe LED heating,and increased power consumption.This article proposes an image tuning method based on facial region brightness weighting,which unifies LED control and facial exposure.It provides a new fll control strategy and adaptive gain adjustment method,improving the quality of collected facial images while significantly reducing LED powerconsumption.

关 键 词：人脸图像 区域亮度加权 图像调优 功耗 

分 类 号：U461[机械工程—车辆工程]