项目名称: 基于压缩感知的CMOS 图像传感器关键技术研究
项目编号: No.61306070
项目类型: 青年科学基金项目
立项/批准年度: 2014
项目学科: 无线电电子学、电信技术
项目作者: 罗韬
作者单位: 天津大学
项目金额: 25万元
中文摘要: 将压缩感知技术引入CMOS图像传感器中,可以在图像光电转换之后,直接对信号进行压缩采样,将信号采集和压缩同时进行,可以有效减少输入数据量,提高转换速度并降低系统功耗。利用CMOS 工艺特征尺寸缩小等新技术工艺,结合现代信息处理和压缩理论,可以在传感器像素阵列级实现冗余信息的集成压缩。本项目基于压缩感知理论,研究在图像传感过程中消减冗余数据的CIS设计理论,研究像素级数字降噪和频域冗余数据压缩的方法,借鉴存储阵列冗余纠错编码等模式提高像素存储稳定度和准确度,结合运动估计进行帧差时域冗余信息消除。最终建立一套基于压缩感知的新型CMOS 图像传感器理论和方法模型,完成核心像素图像信息压缩处理电路的建模、分析和优化,在电路精度、功耗和面积占用实现优化,实现CIS高性能传感成像和高效率数据输出处理。研究成果为小尺寸工艺下新一代高智能化CMOS 图像传感器芯片设计提供可行性理论指导和技术方法,并广泛应用于图像信息采集系统。
中文关键词: 压缩感知;图像传感器;图像融合;图像配准;图像编码
英文摘要: CMOS image sensor (CIS) based on compressed sensing can acquire and compress the signal at the same time,so the data from the source can be reduced. This can improve the speed of A/D conversion and cut down power consumption. Based on CMOS feature sizes shrink and modern information compression theory, we can achieve the integration of redundant information compression. Based on compressed sensing theory , This project research how to reduce redundant data in the process of image sensor of the CIS, how to cut redundant data during frequency domain and pixel digital noise reduction, how to improve pixel storage stability and accuracy from the storage array redundancy error correction coding mode,how to combined with motion estimation eliminate redundant information during time domain. Finally based on compressed sensing theory and CIS model, This project complete key pixel device and circuit modeling, design, analysis and optimization,the circuit implementation technologies that meet the needs of high precision, low power consumption, low area occupied,implement CIS high-performance imaging output and efficient data processing. The research provide guidance and technical method for the new intelligent CMOS image sensor chip design, and is widely applied in image information acquisition system.
英文关键词: compressed sensing;CIS;image fusion;image registration;image coding