针对复杂背景下弱小目标检测的难题,提出一种基于DSP的自适应背景预测弱小目标检测新方法。该方法在DSP为核心的嵌入式图像处理系统平台上,以自适应背景预测算法为基础,在DSP集成开发软件Code Composer Studio 3.3上采用C语言编写弱小...针对复杂背景下弱小目标检测的难题,提出一种基于DSP的自适应背景预测弱小目标检测新方法。该方法在DSP为核心的嵌入式图像处理系统平台上,以自适应背景预测算法为基础,在DSP集成开发软件Code Composer Studio 3.3上采用C语言编写弱小目标检测程序。根据图像的相邻像素的灰度特性选取不同的背景预测模型对连续四帧原始图像进行自适应背景预测得到背景预测图像,背景预测图像与原始图像相减得到残差图像;对残差图像采用交叉差分算法和自适应阈值分割处理得到二值图像;对二值图像采用逻辑与运算和形态学开运算,获得真实弱小目标。实验结果表明,该方法可以有效地检测到弱小目标,且与中值滤波算法相比,该算法预处理时间减少22%,虚警概率降低6%,检测到的目标面积增大2.3倍,更有利于目标点的观察,为工业现场镁合金熔液中弱小目标实时检测奠定了基础。展开更多
针对传统的DSP代码开发周期较长、效率低等缺点,详细介绍了一种快速、高效的DSP代码开发流程。利用MAT-LAB、S imu link、Real-Tim e W orkshop、DSP B locksets工具箱和TI公司的开发工具CCS IDE,在S imu link环境下,用图形化的方式设计...针对传统的DSP代码开发周期较长、效率低等缺点,详细介绍了一种快速、高效的DSP代码开发流程。利用MAT-LAB、S imu link、Real-Tim e W orkshop、DSP B locksets工具箱和TI公司的开发工具CCS IDE,在S imu link环境下,用图形化的方式设计DSP程序,实现代码的自动生成。完成了DSP系统设计到实施的无缝集成,缩短了开发周期。利用该方法所设计的直流力矩电机控制系统已在TMS320F2812开发板上顺利运行。展开更多
Four-point bending tests were conducted up to failure on eleven reinforced concrete (RC) beams and strengthening beams to study the effectiveness of externally pouring ultra high toughness cementitious composites (UHT...Four-point bending tests were conducted up to failure on eleven reinforced concrete (RC) beams and strengthening beams to study the effectiveness of externally pouring ultra high toughness cementitious composites (UHTCC) on improving the flexural behavior of existing RC beams.The strengthening materials included UHTCC and high strength grade concrete.The parameters,such as thickness and length of strengthening layer and reinforcement in post-poured layer,were analyzed.The flexural behavior,failure mode and crack propagation of composite beams were investigated.The test results show that the strengthening layer improves the cracking and ultimate load by increasing the cross section area.Introducing UHTCC material into strengthening not only improves the bearing capacity of the original specimens,but also disperses larger cracks in upper concrete into multiple tightly-spaced fine cracks,thus prolonging the appearance of harm surface cracks and increasing the durability of existing structures.Compared with post-poured concrete,UHTCC is more suitable for working together with reinforcement.The load?deflection plots obtained from three-dimensional finite-element model (FEM) analyses are compared with those obtained from the experimental results,and show close correlation.展开更多
针对工业现场中复杂背景下弱小目标检测的难题,提出了一种基于DSP的双层流水线的弱小目标检测新方法。该方法以双层流水线算法为背景,以半因果支持域背景预测为辅助,以DSP硬件图像处理系统平台为核心。利用C语言编写检测弱小目标的双层...针对工业现场中复杂背景下弱小目标检测的难题,提出了一种基于DSP的双层流水线的弱小目标检测新方法。该方法以双层流水线算法为背景,以半因果支持域背景预测为辅助,以DSP硬件图像处理系统平台为核心。利用C语言编写检测弱小目标的双层流水线算法,通过DSP集成开发软件Code Composer Studio 3.3进行移植、编译、调试及优化,将调试无误的算法程序烧写到DSP中,应用于高温镁熔液中弱小目标的检测。实验结果表明基于DSP的双层流水线方法可以有效检测出弱小目标,为实现工业现场弱小目标的实时检测奠定了基础。展开更多
文摘针对复杂背景下弱小目标检测的难题,提出一种基于DSP的自适应背景预测弱小目标检测新方法。该方法在DSP为核心的嵌入式图像处理系统平台上,以自适应背景预测算法为基础,在DSP集成开发软件Code Composer Studio 3.3上采用C语言编写弱小目标检测程序。根据图像的相邻像素的灰度特性选取不同的背景预测模型对连续四帧原始图像进行自适应背景预测得到背景预测图像,背景预测图像与原始图像相减得到残差图像;对残差图像采用交叉差分算法和自适应阈值分割处理得到二值图像;对二值图像采用逻辑与运算和形态学开运算,获得真实弱小目标。实验结果表明,该方法可以有效地检测到弱小目标,且与中值滤波算法相比,该算法预处理时间减少22%,虚警概率降低6%,检测到的目标面积增大2.3倍,更有利于目标点的观察,为工业现场镁合金熔液中弱小目标实时检测奠定了基础。
文摘针对传统的DSP代码开发周期较长、效率低等缺点,详细介绍了一种快速、高效的DSP代码开发流程。利用MAT-LAB、S imu link、Real-Tim e W orkshop、DSP B locksets工具箱和TI公司的开发工具CCS IDE,在S imu link环境下,用图形化的方式设计DSP程序,实现代码的自动生成。完成了DSP系统设计到实施的无缝集成,缩短了开发周期。利用该方法所设计的直流力矩电机控制系统已在TMS320F2812开发板上顺利运行。
基金Project(50438010) supported by the National Natural Science Foundation of China
文摘Four-point bending tests were conducted up to failure on eleven reinforced concrete (RC) beams and strengthening beams to study the effectiveness of externally pouring ultra high toughness cementitious composites (UHTCC) on improving the flexural behavior of existing RC beams.The strengthening materials included UHTCC and high strength grade concrete.The parameters,such as thickness and length of strengthening layer and reinforcement in post-poured layer,were analyzed.The flexural behavior,failure mode and crack propagation of composite beams were investigated.The test results show that the strengthening layer improves the cracking and ultimate load by increasing the cross section area.Introducing UHTCC material into strengthening not only improves the bearing capacity of the original specimens,but also disperses larger cracks in upper concrete into multiple tightly-spaced fine cracks,thus prolonging the appearance of harm surface cracks and increasing the durability of existing structures.Compared with post-poured concrete,UHTCC is more suitable for working together with reinforcement.The load?deflection plots obtained from three-dimensional finite-element model (FEM) analyses are compared with those obtained from the experimental results,and show close correlation.
文摘针对工业现场中复杂背景下弱小目标检测的难题,提出了一种基于DSP的双层流水线的弱小目标检测新方法。该方法以双层流水线算法为背景,以半因果支持域背景预测为辅助,以DSP硬件图像处理系统平台为核心。利用C语言编写检测弱小目标的双层流水线算法,通过DSP集成开发软件Code Composer Studio 3.3进行移植、编译、调试及优化,将调试无误的算法程序烧写到DSP中,应用于高温镁熔液中弱小目标的检测。实验结果表明基于DSP的双层流水线方法可以有效检测出弱小目标,为实现工业现场弱小目标的实时检测奠定了基础。
文摘针对工业现场复杂背景下镁熔液中弱小目标的检测问题,提出了基于DM642的镁熔液中弱小目标检测算法优化方法,旨在实现检测算法运行的高效性和运算速度的高速性。方法以高性能数字媒体处理器DM642为核心,以DSP集成开发软件Code Composer Studio 3.3为辅助,以工业现场复杂镁熔液中的弱小目标检测为载体,利用C编译器,主要从预处理算法和软件优化两方面对镁熔液中弱小目标的检测算法进行优化研究。结果表明,经过优化的检测算法在速度和稳定性等方面均得到了明显提高,解决了弱小目标检测算法中结构复杂和运算速度慢的问题,为实现工业现场弱小目标的实时检测奠定了基础。
