基于特征向量提取的激光遥感图像模式识别系统设计
|
摘要
针对当前激光遥感图像识别系统在设计上采用单一固定特征提取图像,存在丢失细节且易出现过识别、整体对比度低、细节模糊等问题,设计了特征向量提取的激光遥感图像模式识别系统。硬件上选用ADSBF535作为图像识别模块的核心,采用XC2S200 FPGA作为图像输入模块的硬件支撑,输出像是模块配置单排排针连接P2对应的IO引脚,更便于DSP和LCD的连接。软件根据模糊熵理论结合特征提取结果计算图像提取区间灰度均值,按照灰度值进行标号对灰度直方图进行均衡,依据均衡后的灰度直方图对区间内的灰度采用均衡系数平移,实现区间内灰度细节的填充,完成激光遥感图像模式识别。实验结果表明,所设计系统相比当前激光遥感图像识别系统,有效提高了图像对比度和分辨率,识别了图像局部细节,显著改善了激光遥感图像模式识别效果。
Aiming at the current laser remote sensing image recognition system,a single fixed feature is used to extract the image. There are some problems such as missing details and easy recognition,low overall contrast and blurred details. The laser remote sensing image pattern recognition system based on feature vector extraction is designed. The hardware uses ADS-BF535 as the core of the image recognition module. The XC2 S200 FPGA is used as the hardware support of the image input module. The output is like the module configuration. The single-row pin connection connects the IO pin corresponding to P2,which is more convenient for the connection between DSP and LCD.The software calculates the gray mean value of the image extraction interval according to the feature extraction result,and performs the equalization of the gray histogram according to the gray value. According to the equalized gray histogram,the equalization coefficient is used for the gray level in the interval to realize the gray in the interval. The filling of the detail details completes the pattern recognition of the laser remote sensing image. The experimental results show that compared with the current laser remote sensing image recognition system,the designed system effectively improves the image contrast and resolution,identifies the local details of the image,and significantly improves the pattern recognition effect of the laser remote sensing image.
Aiming at the current laser remote sensing image recognition system,a single fixed feature is used to extract the image. There are some problems such as missing details and easy recognition,low overall contrast and blurred details. The laser remote sensing image pattern recognition system based on feature vector extraction is designed. The hardware uses ADS-BF535 as the core of the image recognition module. The XC2 S200 FPGA is used as the hardware support of the image input module. The output is like the module configuration. The single-row pin connection connects the IO pin corresponding to P2,which is more convenient for the connection between DSP and LCD.The software calculates the gray mean value of the image extraction interval according to the feature extraction result,and performs the equalization of the gray histogram according to the gray value. According to the equalized gray histogram,the equalization coefficient is used for the gray level in the interval to realize the gray in the interval. The filling of the detail details completes the pattern recognition of the laser remote sensing image. The experimental results show that compared with the current laser remote sensing image recognition system,the designed system effectively improves the image contrast and resolution,identifies the local details of the image,and significantly improves the pattern recognition effect of the laser remote sensing image.
引文
[1]刘雪鸥.医学图像模式识别技术的研究及应用[D].山西:太原理工大学,2016.
[2]高丽燕.基于DSP+FPGA的图像识别系统设计与实现[D].南京:南京理工大学,2006.
[3]陈飞.基于GAPSO-BP算法的高阶板形在线模式识别及其检测系统[D].秦皇岛:燕山大学,2016.
[4]刘子维.地震前兆数据异常识别关键技术研究[D].武汉:武汉大学,2016.
[5]张杰峰.基于智能手机的出行方式识别系统的设计和实现[D].广州:华南理工大学,2016.
[6]郑越.基于图像处理和模式识别的图像监控系统初探[J].中国新通信,2018,16:232-235.
[7]段彦卉,韩紫婷,沈天宇,等.基于K60的光电追逐系统的设计与实现[J].汽车实用技术,2018,16:114-117.
[8]杨晓林,张剑平,吴建新,等.基于以太网技术的铁路计量系统设计研究[J].铁道货运,2018,08:42-47.
[9]方浩羽,左仲善,李岩,等.基于单片机的电动自行车充电系统的设计---以河北水利电力学院为例[J].科技与创新,2018,17:57-59.
[10]林青有.基于景区视频的异常事件检测系统的设计与实现[J].价值工程,2018,37(28):228-230.
[11]林卉,于瑞鹏,王李娟,等.基于特征向量的遥感影像自动分类研究[J].计算机工程与应用,2017,53(16):177-181.
[12]魏英娟,郑雄伟,雷兵,等.基于样本模型的高分一号遥感影像云雾自动提取[J].国土资源遥感,2017,29(10):39-45.
[13]李亚飞,董红斌.基于卷积神经网络的遥感图像分类研究[J].智能系统学报,2018,13(4):550-556.
[14]卢学佳,安博文,陈元林,等.基于现代谱估计的光纤周界安防系统模式识别[J].光通信技术,2017,41(5):19-21.
[15]陈洋,范荣双,王竞雪,等.结合相位一致和分水岭变换的高分辨率遥感影像分割方法[J].激光与光电子学进展,2017,54(9):375-380.
[16]刘燕萍,程效军,贾东峰,等.一种序列影像和地面激光点云的配准方法[J].激光杂志,2017,38(10):90-95.
[17]MING D,QIU Y,WEN Z.Applying Spatial Statistics into Remote Sensing Pattern Recognition:with Case Study of Cropland Extraction Based on Ge OBIA[J].Acta Geodaetica Et Cartographica Sinica,2016,45(7):825-833.
[18]CAO C,FAN X,LIU Q,et al.Practical Pattern Recognition System for Distributed Optical Fiber Intrusion Monitoring Based onФ-COTDR[J].ZTE Communications,2017,15(3):52-55.
[2]高丽燕.基于DSP+FPGA的图像识别系统设计与实现[D].南京:南京理工大学,2006.
[3]陈飞.基于GAPSO-BP算法的高阶板形在线模式识别及其检测系统[D].秦皇岛:燕山大学,2016.
[4]刘子维.地震前兆数据异常识别关键技术研究[D].武汉:武汉大学,2016.
[5]张杰峰.基于智能手机的出行方式识别系统的设计和实现[D].广州:华南理工大学,2016.
[6]郑越.基于图像处理和模式识别的图像监控系统初探[J].中国新通信,2018,16:232-235.
[7]段彦卉,韩紫婷,沈天宇,等.基于K60的光电追逐系统的设计与实现[J].汽车实用技术,2018,16:114-117.
[8]杨晓林,张剑平,吴建新,等.基于以太网技术的铁路计量系统设计研究[J].铁道货运,2018,08:42-47.
[9]方浩羽,左仲善,李岩,等.基于单片机的电动自行车充电系统的设计---以河北水利电力学院为例[J].科技与创新,2018,17:57-59.
[10]林青有.基于景区视频的异常事件检测系统的设计与实现[J].价值工程,2018,37(28):228-230.
[11]林卉,于瑞鹏,王李娟,等.基于特征向量的遥感影像自动分类研究[J].计算机工程与应用,2017,53(16):177-181.
[12]魏英娟,郑雄伟,雷兵,等.基于样本模型的高分一号遥感影像云雾自动提取[J].国土资源遥感,2017,29(10):39-45.
[13]李亚飞,董红斌.基于卷积神经网络的遥感图像分类研究[J].智能系统学报,2018,13(4):550-556.
[14]卢学佳,安博文,陈元林,等.基于现代谱估计的光纤周界安防系统模式识别[J].光通信技术,2017,41(5):19-21.
[15]陈洋,范荣双,王竞雪,等.结合相位一致和分水岭变换的高分辨率遥感影像分割方法[J].激光与光电子学进展,2017,54(9):375-380.
[16]刘燕萍,程效军,贾东峰,等.一种序列影像和地面激光点云的配准方法[J].激光杂志,2017,38(10):90-95.
[17]MING D,QIU Y,WEN Z.Applying Spatial Statistics into Remote Sensing Pattern Recognition:with Case Study of Cropland Extraction Based on Ge OBIA[J].Acta Geodaetica Et Cartographica Sinica,2016,45(7):825-833.
[18]CAO C,FAN X,LIU Q,et al.Practical Pattern Recognition System for Distributed Optical Fiber Intrusion Monitoring Based onФ-COTDR[J].ZTE Communications,2017,15(3):52-55.