摘要
由于红外激光成像中的弱小目标,存在信噪比低、无纹理信息和信号弱的特征,使得传统基于背景抑制的目标检测方法,无法将目标从复杂的背景中分割出来,检测准确率低。因此,将基于尺度空间的红外弱小多目标检测算法应用于红外激光成像中弱小目标的定位中,给出了红外激光成像中弱小目标定位硬件总体结构,其由传输控制模块、图像采集模块和弱小目标定位模块组成。采用传输控制模块TMS320F2812传递和处理系统中的图像,依据中央控制单元的命令对红外摄像机进行控制,通过核心芯片为XC95288XL的图像采集模块获取弱小目标图像,依据一定的字节顺序保存在RAM中。通过软件设计中的基于尺度空间的红外弱小多目标检测算法,完成弱小目标的准确定位。实验结果表明,该算法可在不同背景环境中,有效过滤干扰区域,准确检测出实际弱小目标,鲁棒性较高。
Because the dim-small targets in infrared laser imaging has low signal-to-noise ratio and no texture information, and the signal is weak, which makes the traditional target detection method based on background suppression method unable to segment a target from its complex background, the infrared small target detection algorithm based on the scale space is used in small target positioning in the infrared laser imaging, which can rapidly and accurately detect multiple weak small targets in complex background. The overall structure of the hardware for dim-small target positioning in the infrared laser imaging is given. The system is composed of transmission control module, image acquisition module and small target positioning module. The transmission control module TMS320F2812 is adopted to transmit and process the images in the system and control the infrared camera according to the instruction from the central control unit. The dim-small target image is got by the image acquisition module of the core chip XC95288XL and restored in RAM according to a certain byte order. The infrared dim-small multi-target detection algorithm based on the scale space in software design is used to fulfill accurate positioning of dim-small target. The ex- periment results indicate that the algorithm can effectively filter the interference area in the different background environments, accurately detect dim-small targets, and has high robustness.
出处
《现代电子技术》
北大核心
2016年第20期83-86,91,共5页
Modern Electronics Technique
基金
国家自然科学基金(15369627)
