基于DSP的自适应弱小目标检测方法  

ADAPTIVE DIM AND WEAK TARGET DETECTION METHOD BASED ON DSP

作　　者：李天甲[1] 许四祥[1] 姚志生[1] 王洋[1] 

机构地区：[1]安徽工业大学机械工程学院,安徽马鞍山243002

出　　处：《计算机应用与软件》2018年第1期243-245,317,共4页Computer Applications and Software

基　　金：国家自然科学基金项目(51374007);安徽省自然科学基金项目(11040606M104);研究生创新基金项目(2016063)

摘　　要：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。Aiming at the problem of dim and weak target detection in complex background, a new method of adaptive background prediction for dim and weak detection based on DSP is proposed. The method is centered on DSP of the embedded image processing system platform and based on self-adaptive background prediction. The algorithm using for dim and weak target detection was programmed with C language on Code Composer Studio 3.3 which is an integrated development environment for DSP. Firstly, according to the gray characteristics of adjacent pixels, the adaptive background prediction model is selected to get the background image and the residual images were obtained by subtracting the prediction image from the original image. Secondly, the two value image is obtained by cross difference algorithm and adaptive threshold segmentation. Finally, the real dim and weak target was detected combining with the logical and algorithm and open operation of mathematical morphology. Experimental results show that this method can effectively detect the dim and weak target, and the preprocessing time of this method is 22% less than of median filtering algorithm, the probability of fault decreased by 6%, the target point size increased by 2.3 times. It is also laying the foundation of real-time detection of dim and weak target on industrial scene.

关 键 词：DSP 自适应背景 弱小目标Code COMPOSER STUDIO 3.3 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]