In single-frame infrared dim and small target detection,in view of the complex background of an image and the diversity of target morphology,an algorithm is proposed,in which background suppression of the image is carried out,then candidate points are extracted by threshold segmentation,and finally the residual noise points are removed by energy concentration.In the preprocessing part,in view of the diffusion of small targets and the dynamic distribution characteristics of the image plane,the morphology is improved and the background is suppressed by using multi-directional filtering template.In the part of threshold segmentation,a multi-directional gradient threshold segmentation method is proposed.The gradient values of eight directions with different step sizes are calculated,and the number of gradients exceeding the threshold is counted to screen out the target points.Experiments show that the proposed detection method can effectively solve the problems of diffusion and multi-morphological characteristics of dim and small targets.