Research On Infrared Small Target Detection Algorithm Based On Human Vision Mechanism And Gradient Direction Consistency

Key technologies such as infrared detection and precision guidance have high application value in national defense construction.Therefore,the detection of small infrared targets under complex backgrounds has become the focus of research in the military field.Compared with the traditional radar detection system,the infrared search and tracking system itself has the advantages of good concealment and strong anti-interference ability,and can detect targets quickly and effectively in long-distance detection scenarios.Due to the long detection distance,the target occupies fewer pixels in the infrared image,the contrast between the target and the background is low,and there are no features such as color and texture,and it is easy to be submerged by clutter in complex scenes.These factors make the detection of small infrared targets more difficult.In response to the above problems,this paper has conducted in-depth research on target enhancement,background suppression and strong clutter suppression,and proposed an algorithm for detecting small infrared targets in complex scenes.For the local contrast algorithm,the sliding nested window is used to traverse the image to detect potential targets.If the features of the target and the background are not fully utilized,it may cause false alarms.This paper carries out weighting processing on the basis of the local contrast algorithm,and the feature of small targets can be fully utilized through the weighting function.With the background characteristics and the difference information between the target and the background,the ratio form and the difference form are combined to redefine the local contrast,which can effectively enhance the target while suppressing the background.Taking into account that the detection of small infrared targets in complex scenes is interfered by strong edges and other clutters,by analyzing the gradient characteristics of infrared images,it is concluded that the gradient direction at the small target has a high degree of consistency.Edge clutter has no similar characteristics.Based on this,this paper uses the idea of block division to block the image,and calculates the local gradient amplitude and direction.By calculating the ratio of the maximum and minimum gradient average amplitudes in each region,the strong edge background clutter can be further suppressed.The idea makes full use of the characteristics of the same gradient direction around the target.Finally,the target to be detected is obtained by threshold segmentation.Through experimental verification of the algorithm,the proposed algorithm can effectively enhance the target while suppressing background clutter,especially for strong edge clutter.It can be concluded from the related evaluation indicators that the proposed infrared small target detection algorithm can effectively reduce the false alarm rate in complex scenes and has strong robustness compared with other algorithms.