Study On Automatic Interpretation Algorithm Of Infrared Dim Small Target In The Shooting Range

Over the last decade, with the rapid development of infrared imaging technology,The infrared imaging track and measure systems were widely used in the shooting range,infrared image has become an important source of data for the range test, the infraredimage interpretation model also requires changes ensued, the traditional manualinterpretation method has been unable to meet the mission requirements, automatic,high-precision, high efficiency range infrared images for automatic interpretationbecome an urgent problem to be solved. In this practical problem, research the range ofinfrared dim small target automatic detection technology and automatic interpretation ofthese techniques to work.Firstly, the infrared dim small target detection scenarios, analysis of the currentresearch situation, discusses the main advantages and disadvantages of the method ofinfrared dim small target detection.Secondly, describe the infrared moving dim small target tracking mode in theshooting range, then from the spatial, frequency and temporal to study thecharacteristics of the target range infrared dim small target image and the image in theface of the tripartite filtering process to compare the filtering effect. To be improved onthe basis of the classical morphological filtering based on the realization of themorphological band-pass filter, choose this method as a pretreatment method.Again, for the completion of the pretreatment images CFAR thresholdsegmentation, image segmentation after doing centroid extraction, given the targettracking mode is typical of people in the loop control target in the temporal withouttrace feature, no fixed speed, no fixed direction characteristics associated trace ideasrelated applications based on the data associated with the nearest neighbor correlationdetection algorithm neighborhood association for target detection applications based onideological trajectory, the algorithm can be achieved.Finally, the performance and applicability of the algorithm to be evaluated. First,apply a high signal to noise ratio of the fast moving object image detection algorithmperformance, and then apply a lower signal to noise ratio, longer sequence length slowmoving target detection algorithm performance image, the algorithm of the above twosets of image the interpretation of detection target, the detection rate was88%. But forlower-noise ratio to detect the image cannot be completed, and we make an assessmentof the applicability of the algorithm, and in the final of the improved algorithm forprospects.