基于红外序列图像的特定目标智能检测

红外成像探测技术是目前以及未来军事作战需要倚仗的重要技术。以卫星红外序列遥感图像为处理对象，采用人工智能相关技术对图像数据进行目标检测。针对序列图像检测到的特定目标进行关联分析，并通过目标灰度提取以及运动变化信息来生成特定目标的轨迹报告。实验结果表明，多目标检测的平均正确率为90%,部分场景可以达到95%,误检率为7%;多目标跟踪的平均正确率为91%,误检率为6%,为及时发现和预警特定目标提供了信息支撑。

Intelligent Detection of Specific Target Based on Infrared Sequence Images

Infrared imaging detection technology is an important technology that military operations need to rely on at present and in the future. In this paper, satellite infrared remote sensing image is taken as the processing object, and artificial intelligence related technology is used to detect the target of image data. Based on the correlation analysis of the specific target detected in the sequence image, the trajectory report of the specific target is generated according to the extracted gray scale and motion change information of the target. The experimental results show that the average accuracy rate of multi-target detection is 90%, which can reach 95% in some scenes, and the error detection rate is 7%. The average accuracy of multi-target tracking is 91%, and the error detection rate is 6%. They provide information support for timely detection and early warning of specific targets.