红外与激光工程
紅外與激光工程
홍외여격광공정
INFRARED AND LASER ENGINEERING
2014年
4期
1080-1086
,共7页
徐富元%顾国华%陈钱%王长江%杨蔚
徐富元%顧國華%陳錢%王長江%楊蔚
서부원%고국화%진전%왕장강%양위
转动红外探测器%远距离运动目标检测%光流法%运动信息相关性
轉動紅外探測器%遠距離運動目標檢測%光流法%運動信息相關性
전동홍외탐측기%원거리운동목표검측%광류법%운동신식상관성
rotational infrared detector%remote motion target detection%optical flow%motion information correlation
地面远距离运动目标检测是红外成像防御中的关键技术之一。地面远距离运动目标检测指利用红外探测器对远距离运动目标实现自动探测和捕获。由于目标距离较远,在红外探测器上成像面积较小,同时远距离运动目标容易被遮挡,很难将目标从复杂的地面背景中提取出来。提出了一种能够在面阵红外探测器转动的条件下实现远距离地面运动目标检测方法。首先利用图像信息计算红外探测器的运动补偿参数并通过背景更新获取目标一次检测结果,然后采用光流法获取背景和目标的运动信息,通过计算背景和目标之间的运动信息相关性实现目标最终检测。实验采用多种场景对文中算法进行验证,结果表明该方法通过转动红外探测器可以扩大目标搜索区域,利用背景和目标之间运动信息的相关性有效地克服了目标遮挡、目标重叠和视差等问题。
地麵遠距離運動目標檢測是紅外成像防禦中的關鍵技術之一。地麵遠距離運動目標檢測指利用紅外探測器對遠距離運動目標實現自動探測和捕穫。由于目標距離較遠,在紅外探測器上成像麵積較小,同時遠距離運動目標容易被遮擋,很難將目標從複雜的地麵揹景中提取齣來。提齣瞭一種能夠在麵陣紅外探測器轉動的條件下實現遠距離地麵運動目標檢測方法。首先利用圖像信息計算紅外探測器的運動補償參數併通過揹景更新穫取目標一次檢測結果,然後採用光流法穫取揹景和目標的運動信息,通過計算揹景和目標之間的運動信息相關性實現目標最終檢測。實驗採用多種場景對文中算法進行驗證,結果錶明該方法通過轉動紅外探測器可以擴大目標搜索區域,利用揹景和目標之間運動信息的相關性有效地剋服瞭目標遮擋、目標重疊和視差等問題。
지면원거리운동목표검측시홍외성상방어중적관건기술지일。지면원거리운동목표검측지이용홍외탐측기대원거리운동목표실현자동탐측화포획。유우목표거리교원,재홍외탐측기상성상면적교소,동시원거리운동목표용역피차당,흔난장목표종복잡적지면배경중제취출래。제출료일충능구재면진홍외탐측기전동적조건하실현원거리지면운동목표검측방법。수선이용도상신식계산홍외탐측기적운동보상삼수병통과배경경신획취목표일차검측결과,연후채용광류법획취배경화목표적운동신식,통과계산배경화목표지간적운동신식상관성실현목표최종검측。실험채용다충장경대문중산법진행험증,결과표명해방법통과전동홍외탐측기가이확대목표수색구역,이용배경화목표지간운동신식적상관성유효지극복료목표차당、목표중첩화시차등문제。
The ground remote motion target detection is one of key techniques in infrared imaging defense. The ground remote motion target detection is a method which can automatic detection and capture the remote motion target by Infrared detector. It is difficult to extract the target from the complex background, because the target is distance, imaged area is small on infrared detector and easily blocked. A method can realize detect the ground remote motion target by rolling the area array infrared detector was presented. First, the motion compensation parameters were calculated through the image information and target first detection result was obtained through background update, then background and target motion information were obtained through the optical flow and target final detection result was obtained through calculating the relationship between background and target motion information. Experiments validate the proposed algorithm using a variety of scenarios; the results show that this method can expand the target search area by the rotating infrared detectors. Using the relevance of the movement information between background and targets effectively overcome the problems of the target obscured, goals overlap and parallax.