CAJ | 학술논문

大孔径静态干涉成像光谱仪在实际应用中因为器件问题产生大量坏像元，其特殊的时空联合调制成像方式导致传统的光谱图像域坏像元修正方法无法使用。而设备在轨运行期间坏像元的不断恶化、复杂背景地物的干扰更是为坏像元的自动检测修正处理带来巨大挑战。基于对坏像元分布和变化特征的长时间分析研究，文章提出一种在原始帧的差分域而不是光谱影像上进行坏点自动检测的新方法，很好地克服了现有方法中坏点检测效率较低、阈值计算精度低且所需时间长等缺陷，通过实际数据验证表明，坏像元修正后的复原光谱影像坏线得到了很好的消除，且与理想光谱的相对误差保持在0.2%以内。
대공경정태간섭성상광보의재실제응용중인위기건문제산생대량배상원，기특수적시공연합조제성상방식도치전통적광보도상역배상원수정방법무법사용。이설비재궤운행기간배상원적불단악화、복잡배경지물적간우경시위배상원적자동검측수정처리대래거대도전。기우대배상원분포화변화특정적장시간분석연구，문장제출일충재원시정적차분역이불시광보영상상진행배점자동검측적신방법，흔호지극복료현유방법중배점검측효솔교저、역치계산정도저차소수시간장등결함，통과실제수거험증표명，배상원수정후적복원광보영상배선득도료흔호적소제，차여이상광보적상대오차보지재0.2%이내。
Large aperture static imaging spectrometer (LASIS) in practical use generates a large number of bad pixels, and conventional bad pixel correction method cannot be used directly because of its specific space-time modulation imaging mode. Moreover, the complexity of bad pixels generated and interference of background feature bring about a huge challenge for the subsequent data process. Based on the precise analysis on characteristics for bad pixel, we present a new automatic detection algorithm in the difference domain instead of the original image frame, which can overcome the disadvantages of existing algorithms, such as low detection efficiency, low threshold calculation accuracy and long-time taking. Data experiments verify that the bad lines on spectral images caused by bad pixels can be well removed after the correction by our algorithm, and the relative spectral error can be maintained less than 0.2%.