- 期刊
Reliability Assessment of Gaussian Estimates Derived from Small-footprint Waveform Lidar
空載波形光達資料之高斯估值可靠度評估
摘要
目前在處理大足跡或小足跡波形光達資料，高斯分解回波技術扮演相當重要之角色，當以高斯函數擬合所接收之回波或地表反應，可獲算得各回波之高斯係數，主要包含對應為回波時間點之高斯波峰值、回波寬度與回波振幅值，這些屬性隱含雷射掃描目標物之特徵。然而在某些情況下，所估算得的參數具不確定性，目前僅少數研究評估高斯估算值之可靠度或不確定性，在應用波形高斯估值時，實必須考量此不確定性之指標值。本文主在研究此議題，調查波形複雜度對高斯估值可靠度之影響，測試資料為包含模擬波形資料與Riegl LMS-Q560真實波形資料，另掃描當天設置所設計之目標物於野外現地，觀察波形形狀隨多重回波間之距離與振幅大小之影響。模擬資料發現高斯波寬估值之均分根誤差會隨著多重回波間之距離縮短（＜6ns）而增加，回波時間距離之均方根誤差則可維持一致之精度（0.04ns）。
並列摘要
The Gaussian decomposition technique has to date played an important role in processing both large- and small-footprint lidar waveforms. When fitting Gaussian functions to received waveforms or the surface response estimated by a deconvolution process, Gaussian coefficients for each detected return can be estimated. These are the temporal position of the Gaussian peak, pulse width and amplitude, which indicate feature characteristics. However, in some circumstance, the estimates may not be fully certain. Little attention has been paid to assessing the reliability or uncertainty of Gaussian estimates. It is necessary to take such indicators into account when application of multiple waveform features is attempted. This study aims to fill this research gap. Whether the reliability of the estimates is affected by the complexity of the waveform shape was investigated. Waveform data collected from simulation experiment and a Riegl LMS-Q560 field campaign was analyzed. Several targets were designed and set in the field to observe how the shape of waveforms varies with the separation distance between two returns and their amplitude magnitude. It was found that the RMSE values for the pulse width estimates based on data simulation were increasingly large when the separation distance was decreased (＜ 6 ns). The RMSE values for the range estimates based on data simulation remained consistently small (~ 0.04 ns) when decreasing the separation distance.