• Infrared and Laser Engineering
  • Vol. 52, Issue 2, 20220376 (2023)
Liying Wang1, Ze You1, Ji Wu2, and Mahamadou CAMARA1
Author Affiliations
  • 1School of Geomatics, Liaoning Technical University, Fuxin 123000, China
  • 2Heilongjiang Institute of Geomatics Engineering, Harbin 150081, China
  • show less
    DOI: 10.3788/IRLA20220376 Cite this Article
    Liying Wang, Ze You, Ji Wu, Mahamadou CAMARA. Airborne MS-LiDAR data classification by combining NDRI features and spatial correlation[J]. Infrared and Laser Engineering, 2023, 52(2): 20220376 Copy Citation Text show less
    Titan MS-LiDAR point cloud of Area1 and its processing results. (a) C1 intensity; (b) C2 intensity; (c) C3 intensity; (d) Points colored by elevation; (e) Merged point cloud (R = C1 intensity, G = C2 intensity and B = C3 intensity);(f) 3D land cover classified point cloud
    Fig. 1. Titan MS-LiDAR point cloud of Area1 and its processing results. (a) C1 intensity; (b) C2 intensity; (c) C3 intensity; (d) Points colored by elevation; (e) Merged point cloud (R = C1 intensity, G = C2 intensity and B = C3 intensity);(f) 3D land cover classified point cloud
    Titan MS-LiDAR point cloud of Area2 and its processing results. (a) C1 intensity; (b) C2 intensity; (c) C3 intensity; (d) Points colored by elevation; (e) Merged point cloud (R = C1 intensity, G = C2 intensity and B = C3 intensity);(f) 3D land cover classified point cloud
    Fig. 2. Titan MS-LiDAR point cloud of Area2 and its processing results. (a) C1 intensity; (b) C2 intensity; (c) C3 intensity; (d) Points colored by elevation; (e) Merged point cloud (R = C1 intensity, G = C2 intensity and B = C3 intensity);(f) 3D land cover classified point cloud
    Flowchart of the proposed algorithm
    Fig. 3. Flowchart of the proposed algorithm
    Relationships between wavelengths and reflectance properties of objects[26]
    Fig. 4. Relationships between wavelengths and reflectance properties of objects[26]
    NDRI statistical histogram for ground points of Area1
    Fig. 5. NDRI statistical histogram for ground points of Area1
    NDRI statistics histogram for different objects of Area1. (a) Grass; (b) Road; (c) Tree; (d) Building
    Fig. 6. NDRI statistics histogram for different objects of Area1. (a) Grass; (b) Road; (c) Tree; (d) Building
    Differentiability of different NDRI used for distinguishing from objects of Area1. (a) C1−C2; (b) C1−C3; (c) C2−C3
    Fig. 7. Differentiability of different NDRI used for distinguishing from objects of Area1. (a) C1−C2; (b) C1−C3; (c) C2−C3
    ParameterSpecification
    WavelengthsC1: 1550 nm, C2: 1064 nm, C3: 532 nm
    Forward anglesC1: 3.5°, C2: 0°, C3: 7°
    Beam divergenceC1: 0.35 mrad, C2: 0.35 mrad, C3: 0.7 mrad
    AltitudeTopographic: 300-2000 m above ground level (AGL), all channels
    Pulse repetition frequency 50-300 kHz/channel; 900 kHz total
    Scan angle (FOV)Programmable; 0-60° max
    Scan frequencyProgrammable; 0-210 Hz
    Swath width0-115% AGL
    AccuracyHorizontal: 1/7500×altitude, 1σ; Vertical: < 5-10 cm, 1 σ
    Laser range precision<0.008 m, 1σ
    Table 1. Specifications of Titan
    Experimental areaType I errorType II errorTotal error
    Area15.24%1.87%3.90%
    Area22.61%4.25%3.26%
    Table 2. The errors of the filtering results
    NDRIArea1Area2
    Ground pointsNon- ground points Ground pointsNon- ground points
    C2−C3 0.3440.3730.2960.367
    C1−C3 0.2270.2840.3440.373
    C1−C2 0.4050.2620.2310.258
    Table 3. NDRI threshold t*
    NDRIArea1Area2
    OAKappaOAKappa
    C2−C3 90.45%0.869%89.89%0.853%
    C1−C3 88.59%0.841%86.98%0.812%
    C1−C2 80.68%0.738%83.07%0.752%
    Table 4. Classification accuracy of different NDRI indexes
    Classification dataReference dataTotal rowUser’s accuracy
    RoadsGrassTreesBuildings
    Roads24501 (A1) 1524 (A2) 939 (A3) 746 (A4) 27710 (Ta) 88.42% (Ua=A1/Ta)
    Grass1386 (B1) 40260 (B2) 1137 (B3) 1128 (B4) 43911 (Tb) 91.69% (Ub=B2/Tb)
    Trees234 (C1) 360 (C2) 24019 (C3) 1594 (C4) 26207 (Tc) 91.65% (Uc=C3/Tc)
    Buildings40 (D1) 248 (D2) 2017 (D3) 18750 (D4) 21055 (Td) 89.05% (Ud=D4/Td)
    Total column26161 (T1) 42392 (T2) 28112 (T3) 22218 (T4) 118883
    Producer’s accuracy 93.65% (P1=A1/T1) 94.97% P2=B2/T285.44% (P3=C3/T3) 84.39% (P4=D1/T4)
    Overall accuracy: 90.45%; Kappa statistic: 0.869
    Table 5. Confusion matrix of the land cover classification result (Area1)
    Classification data Reference dataTotal rowUser’s ccuracy
    RoadsGrassTreesBuildings
    Roads53 97411 4221 13069567 22180.29%
    Grass9 165150 1016 7122 175168 15389.26%
    Trees23961993 4842 09596 43796.94%
    Buildings2822495 22235 43541 18886.03%
    Total column63 660162 391106 54840 400372 999
    Producer’s accuracy84.78%92.43%87.74%87.71%
    Overall accuracy: 89.89%; Kappa statistic: 0.853
    Table 6. Confusion matrix of the land cover classification result (Area2)
    AuthorsAlgorithm principleFeaturesOAKappa
    Our paperStep-by-step separationSpatial;NDRI90.17%0.861
    Fernandez-Diaz et al[16]Mahalanobis distanceFive structural; three intensity images82.33%0.77
    Chen et al[13]SVM;k-NN re- classificationSpectral reflectance of four channels; five vegetation indexes; neighborhood spatial 87.19%N/A
    Fernandez- Diaz et al[16]Maximum likelihoodFive structural; two intensity images90.22%0.870
    Zou et al[20]Decision treePseudo normalized difference vegetation index; ratio of green; ratio of returns counts; difference of elevation between maximum elevation of first returns and minimum elevation of last returns 91.63%0.895
    Chen et al[13]Random forest7 spectral features; 11 geometric features93.00%N/A
    Wang and Gu[9]SVMSpatial location; spectral; neighborhood geometric; spectral structures; geometric-spectral 94.76%0.935
    Table 7. Accuracy comparison between the proposed algorithm and the other classical algorithms
    Liying Wang, Ze You, Ji Wu, Mahamadou CAMARA. Airborne MS-LiDAR data classification by combining NDRI features and spatial correlation[J]. Infrared and Laser Engineering, 2023, 52(2): 20220376
    Download Citation