• Optics and Precision Engineering
  • Vol. 31, Issue 7, 1096 (2023)
Zhipeng WANG and Xiaofei WANG*
Author Affiliations
  • Electronic Engineering College, Heilongjiang University, Harbin, 150080China
  • show less
    DOI: 10.37188/OPE.20233107.1096 Cite this Article
    Zhipeng WANG, Xiaofei WANG. Automatic extraction of red raspberry planting areas using time series multispectral images[J]. Optics and Precision Engineering, 2023, 31(7): 1096 Copy Citation Text show less
    Study area overview
    Fig. 1. Study area overview
    Flow chart for planing area extraction of red raspberry using CART
    Fig. 2. Flow chart for planing area extraction of red raspberry using CART
    NDVI curves before and after smoothing
    Fig. 3. NDVI curves before and after smoothing
    Extraction results of raspberry planting area in study area by different methods
    Fig. 4. Extraction results of raspberry planting area in study area by different methods
    Results of raspberry extraction in the test area
    Fig. 5. Results of raspberry extraction in the test area
    品 种萌芽期展叶期现蕾期始花期终花期果实成熟期落叶期
    夏果型4.245.56.76.157.17.189.18
    秋果型4.255.88.78.128.198.2610.1
    Table 1. Phenological dates of red raspberry in study area
    波 段波段号波长/nm中心波长/nm空间分辨率/m

    可见光

    (400~740 nm)

    B1430~45744360
    B2440~53849010
    B3537~58256010
    B4646~68466510
    B5694~71370520
    B6731~74974020

    近红外

    (740~1 300 nm)

    B7769~79778320
    B8760~90884210
    B8a848~88186520
    B9932~95894560

    短波红外

    (1 300~2 500 nm)

    B101 337~1 4121 37560
    B111 539~1 6821 61020
    B122 017~2 3202 19020
    Table 2. Band information from Sentinel-2 satellite
    地物类型分离度
    旱田地与建筑1.880 339 68
    旱田地与水田地2.000 000 00
    旱田地与林地2.000 000 00
    水田地与建筑1.991 178 63
    水田地与林地1.999 990 74
    水体与建筑1.996 851 72
    水体与林地1.999 266 07
    水体与旱田地1.999 998 23
    树莓与建筑1.998 156 00
    树莓与林地1.999 300 53
    树莓与旱田地1.999 997 83
    树莓与水体1.999 999 95
    树莓与水田地2.000 000 00
    Table 3. Separation of some land types
    提取区域实验方案数 据总精度(OA)Kappa
    研究区最佳时相94.47%0.93
    最佳时相+NDVI时序96.64%0.95
    面向对象分类94.81%0.93
    支持向量机分类93.89%0.92
    实验区最佳时相96.06%0.93
    最佳时相+NDVI时序98.26%0.97
    Table 4. Accuracy evaluation
    地 区

    实际种植

    面积/亩

    本文最优方法提取

    面积/亩

    面向对象分类提取

    面积/亩

    支持向量机提取

    面积/亩

    尚志市25 00025 46027 56328 425
    一面坡镇3 5003 5413 7243 859
    亮河镇8 1008 0178 2237 958
    马延乡东兴村150153171118
    宾县常安镇3 1803 2053 3273 466
    宾县宁远镇1 2401 2621 3431 572
    Table 5. Comparison of extracted red raspberry planting area and actual planting area
    Zhipeng WANG, Xiaofei WANG. Automatic extraction of red raspberry planting areas using time series multispectral images[J]. Optics and Precision Engineering, 2023, 31(7): 1096
    Download Citation