Georeferencing API
    Preparing search index...

    Interface Fit

    Serializable fitted transformation and training-point diagnostics. Treat coefficients as opaque and evaluate with forward and backward.

    interface Fit {
        backward: "reverse-polynomial" | "analytic-or-newton";
        condition: number;
        knots?: XY[];
        model: Model;
        residuals: {
            distance: number;
            id: string;
            pixels: number | null;
            vector: XY;
        }[];
        reversePolynomial?: { x: number[]; y: number[] };
        rmse: number;
        seed: number[][];
        source: Normalization;
        target: Normalization;
        x: number[];
        y: number[];
    }
    Index
    backward condition knots? model residuals reversePolynomial? rmse seed source target x y
    backward: "reverse-polynomial" | "analytic-or-newton"

    Backward-mapping convention used for preview/export and pixel residuals.

    condition: number

    Condition estimate of the normalized fitting system; values above 1e10 are rejected.

    knots?: XY[]

    Normalized source GCP locations for TPS radial terms; absent for other models.

    model: Model

    Model used to produce these coefficients.

    residuals: { distance: number; id: string; pixels: number | null; vector: XY }[]

    Diagnostics for enabled GCPs only, in input order.

    Type Declaration

    • distance: number

      Euclidean length of the forward residual in working-CRS units.

    • id: string

      Stable GCP identifier.

    • pixels: number | null

      Distance between the backward-mapped target and source point in original pixels, or null when inversion fails.

    • vector: XY

      Forward residual T(image) - target in working-CRS units.

    reversePolynomial?: { x: number[]; y: number[] }

    Separately fitted reverse polynomial for GDAL/QGIS-compatible polynomial raster sampling.

    Type Declaration

    • x: number[]

      Normalized reverse x coefficients in polynomial basis order.

    • y: number[]

      Normalized reverse y coefficients in polynomial basis order.

    rmse: number

    Training RMSE: sqrt(sum(distance²) / enabledCount) in working-CRS units. This is not an independent accuracy estimate.

    seed: number[][]

    Reverse affine least-squares coefficients used to initialize Newton inversion.

    source: Normalization

    Normalization of canonical image pixel coordinates.

    target: Normalization

    Normalization of target coordinates in the working CRS.

    x: number[]

    Normalized x-output coefficients; for projective fits, the complete row-major 3×3 homography.

    y: number[]

    Normalized y-output coefficients; unused for projective fits.

    Settings
    Member Visibility

    On This Page
    Properties