AutoMovie API
    Preparing search index...
    Operation: (
        fundamental: readonly number[],
        firstPoints: readonly (readonly [number, number])[],
        secondPoints: readonly (readonly [number, number])[],
        first: ReturnType<IAutoMovieRecoveryCameraChartCreationOperation.Operation>,
        second: ReturnType<
            IAutoMovieRecoveryCameraChartCreationOperation.Operation,
        >,
        minimumDesignRatio: number,
        maxSweeps: number,
    ) => {
        allPositiveDepthCounts: number[];
        fundamental: number[];
        medianParallaxRadians: number | null;
        points: ReturnType<
            IAutoMovieRecoveryCameraPointTriangulationOperation.Operation,
        >[];
        positiveDepthFraction: number;
        relativeProjectionDiscrepancy: number;
        rotation: number[];
        translation: [number, number, number];
        uniqueWinner: boolean;
    }

    Measure all four calibrated pose branches on one fixed correspondence set. The largest positive-depth count supplies a candidate; equal winning counts remain explicitly ambiguous even though deterministic order names a candidate. Parallax is summarized only for finite points positive in both cameras.

    Type Declaration

      • (
            fundamental: readonly number[],
            firstPoints: readonly (readonly [number, number])[],
            secondPoints: readonly (readonly [number, number])[],
            first: ReturnType<IAutoMovieRecoveryCameraChartCreationOperation.Operation>,
            second: ReturnType<
                IAutoMovieRecoveryCameraChartCreationOperation.Operation,
            >,
            minimumDesignRatio: number,
            maxSweeps: number,
        ): {
            allPositiveDepthCounts: number[];
            fundamental: number[];
            medianParallaxRadians: number | null;
            points: ReturnType<
                IAutoMovieRecoveryCameraPointTriangulationOperation.Operation,
            >[];
            positiveDepthFraction: number;
            relativeProjectionDiscrepancy: number;
            rotation: number[];
            translation: [number, number, number];
            uniqueWinner: boolean;
        }
      • Parameters

        Returns {
            allPositiveDepthCounts: number[];
            fundamental: number[];
            medianParallaxRadians: number | null;
            points: ReturnType<
                IAutoMovieRecoveryCameraPointTriangulationOperation.Operation,
            >[];
            positiveDepthFraction: number;
            relativeProjectionDiscrepancy: number;
            rotation: number[];
            translation: [number, number, number];
            uniqueWinner: boolean;
        }

        • allPositiveDepthCounts: number[]

          Positive-depth support count for each candidate pose in candidate order.

        • fundamental: number[]

          Row-major 3 by 3 fundamental matrix in the declared pixel charts; homogeneous scale has no metric meaning.

        • medianParallaxRadians: number | null

          Median accepted ray angle in radians, or null when no usable population exists.

        • points: ReturnType<IAutoMovieRecoveryCameraPointTriangulationOperation.Operation>[]

          Triangulated correspondence results in input order, retaining depth and pixel-error qualifications.

        • positiveDepthFraction: number

          Fraction of evaluated correspondences having positive depth in both cameras.

        • relativeProjectionDiscrepancy: number

          Relative discrepancy between the projected essential model and the estimated epipolar model; dimensionless.

        • rotation: number[]

          Row-major 3 by 3 camera rotation in the operation's declared source-to-camera convention.

        • translation: [number, number, number]

          Camera translation in the operation's admitted reconstruction scale; two-view direction alone has no metric length.

        • uniqueWinner: boolean

          Whether one pose has uniquely greatest admitted positive-depth support.