AutoMovie API
    Preparing search index...

    Function linearizeRecoveryPoseResiduals

    • Signed pixel Sampson residuals and analytic five-coordinate local Jacobian. With e=x2-transpose F x1 and d the joint epipolar-normal norm, each derivative is de/d - (e/d)sum((line/d)(derivativeLine/d)). This differentiates both the incidence and its pixel normalization; zero/undefined lines refuse the fit.

      Parameters

      • model: {
            derivatives: number[][];
            matrix: number[];
            tangent: [number, number, number][];
        }
        • derivatives: number[][]

          Derivative matrices of the current epipolar model, ordered by the declared pose tangent coordinates.

        • matrix: number[]

          Row-major 3 by 3 epipolar model in the declared image charts; homogeneous scale is arbitrary.

        • tangent: [number, number, number][]

          Local translation-direction tangent basis used with the current rotation increments.

      • first: readonly (readonly [number, number])[]
      • second: readonly (readonly [number, number])[]

      Returns { jacobian: number[][]; residuals: number[] }

      • jacobian: number[][]

        Residual derivative rows in the current pose tangent-coordinate ordering.

      • residuals: number[]

        Signed pixel residuals aligned with the supplied correspondence and Jacobian row order.