Function sobel_l1_magnitude_u8 

pub fn sobel_l1_magnitude_u8(
    input: ImageView<'_, u8, 1>,
    border: BorderMode<u8, 1>,
) -> Result<Image<i16, 1>, VisionError>

Computes exact 3×3 Sobel L1 magnitude (|Gx| + |Gy|) for grayscale u8.

The non-negative signed i16 output ranges from 0 through 2040. Fusing the paired derivatives and magnitude avoids materializing two gradient images.