Applies the given reducer to the neighborhood around each pixel, as determined by the given kernel. If the reducer has a single input, it will be applied separately to each band of the collection; otherwise it must have the same number of inputs as the input image has bands.

The reducer output names determine the names of the output bands:

- Reducers with multiple inputs will use the output names directly.
- Reducers with a single input will prefix the output name with the input band name (e.g., '10_mean', '20_mean').

Reducers with weighted inputs can have the input weight based on the input mask, the kernel value, or the smaller of those two.

| Usage | Returns |
|---|---|
| `Image.reduceNeighborhood(reducer, kernel, *inputWeight*, *skipMasked*, *optimization*)` | Image |

| Argument | Type | Details |
|---|---|---|
| this: `image` | Image | The input image. |
| `reducer` | Reducer | The reducer to apply to pixels within the neighborhood. |
| `kernel` | Kernel | The kernel defining the neighborhood. |
| `inputWeight` | String, default: "kernel" | One of 'mask', 'kernel', or 'min'. |
| `skipMasked` | Boolean, default: true | Mask output pixels if the corresponding input pixel is masked. |
| `optimization` | String, default: null | Optimization strategy. Options are 'boxcar' and 'window'. The 'boxcar' method is a fast method for computing count, sum or mean. It requires a homogeneous kernel, a single-input reducer and either MASK, KERNEL or no weighting. The 'window' method uses a running window, and has the same requirements as 'boxcar', but can use any single input reducer. Both methods require considerable additional memory. |