# Auto-detect with manual review

When using CameraSnap in the **Auto-detect with review** mode, **CameraSnap** finds matching features (generally referred to as keypoints), then displays these feature matches in the **CameraSnap** **Match Explorer**:

<figure><img src="https://2222094320-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FdEevfLZRIk38LUPwDa4V%2Fuploads%2Fz0GOpoNfVs5I2Qg3uMsU%2Fimage.png?alt=media&#x26;token=bf47e164-9bb2-4c22-8944-060abf5f9ab3" alt=""><figcaption></figcaption></figure>

This allows the user to remove or add matches.  Matches shown in the table are a feature match between two images. The image numbers and receptor numbers associated with the match are displayed in the leftmost columns of the table. To manually add a match, so the [Manually-Created mode](https://docs.phoenixlidar.com/spatialexplorer-8-and-9/user-interface/toolbars/workflow/camera-snap/manually-created-matches) (the process is the same).

**Approach Distance** is one of the most useful fields in the Match Explorer. Approach Distance is calculated by projecting a ray from each image to the keypoint (or feature), and computing the difference in position of the rays. Generally, the smaller approach distance, the better the match.&#x20;

{% hint style="info" %}
You can sort the matches by clicking on the column headers. A quick and easy way to filter out bad matches is to sort by **Approach Distance**, then select and remove all matches with an approach distance greater than approximately **0.10 m**.&#x20;
{% endhint %}

Try to retain at least 20 matches, however ideally you have much more. Once you have filtered the matches, you can click **Calibrate** and CameraSnap will proceed with Calibration using the available feature matches. Review the [CameraSnap report](https://docs.phoenixlidar.com/spatialexplorer-8-and-9/user-interface/toolbars/workflow/camera-snap/camerasnap-reports) to ensure a proper calibration was achieved.&#x20;
