Skip to main content

BriefCam White Papers

Searching for License Plates in BriefCam

Last Updated: 2 minute read
Version2024r2
LanguageEnglish

Using BriefCam, you can search for license plates across sources. You can search using existing vehicles in the video or by watchlists.

Watchlists

One of the ways to search for plates in videos is by using watchlists, which are created by manually inputting license plate numbers or uploading a CSV file with two columns (License Plate Number and Description).

Watchlist Size

The size of the watchlist impacts accuracy and performance.

Intuitively, one can understand that the larger the watchlist – the higher the chance of false positives (as there is more chance that the watchlist will contain a plate similar to the plate in the video).

Fuzzy Search and Matching

LPR readings (especially “in the wild” LPR readings from surveillance cameras) are not always accurate and can have transcription errors with some of the transcribed characters.

The chances of having two vehicles with similar license plates in videos is very low. Therefore, BriefCam considers two similar plates as the same plate when searching and matching, depending on the number defined in the Mismatches allowed selector, which acts as the tolerance level for license plates.

BriefCam uses a similarity measure called Levenshtein distance, which represents the number of substitutions needed to transform one string into another.

For example, the Levenshtein distance between "kitten" and "sitting" is 3, since the following three substitutions are needed, and there is no way to do it with fewer than three substitutions:

  1. kitten → sitten (substitution of "s" for "k")

  2. sitten → sittin (substitution of "i" for "e")

  3. sittin → sitting (insertion of "g" at the end).

During the matching process, different values set in the Mismatches allowed selector (from 0 to 8), allow strings with different Levenshtein distances to be matched.

As the number of objects in the case grows, the number of false matches might increase.

Note

The Mismatches allowed selector can be used in REVIEW when searching for a plate and in RESPOND when defining the rule. It is not available in RESEARCH and matching between two plates in RESEARCH may not succeed if one of them was inaccurately read. On good quality cameras this is expected to average-out when many pairs are matched.

Wildcards

Wildcards can be used when searching for a license plate either manually or in a watchlist.

The supported wildcards are:

Wildcard

Description

?

Replaces any single character or number. Question marks can be used anywhere in the string. The string ???JK will match any license plate that has J and K as the fourth and fifth character. For example, the following will be found for this search: ABCJK, 251JK and more.

*

Replaces any number (zero or more) of characters and numbers. Asterisks can be used at the beginning or end of the string.

For example: The string *JK will match with the following: 0974JK, 25JK, JK and more.

Note that the string *JK? will match with the following: 0974JK2, 25JK0, JK1 and more.