Вы находитесь на странице: 1из 2

1.

Real-time Signal Light Detection (1)

Architecture: Grab image, color extraction, clustering, circularity check,


frame check, robust filtering, detection result
Color extraction: using statistical processing (Average value
of red, green, blue and value of standard deviation are determined for
each sample image of
signal light. Select pixels within range that is determined from average
value and standard
deviation and constant)
Clustering: K-means centroid clustering
Circularity check: Haralick algorithm
Frame check: determine if traffic light
Detection success rate is with 93.9%.
2.

Traffic Light Mapping and Detection (37259)


- Map of traffic lights

3. Robust and Real-Time Traffic Lights Recognition in Complex Urban


Environments (18756891%2E2011%2E9727889)
HSI color model to analyze colors of traffic lights.
Filtering of Candidate Regions: First, the noise removal is carried out on the
image using the morphology operation, and scatter small blocks are removed and connected regions are
preserved accurately. Secondly, each candidate region is tested according to their shape feature and
validated or discarded depending on whether they satisfied the dimension property and the border
property of the real traffic lights.
Filtering based on shape feature: dimension based thresholding, border property based tresholding

Template Matching Using Normalized Cross Correlation: The template matching


degree is evaluated by
calculating the normalized cross-correlation coefficient between the selected templates and the
candidate region. If the normalized cross-correlation value is higher than a fixed threshold, the region is
considered as corresponding to a traffic light.
Otherwise it is rejected.
Precision: 96-97%
4.

ArticleFawzi

we decided to base our algorithm on the light property emission of traffics lights
First Step First a robust Spot Light Detection (SLD) is executed on the whole grayscale source image in order to
detect all visible spots lights
Second Step After an intermediate association step which goal is to define relevant candidates region from the
previously detected spots, an Adaptive Template Matcher (ATM) is executed
Third Step Finally, a simple validation step is used to filter the candidates marked by the ATM.
Spot Light Detection: top-hat morphological operator to detect the bright areas and then applies shape filtering and
region growing algorithm to keep only the consistent areas

5.

ijicic-10-08060

Detection of traffic lights:


1.
2.

H-S color segmentation


Genetic approach to aprox ellipse detection

Traffic light color detection: spatial texture layouts


1.

Extraction of the entire traffic light unit

Feature extraction and recognition


Results: 94%

Вам также может понравиться