Method and system for detecting lane pattern
The present disclosure is related in general to field of machine learning and image processing, that provides a method and system for detecting lane pattern. A lane classification system receives an input image comprising lane markings and co-ordinates of the lane markings, which is divided into two portions, a first image portion and a second image portion based on the lane markings and the co-ordinates of the lane markings. Further, the first image portion and the second image portion are resized into a predefined template size. The lane classification system detects a lane pattern in the two portions by analysing each of the one or more lane markings located in the resized first image portion and the second image portion, independently, based on historical lane data.
1. A lane classification system for detecting lane pattern, the lane classification system comprising:
an image sensor mounted to a host vehicle, to capture images of an area in front of the host vehicle;
a processor communicatively connected to the image sensor and configured to:
receive a plurality of training images from the image sensor, wherein each of the plurality of training images comprises one or more lane markings and co-ordinates of the one or more lane markings;
divide each of the plurality of training image into two portions, a first image portion and a second image portion, wherein the first image portion is determined based on the one or more lane markings and the co-ordinates of the one or more lane markings located on a left side of each of the plurality of training images and the second image portion is determined based on the one or more lane markings and the co-ordinates of the one or more lane markings located on a right side of each of the plurality of training images;
resize the first image portion and the second image portion into a predefined template size;
train a first machine learning model using the resized first image portion of each of plurality of training images and a second machine learning model using the resized second image portion of each of plurality of training images;
receive an input image from the image sensor and divide the input image into the first image portion of the input image and the second image portion of the input image;
detect a lane pattern of the input image by inputting the first image portion of the input image to the first machine learning model and the second image portion of the input image to the second machine learning model; and
control a navigation of the host vehicle based on the detected lane pattern.
2. The lane classification system as claimed in claim 1 , wherein, to detect the lane pattern, the processor is further configured to:
determine a binary value corresponding to each of the one or more lane markings located in the first image portion and the second image portion based on the plurality of training images, wherein the binary value is determined using each of the trained first and second machine learning models; and
detect the lane pattern based on the binary value determined for each of the one or more lane markings.
3. The lane classification system as claimed in claim 2 , wherein the training images are stored in a database associated with the lane classification system.
4. The lane classification system as claimed in claim 1 , wherein the lane pattern comprises at least one of a solid lane or a dashed lane.
5. The lane classification system as claimed in claim 1 , wherein the input image is a ground truth (GT) image that is generated from an original image using a feature extraction and prediction model.
6. An Advanced Driver-Assistance System (ADAS) comprising:
a lane classification system;
a display device; and
a vehicle controlling system;
wherein the lane classification system comprises:
an image sensor mounted to a host vehicle; to capture images of an area in front of the host vehicle; and
a processor communicatively connected to the image sensor and configured to:
receive a plurality of training images from the image sensor, wherein each of the plurality of training images comprises one or more lane markings and co-ordinates of the one or more lane markings;
divide each of the plurality of training images into two portions, a first image portion and a second image portion, wherein the first image portion is determined based on the one or more lane markings and the co-ordinates of the one or more lane markings located on a left side of each of the plurality of training images and the second image portion is determined based on the one or more lane markings and the co-ordinates of the one or more lane markings located on a right side of each of the plurality of training images;
resize the first image portion and the second image portion into a predefined template size;
train a first machine learning model using the resized first image portion of each of plurality of training images and a second machine learning model using the resized second image portion of each of plurality of training images;
receive an input image from the image sensor and divide the input image into the first image portion of the input image and the second image portion of the input image; and
detect a lane pattern of the input image by inputting the first image portion of the input image to the first machine learning model and the second image portion of the input image to the second machine learning model,
the display device associated with the lane classification system is configured to display the detected lane pattern; and
the vehicle controlling system associated with the lane classification system is configured to control navigation of the host vehicle, in real-time, based on the detected lane pattern.
7. A method of detecting lane pattern, the method comprising:
receiving, by a processor communicatively connected to an image sensor mounted to a host vehicle, a plurality of training images, wherein each of the plurality of training images comprises one or more lane markings and co-ordinates of the one or more lane markings;
dividing, using the processor, each of the plurality of training images into two portions, a first image portion and a second image portion, wherein the first image portion is determined based on the one or more lane markings and the co-ordinates of the one or more lane markings located on a left side of each of the plurality of training images and the second image portion is determined based on the one or more lane markings and the co-ordinates of the one or more lane markings located on a right side of each of the plurality of training images;
resizing, using the processor, the first image portion and the second image portion into a predefined template size;
training, using the processor, a first machine learning model using the resized first image portion of each of plurality of training images and a second machine learning model using the resized second image portion of each of plurality of training images;
receiving, using the processor, an input image from the image sensor and divide the input image into the first image portion of the input image and the second image portion of the input image;
detecting, using the processor, a lane pattern of the input image by inputting the first image portion of the input image to the first machine learning model and the second image portion of the input image to the second machine learning model; and
controlling, using the processor, a navigation of the host vehicle based on the detected lane pattern.
8. The method as claimed in claim 7 , wherein detecting the lane pattern further comprises:
determining, using the processor, a binary value corresponding to each of the one or more lane markings located in the first image portion and the second image portion based on the plurality of training images, wherein the binary value is determined using each of the trained first and second machine learning models; and
detecting, using the processor, the lane pattern based on the binary value determined for each of the one or more lane markings.
9. The method as claimed in claim 7 , wherein the training images are stored in a database associated with the lane classification system.
10. The method as claimed in claim 7 , wherein the lane pattern comprises at least one of a solid lane or a dashed lane.
11. The method as claimed in claim 7 , wherein the input image is a ground truth (GT) image that is generated from an original image using a feature extraction and prediction model.