Image processing system and method for recognition performance enhancement of CIPV

1 . An image processing system, embedded in mounted on a vehicle and having a predetermined amount of computation resources, for recognition performance enhancement of a closest in-path vehicle (CIPV) without exceeding the computation resources of the image processing system, the system comprising:

a size adjusting unit configured to convert a size of input forward image data of the vehicle, which input forward image data is obtained from a camera mounted on the vehicle, to provide converted forward image data of a predetermined size to reduce an amount of computation necessary for recognition of an object in the input forward image data;

an object recognition unit configured to analyze the converted forward image data and to recognize the object included in the converted forward image data; and

a result processing unit configured to inversely convert a size of the recognized object in response to an operation of the size adjusting unit to resolve image distortion included in the converted image data and, following inversely converting the size of the recognized object, to analyze an actual distance between the recognized object and the vehicle without exceeding the computation resources of the image processing system,

wherein the image processing system is part of an advanced driver assistance system (ADAS) configured to be used for semi-autonomous or autonomous driving of the vehicle.

2 . The system of claim 1 , wherein the size adjusting unit is configured to:

set a first predetermined value as a first decimation factor of a first predetermined region range of first predetermined coordinates of the forward image data, and

set a second predetermined value as a second decimation factor of a second predetermined region range of second predetermined coordinates of the forward image data,

wherein the second predetermined value is a larger decimation factor than the first predetermined value.

3 . The system of claim 2 , wherein the size adjusting unit is configured to set a decimation factor which is increased from a value that exceeds the first predetermined value to the second predetermined value as a region range is moved further away from the first region range of the first predetermined coordinates based on the first region range of the first predetermined coordinates.

4 . The system of claim 2 , wherein the size adjusting unit is configured to set a value between the first predetermined value and the second predetermined value as a decimation factor of a region range of coordinates between the first region range of the first predetermined coordinates and the second region range of the second predetermined coordinates.

5 . The system of claim 2 , wherein the first region range of the first predetermined coordinates is a predetermined CIPV region.

6 . The system of claim 2 , wherein the result processing unit includes:

an inverse conversion unit configured to inversely convert the size of the recognized object in consideration of a region range where the recognized object is positioned, in response to an operation of the size adjusting unit; and

an analysis unit configured to analyze a distance between the vehicle and the recognized object by converting image coordinates of the object inversely converted by the inverse conversion unit into actual distance coordinates in consideration of a predetermined factor based on a basic specification of a device used for acquiring the forward image data.

7 . An image processing method for recognition performance enhancement of a closest in-path vehicle (CIPV) using an image processing system, embedded in a vehicle and having a predetermined amount of computation resources, for recognition performance enhancement of CIPV in which each operation is performed by a computation processing device without exceeding the computation resources of the image processing system, the method comprising:

converting an image size of input forward image data of a vehicle, which input forward image data is obtained from a camera mounted on the vehicle, to provide converted forward image data of a predetermined size to reduce an amount of computation necessary for recognition of an object in the input forward image data;

recognizing the object by analyzing the converted forward image data and recognizing the object included in the converted forward image data; and

inversely converting a size of the recognized object in response to an operation in the converting of the image size to resolve image distortion included in the converted image data and, following inversely converting the size of the recognized object, analyzing an actual distance between the recognized object and the vehicle without exceeding the computation resources of the image processing system,

wherein the image processing system is part of an advanced driver assistance system (ADAS) configured to be used for semi-autonomous or autonomous driving of the vehicle.

8 . The method of claim 7 , wherein in the converting of the image size:

a first predetermined value is set as a first decimation factor of a first predetermined region range of first predetermined coordinates of the forward image data, and

a second predetermined value is set as a second decimation factor of a second predetermined region range of second predetermined coordinates of the forward image data,

the region range of the first predetermined coordinates is a predetermined CIPV region, and

the second predetermined value is a larger decimation factor than the first predetermined value.

9 . The method of claim 8 , wherein, in the converting of the image size, a value between the first predetermined value and the second predetermined value is set as a decimation factor of a region range of coordinates between the first region range of the first predetermined coordinates and the second region range of the second predetermined coordinates.

10 . The method of claim 8 , wherein, in the converting of the image size, a decimation factor is set to be increased from a value that exceeds the first predetermined value to the second predetermined value as a region range is moved further away from the first region range of the first predetermined coordinates based on the first region range of the first predetermined coordinates.

11 . The method of claim 8 , wherein the method further includes:

performing inverse conversion of inversely converting the size of the recognized object in consideration of a region range where the recognized object is positioned, in response to the operation in the converting of the image size; and

performing a distance analysis of analyzing a distance between the vehicle and the recognized object by converting image coordinates of the inversely converted object into actual distance coordinates in consideration of a predetermined factor based on a basic specification of a device used for acquiring the forward image data.

12 . The system of claim 1 , wherein the result processing unit is configured to inversely convert the size of the recognized object in response to an operation of the size adjusting unit to an original image size of the recognized object included in the input forward image data to analyze an actual distance between the recognized object and the vehicle.

13 . The system of claim 1 , wherein:

the size adjusting unit converts the size of the input forward image data in a manner which disposes a greater number of pixels in a first predetermined region range of first predetermined coordinates of the forward image data than a number of pixels disposed in a second predetermined region range of second predetermined coordinates of the forward image data; and

the object recognition unit is configured to analyze the converted forward image data and to recognize an object included in the first predetermined region range of first predetermined coordinates in the converted forward image data.

14 . The system of claim 13 , wherein the result processing unit is configured to inversely convert the size of the recognized object in response to an operation of the size adjusting unit to an original image size of the recognized object included in the input forward image data to analyze an actual distance between the recognized object and the vehicle.

15 . The system of claim 1 , wherein the size adjusting unit is configured to convert the size of input forward image data to reduce an entire image size while maintaining a full high definition (FHD) resolution of a number of pixels in a CIPV region of the input forward image data.