GstCUDA - Example - opencvfilter

GstCUDA
Features and Limitations
Supported Platforms
Evaluating GstCUDA
Building and Installation Guide
Project Structure
GstCUDA Framework
libGstCUDA API GstCUDABaseFilter GstCUDABaseMISO GstCUDABaseSIMO GstCUDABaseMIMO
GstCUDA Quick Prototyping Elements
cudafilter cudamux cudademux cudamimo
GstCUDA Algorithm Developer’s Guide
GstCUDA Add-Ons
cudadebayer
Examples
cudafilter cudafilter: NVMM direct mapping cudafilter: Unified memory allocator cudamux cudamux: NVMM direct mapping cudamux: Unified memory allocator cudadebayer opencv opencvwarp opencvfilter
Performance Profiling
Contact Us

Introduction

The opencvfilter element exposes a series of image filters from the popular computer vision framework OpenCV. The following figure shows some of the available filters applied to the same image.

Figure 1. Example of opencv available filters.

Element properties

• size
  

The size of the filter. Depending on the filter, it could either be the kernel size or the aperture size. Must be odd.
Type: Integer
Range: 1 - 2147483647
Flags: readable, writable
Default: 3

• filter
  

Type of the filter to create. Available options:

box: Normalized box filter

box-max: Maximum filter

box-min: Minimum filter

column-sum: Vertical 1D box filter

deriv: General deriv filter

gaussian: Gaussian filter

laplacian: Laplacian filter

linear: General linear filter

morphology: Morphological filter

row-sum: Horizontal 1D box filter

scharr: Scharr filter

separable: Separable linear filter

sobel: Sobel filter

Flags: readable, writable
Default: "gaussian"

• dx
  

Derivative order in respect of X (for deriv, sobel and scharr).
Type: Integer
Range: 0 - 2
Flags: readable, writable
Default: 0

• dy
  

Derivative order in respect of Y (for deriv, sobel and scharr).
Type: Integer
Range: 0 - 2
Flags: readable, writable
Default: 0

• scale
  

Optional scale factor for the computed values (for deriv, laplacian, sobel and scharr).
Type: Integer
Range: 1 - 2147483647
Flags: readable, writable
Default: 1

• iterations
  

Number of times for the morphological operations to be applied (for morphology).
Type: Integer
Range: 1 - 2147483647
Flags: readable, writable
Default: 1

• morph-op
  

Type of morphological operation (for morphology). Available options:

erode: Normalized box filter

dilate: Maximum filter

open: Minimum filter

close: Vertical 1D box filter

Flags: readable, writable
Default: "erode"

• kernel
  

-2D matrix of NxM to use as kernel (for linear and morphology filters).
-2D matrix of 1xM to use as a row kernel (for separable filter).
Type: GstValueArray of GValues of type GstValueArray
Flags: readable, writable
Default: <<1.0,0.0,0.0>,<0.0,1.0,0.0>,<0.0,0.0,1.0>>

Examples

The following examples pipe will help you apply different types of filters to the input video stream from your camera. These pipes were tested on both, the Nvidia Jetson TX1 and TX2.

Border Detection

gst-launch-1.0 v4l2src ! nvvidconv ! opencvfilter filter=sobel dx=0 dy=1 size=1 ! queue ! nvvidconv ! autovideosink

Linear Filter with Custom Kernel

gst-launch-1.0 v4l2src ! nvvidconv ! opencvfilter filter=linear kernel="<<0.0,0.0, 0.0>,<0.0,0.5,0.0>,<0.0,0.0,0.0>>" ! queue ! nvvidconv ! autovideosink

Separable Filter with Custom Row and Column Kernels

gst-launch-1.0 v4l2src ! nvvidconv ! opencvfilter filter=separable kernel="<<0.0,0.5,0.0>>" kernel2="<<0.2,0.2,0.7,0.5>>" ! queue ! nvvidconv ! autovideosink