Products of COVISYS Computer Vision Systems GmbH

covisoft sdk

The covisoft sdk software development kit is written in C++ using both OpenMP and OpenCL technologies. It delivers a very intuitive interface based on simple C++ functions and guarantees optimized image processing and analysis exploiting the CPU and GPU parallelization capabilities present on actual personal computers. With covisoft sdk, COVISYS Computer Vision Systems GmbH follows the trend of using standard hardware in computer vision without any significant lost in processing performance compared to dedicated systems. The modularity of covisoft sdk simplifies the development of new applications and assures that software based on it, may be integrated in nearly every standard hardware. With this library, you will be able to write high efficient software without any OpenMP / OpenCL previous knowledge. The covisoft sdk set of C++ functions are developed to automatically deal with the OpenMP threading system or to internally code, configure and communicate with the OpenCL kernel architecture while ensuring that the developers do not have to be concerned with these processes.

Only a prefix makes the difference between function implementations.

  • - cpu... for C++ coded functions.
    COVISOFT_API(void) cpuGreySobelEdgeDetector(const byte* from, byte* to, uint width, uint height);
  • - mpu... for multithread / multicore implementation based on OpenMP.
    COVISOFT_API(void) mpuGreySobelEdgeDetector(const byte* from, byte* to, uint width, uint height);
  • - gpu... for multithread / gpu implementation based on OpenCL.
    COVISOFT_API(void) gpuGreySobelEdgeDetector(const byte* from, byte* to, uint width, uint height);

  • The covisoft sdk consists of numerous image processing and analysis functions that solve problems like noise reduction, convolution computation, image enhancement, anisotropic diffusion, alpha blending, feature extraction, transformations (distance, FFT, IFFT, DCT, DST, wavelets, Hough, radon, integral image, census, rank, LBP), color space conversions, bayer demosaicing, morphology operations, texture analysis, movement segmentation, binarization / thresholding, color and gray level segmentation, moments extraction, arithmetic, logical, geometrical and derivative operations, measurement of image quality and statistical parameters, artificial noise generation, noise estimation, generation of tv effects, similarity computation, calculation of correlation coefficients, stereo vision, overlay of geometrical primitives and fonts, interpolation, resampling, contour extraction, bounding box and convex hulls estimation, object detection, blob labeling, computation of blob geometrical features,...

    A list with all the operators could be found here.