Research & Development
Computer Vision
Computer vision is the branch of computer science that emulates the human visual sense using optical cameras. Computer Vision has numerous applications, ranging from industrial quality control to security applications and precise contactless measurements.
Camera Calibration
During the camera calibration process, the intrinsic parameters, describing the geometry of the lens and its relation with the camera sensor, as well as the extrinsic parameters, relating multiple cameras, when available, are computed. These parameters allow correcting the geometric deformations induced by the camera lens, as well as enable carrying out measurements based on the images.
Original Image
Geometrically Rectified and Color-Corrected Image
Geometrically Rectified Image
Machine Learning
During the last years Machine Learning, and its most commonly known Deep Learning branch, have become a trend to solve multiple computer vision-related problems. The power of Deep Learning relies on the capability of learning based on (large) sets of labeled training data. Object recognition, segmentation, and classification are some of the most well-known tasks Deep Learning is tailored for.
Original Image
Color-Corrected Image
Segmented and Classified Fish
Image Preprocessing
Image preprocessing is a fundamental step when dealing with images, and becomes especially relevant in the case of underwater imagery, due to the additional challenges imposed by the medium. In this process, the visual but also the informative quality of the images is improved, to provide any further processing with the most accurate data possible.
Color Correction
Among the image preprocessing steps, color correction is the one where the images acquired underwater are modifyed to recover the original colors of the scene as if this had been taken on air. Girona Vision Research owns the tools and the knowledge to perform these steps as a part of their typical processing pipeline.
Original Image
Color-Corrected Image
Underwater Image Dehazing
Image dehazing is a technique used to improve the clarity of underwater, but also terrestrial, imaging, when the medium has a reduced visibility distance due to the presence of particules in suspensions.
Original Image
Image After Dehazing