About BoofCV
BoofCV is an open source Java library for real time computer vision, written from scratch and released under the Apache 2.0 license for both academic and commercial use. It covers low level image processing, camera calibration, feature detection and tracking, structure from motion, fiducial detection and recognition, organized into packages for image processing, features, geometric vision, calibration, recognition, visualization and input and output.
It is aimed at developers and researchers who need dependable vision algorithms inside their own software rather than a hosted service. The capability list is broad: convolution, derivatives, thresholding, color spaces, image pyramids, Fourier and wavelet transforms, fisheye and equirectangular camera models, line and interest point detection, motion detection, contour and polygon fitting, template matching, and QR and Micro QR code reading. The project also ships prebuilt standalone applications, including a camera calibration tool, calibration target and square fiducial generators, a QR code generator, batch lens undistortion and batch image downsizing, along with browser based tools for creating calibration targets and scanning barcodes.
BoofCV is free to download and use. There is no paid tier, and the project documents a bounty program and contribution guide for anyone who wants to fund or add features.
The current release is v1.2.3, dated July 1, 2025. Documentation includes a manual, JavaDoc, a FAQ, examples, performance and validation reports, and a road map, and a companion C++ port called BoofCPP is also available. Although it is listed among AI tools, BoofCV is mainly classical computer vision rather than a machine learning framework.