Detection of Nano-particles Based on Machine Vision

Information extraction from images is a key problem. In this paper, a fast, high-precision, multi-target image processing method based on machine vision is proposed to detect nanoparticles. The proposed methods include the following steps: image pre-processing, feature extraction, and result output. An auxiliary-Calibration plate is used as the auxiliary need to determine the measurement accuracy. Through the analysis and calculation of the characteristic scales of nanoparticles, our results are highly consistent with the experiment, indicating the accuracy of our method and algorithm. From the input test images to the output results, the algorithm only takes less than a second. For the image including more complexed structures, it can also be accurately detected. Furthermore, the algorithm is transferable and extensible, and can be applied in wider image processing fields.