Coffee Quality Analyzer

1.application areas: processing plant, storage warehouse, quality inspection center;

2. Detection scope: raw sprouts, dried beans, grass seeds, worm-etched beans, red mold beans, damaged beans, spotted beans, impurities, etc;

3. Key technologies:

1) Automatic binarization: The use of deep neural network to segment the foreground background of the image, compared with the traditional two-value method, can be applied to a variety of lighting conditions, and.coffee bean edge segmentation is smoother, with the advantages of fast speed and high robustness.

2) Adhesion coffee bean segmentation algorithm: The method based on the connected domain cannot divide the adhesion.coffee beans, the use of deep neural networks to the adhesion of coffee beans instance-level segmentation, can achieve.1000fpsspeed, the adhesion of coffee beans can be processed in real time.

3) coffee bean attribute recognition algorithm: Using lightweight neural networks and integrating semi-supervised learning methods, the model can be iteratively optimized with only a small amount of data labeled by the user, with the advantages of high precision, high speed and convenient deployment.