Detection requirements:
Detect foreign matter (spots, hair, impurities) in bottle liquid.
The image signal of the detected target is obtained through image acquisition, and the image signal is converted into a digital signal through A/D and then transmitted to a special image processing system. The target features are obtained by calculation and analysis based on the distribution information of brightness, pixels and color. Finally, the result of the system operation is output according to the discrimination criteria previously set, and the corresponding peripheral mechanism is controlled and driven by the control system to perform certain processing.
The image acquisition, image preprocessing and foreign matter detection processes are implemented to simulate the entire production line in the figure. The online detection of foreign bodies in bottled liquids is realized, and unqualified products are automatically eliminated, which greatly improves the accuracy and speed of online detection in terms of detection methods.
Bubble interference：
Hair:
Foreign matter:
Solution description:
The foreign matter detection method for bottle liquid adopts the moving target detection method. We studied the foreign matter detection method under relatively regular motion. By quickly and accurately obtaining the real-time image information of the liquid in the bottle liquid, we can detect, analyze and make specific judgments on foreign matter in the bottle liquid to ensure the quality of the liquid.
If the posture of the moving target changes or the surrounding scene changes, these feature changes need to be modeled and detected. The dirtiness of the glass bottle and the bubbles in the dropper are the key issues detected by the detection system, which directly affect the effectiveness and accuracy of the entire detection system. Blank detection will affect the system’s judgment on the entire bottle liquid foreign matter detection status.
