Cocoa bean air separation method by using artificial intelligence detection, automatic foreign objects removal, and dynamic control

Abstract

This invention relates to a cocoa bean air separation method that combines artificial intelligence detection, automatic foreign object removal, and dynamic control. It is designed for air separation of pre-roasted and crushed cocoa beans into kernels and shells. The method's preparation step includes an intelligent processing system comprising an air separation module, an artificial intelligence detection module, an anomaly removal module, and a dynamic control module. The air separation module then preliminarily filters out most of the cocoa shells (preliminary screening step). Next, the artificial intelligence detection module performs multi-angle image capture and spectral data acquisition (detection and classification step). Finally, the dynamic control module, based on the returned data, drives the anomaly removal module in real-time to remove foreign objects (foreign object removal step), and dynamically adjusts the air separation module (advanced air separation step), thereby achieving high-precision, adaptive, and automated separation.

Claims

1. A method for cocoa bean air separation combining artificial intelligence detection, automatic foreign object removal, and dynamic control, wherein the method performs air separation processing on roasted cocoa beans, wherein the cocoa beans are pre-crushed into a mixture containing cocoa kernels and cocoa shells, the method comprising: a preparation step, which includes an intelligent processing system having an air separation module, an artificial intelligence detection module connected to the air separation module, an anomaly removal module connected to the artificial intelligence detection module, and a dynamic control module for controlling the air separation module, the artificial intelligence detection module, and the anomaly removal module, wherein... The AI ​​detection module has at least one high-speed camera and at least one spectral sensor, while the anomaly removal module has a foreign matter removal component and a collection tank connected to the foreign matter removal component; a preliminary screening step, in which the air separation module uses airflow to preliminarily sieve away most of the cocoa shells in the mixture, thereby obtaining a preliminary screened product with most of the cocoa shells removed; a detection and classification step, in which the high-speed camera captures multi-angle images of the preliminary screened product as it passes through the AI ​​detection module, and the spectral sensor collects spectral data of the preliminary screened product to achieve real-time and comprehensive detection, and classifies the particles and impurities contained in the preliminary screened product, while the images captured by the high-speed camera and the spectral data collected by the spectral sensor are transmitted to the dynamic control module; A foreign object removal step: when the dynamic control module determines, based on the received image and spectral data, that the preliminary screened material contains inferior cocoa kernels or impurities, the dynamic control module drives the foreign object removal component to separate the inferior cocoa kernels or impurities from the preliminary screened material and introduce them into the collection tank for collection; and an advanced air separation step: when the dynamic control module determines, based on the received image and spectral data, that the preliminary screened material contains too many cocoa shells or that the proportion of mixed impurities exceeds a preset threshold, the dynamic control module further adjusts the operating data of the air separation module to remove the cocoa shells and mixed impurities contained in the preliminary screened material in real time.

2. According to the cocoa bean air separation method combining artificial intelligence detection, automatic foreign object removal and dynamic control as described in Request 1, an optimization step is added after the advanced air separation step. The optimization step records the batch data of the cocoa beans, the image and spectral data of the preliminary screening material, the corresponding operating data of the air separation module, the data of the substances removed by the abnormal rejection module, and the final separation quality detection data through the aforementioned data, and trains and optimizes the dynamic control module through the aforementioned data.

Images