An embedded module-based liquor segmentation detection device and method thereof

Abstract

The invention discloses a liquor segmentation detection device based on an embedded module. The device is composed of wine receiving equipment, industrial cameras, and NPU-based embedded modules. Using computer vision technology, it can quickly and stably judge the current liquor segmentation based on the hop image, and identify abnormal situations. The wine receiving equipment reduces the interference caused by water impact, ambient light and reflection through special structural design, side fill light, overflow, etc., strengthens the edge characteristics of hops, and avoids blur problems caused by insufficient depth of field of the camera; uses NPU-based The embedded module is used as a computing unit to run liquor segmentation and anomaly detection algorithms based on computer vision technology. Each step of the algorithm runs on different computing units and runs concurrently through the pipeline. Compared with large equipment such as workstations and servers, it has the advantages of small size, low cost, low power consumption, and easy installation and maintenance. The detection device can be connected to the wine picking system through the communication interface, so as to realize the automatic quantity and quality picking of wine.

Description

technical field [0001] The invention belongs to the technical field of brewing, and in particular relates to a liquor segmentation detection device based on an embedded module and a method thereof. Background technique [0002] During the distillation process of liquor brewing, due to the different volatility and boiling points of the components in the fermented grains, the order of distillation is also different. In the initial stage of distillation, aldehydes with low boiling point and volatile are distilled first; after that, lipids, alcohols and acids with low boiling point and more volatile are distilled next, and these substances are rich in aroma components , is the main part that constitutes the aroma components of liquor, and is also the part with the best wine quality; what is then distilled is higher boiling point and less volatile high-grade fatty acid esters, fusel oils, etc., showing a sour taste, excessively mixed into the previous stage Alcohol can easily af...

AI Technical Summary

Problems solved by technology

However, there are two problems in "watching flowers and picking wine": first, there are large differences in the judgment standards of each wine picker, and the alternation between the three-stage wine and the tail wine changes quickly, and manual flower picking can easily lead to unstable product quality; The second is to rely on the wine pickers to continuously observe the hops to make judgments, which wastes manpower and hinders the implementation of automation in liquor production

These two schemes directly measure flavor substances and alcohol content, have high reliability, and use online detection instruments, but have the disadvantages of high equipment cost and the fact that the measuring equipment will directly contact the wine liquid

The invention patent "A Wine Picking System for Industrial Production" (Application No. 201610979934.3) integrates various information such as temperature, pressure, spectrum, concentration, etc., and has higher reliability, but the relative cost is also higher, and a variety of equipment It is also difficult to maintain

This method has the following problems: the hop tank is open, which easily leads to splashing and volatilization of the wine liquid; the detailed description of the method is insufficient, including camera selection, wine receiving equipment design, microcomputer selection, etc.

This method of liquor segmentation has high precision and fast speed, but it also has some disadvantages: high equipment cost, high power consumption, and difficult installation and deployment; the coaxial placement of the fill light and the camera will cause reflections on the liquid surface; the wine bowl is open The wine is easy to splash and volatilize; the wine flows into the pipe from the hole at the bottom of the wine bowl, and the liquid level of the wine fluctuates greatly

Although this type of equipment has sufficient computing power and can run high-precision wine-picking algorithms based on computer vision, it has some disadvantages: 1) The cost of computing equipment is relatively high, such as the one selected in "Research on Automatic Wine-picking Methods Based on Image Classification Algorithms" The server ThinkSystem SR650 costs about 15,000 yuan; 2) The equipment is large in size and complicated to install, and needs to be placed in the central control room, and the industrial cameras on the production line are connected with network cables or optical fibers; 3) The power consumption of the equipment is high, which will bring additional electricity consumption

This method has high precision, but the preprocessing time is long, and abnormal conditions are not considered

Images