Study Of Inspection System For Bran Specks In Wheat Flour Based On Machine Vision

With the improvement of material living standards, people are demanding of improving the quality of wheat flour, so classification of wheat flour is much more refined. Processing degree of wheat flour is the main basis of evaluating wheat flour grade, it is denoted by wheat flour color and the size and distribution density of bran specks in wheat flour. Five methods(dry method, dry hot method, wet method, wet hot method and steaming method) are specified in current national standards to determine the processing degree of wheat flour, and these methods have something in common: the color of flour and dark bran specks in flour are evaluated based on human eyes, use control flour as a reference. These organoleptic test methods not only are time-consuming and strenuous, but also are subjective and not easy to be quantified.Along with the increasing maturity of machine vision technology, applying this technology has also made some achievements in terms of grain quality detection. Wheat bran specks instrument use objective examination to detect bran specks ratio that could assess wheat flour processing degree. Testing precision and speed are improved obviously by using this equipment. But the bran speck instrument needs to connect a PC to realize image preservation, processing, display, and other functions. There are always problems of high cost and poor portability in the use of PC-based system. In view of this, in this paper a micro intelligent wheat flour bran speck detection system is researched by using machine vision technology and the embedded system technology.This paper studies the composition of embedded machine vision system, proposes the scheme of micro intelligent wheat flour bran speck detection system including the light source, lens, camera, data interface, embedded processing platform, image processing software, graphic interactive interface. The design of hardware and software part in embedded processing platform is analyzed in detail. Hardware part includes power circuit, memory circuit, communication interface circuit, and software part has image acquisition methods and image processing algorithm in embedded system platform, and the system uses V4L2 that provides a video input interface to get wheat flour images captured by camera. Finally the system applies OpenCV library to pretreat the flour images such as graying, histogram equalization and binarizationan, extract characteristic of binary image, calculate bran specks area proportion, and adopts a new method based on connected domain to search bran specks number.Application software design of the entire system is completed in Qt programming software, and the system function is realized by the human-computer interaction using LCD touch screen. The system has the advantages of easy operation, stable image acquisition, detection speed and high precision. The system also has the characteristics of low cost, small size, low power consumption, high reliability, scalability, etc., and can be extended to other areas of machine vision applications.