The Microscopic Automatic Analysis System Of Inclusions Based On OpenCV

Inclusion analysis is an important method to study geology with important guiding significance for the diagenesis and mineralization, resources exploration and the analysis of ancient fluid composition,majorly including inclusions observation and inclusion temperature measurement.The principle of most existing popular inclusion analysis method:due to molecular thermal effects, inclusions with free motion in the sample of rock slice could be observed using high-power microscope, and then heated up the sample artificially until the inclusions disappeared, recording this point temperature called homogeneous temperature points. Cons:the human eye is not sensitive to the movement of small objects; because of the large proportion of microscope, single observation area is limited, then manually moving the stage is needed for observing panoramic area; the geologial information of slice cannot be saved and analyzed integrallty.This paper proposed improvement aiming at inclusions naked-eye observation and limited observation area, and designed the system of inclusions automatically recognition and panoramic mosaic for the microscope images of rock slice sample by applying image processing technology.Using Intel’s OpenCV library as a basic algorithm support for image processing, implement motion detection algorithm for high-definition video microscopic images obtained by CCD camera to recognize movement of inclusions.And with micro-electric loading platform, automatically controlled by computer, inclusions identification and recording in the whole rock sample sheet could be realized automatically. Using image splicing technology, during the process of recognizing inclusions, stitching the microscope image of rock slice sample at the same time, the synthesis of rock slice from the a complete panorama, save all the geological information of the wafer, convenient later viewing and sharing.This paper detailed the basic principles of motion detection algorithm, and proposed improvement due to motion characteristics of inclusions.The specific process included image preprocessing, grayscale conversion, background subtraction operation, etc., setting a threshold for the differential results binarization transformation, and then the binary image edge detection and contour drawing to mark the location of inclusions. For the mosaic of rock slice image captured by high-power microscope, this paper used the automatic panoramic stitching feature-based algorithm. The algorithm took into account the microscopic image acquisition process and the microscopic characteristics of the image itself, using the Scale Invariant Feature Transform (SIFT) to extract the basic characteristics of the image, and then the priority (BBF) image feature extraction algorithm based on the potential optimal node optimization matching pairs,Using Random sampling consensus algorithm (RANSAC) to purify the early matches, and calculating the transformation matrix for the movement between images using affine transformation model, completing image registration and fusion. Both inclusions Motion detection algorithm in rock slice sample and stitching algorithms were programmed by using MFC and calling the OpenCV library, to reduce programming effort and improve development efficiency and reliability of the program running.At the same time,this paper designed a miniature motorized stage driven by stepping motor instead of manual operation,and programed the driving code to realize digital control. The application running at computer controlled the stage via RS232 interface, and completed panorama stitching work for sheet rock samples and inclusions automatical recognition.Through measurement,the inclusions automatical analysis system designed by this paper could realize continuous automatic recognition of inclusions under a single average recognition time horizon of about 30S;At the same time during the process of inclusions automatic recognition, the stiching work for microscopic image sequences from samples of rock slice was realized,with ideal stiching effect,no obvious visual traces,keeping the details of the image well.