Research On Image Key Technology Of Digital Stereo Light Microscope

Traditional measurement methods mostly are contact measurement, it will casue abrasion to tiny objects. However, it’s not necessary to contact with objects by utilizing microscopic visual measurement methods. It only needs to process captured images of tiny objects to complete measurement. Therefore, it protects tiny objects effectively. Because of the limitation of depth of digital stereo light microscope, low level manufacturing technology and micro-operation, traditional macroscopical binocular three-dimensional reconstruction methods are hard to perform effectively in digital stereo light microscope. Image information consists of primary information and secondary information. It can filter secondary information and save a lot of reconstruction time if the primary information were extracted automatically. Moreover, it is beneficial to subsequent objects’ analysis and recognition. Traditional autofocusing methods need capture a large number of images to find the clearest image as target image or adjust focus by combing hands with eyes to capture microscopic images. It turns out that it need spend a lot of time each time.The paper mainly researches on following three problems. Firstly, due to camera parameter calibration process is difficult and cumbersome in stereo microscopic image reconstruction and small deviation of actual stereo microscopic parallel optic path causes the vertical displacement in camera’s imaging of the same object, a three-dimensional reconstruction method for tiny object in digital stereo light microscope is proposed in this paper, which regards stereo microscopic system as a whole. The proposed method uses simple three-dimensional reconstruction model to recover three-dimensional coordinates of test object by considering the characteristics of parallel binocular imaging principle and only need calibrate two three-dimension reconstruction parameters. Experimental results demonstrate that for the proposed method, its measurement error in the directions X and Y is about 3% and while one in the direction Z is 3.2%-9.1%, which indicates that the proposed method can be used to measure three-dimensional size of micro object within tolerance error.Secondly, Common image segmentation methods only consider the single image and cannot extract welded components automatically and exactly. GrabCut algorithm need user to draw a rectangle which labels possible foreground region and determines general location to segment desired object. In this paper, a novel automatic GrabCut algorithm is proposed for Printed Circuit Board(PCB) welding component under stereo optical microscope. The proposed method considers the characteristic of welded component locating on the foreground region of PCB and uses unparallel stereo microscopic image to obtain foreground mask of disparity map of microscopic image. Rectified microscopic image is segmented by initializing GrabCut’s trimap T with foreground mask of disparity map of microscopic image. The experimental results show that the proposed automatic Grabcut method can extract PCB welding component automatically and accurately without user intervention.Finally, Common autofocusing methods only use monocular image information and there is not fixed index to determine sharpness for arbitrary image. Thus, it is necessary to capture a series of images and spend a lot of time. For these questions, according to binocular images information of stereo light microscope and linear relationship between disparity of stereo microscopic image and height of motorized translation stage, the paper proposes a disparity servoing-based fast autofocusing approach for stereo light microscope, which utilizes disparity to measure image sharpness. The experimental results show that the proposed method only needs a small number of iterations and little time to reach focus position, and the disparity error is less than 0.5 pixel.