2D And 3D Geometric Measurement Methods Based On Image

The detection of image is an important method to obtain the two-dimensional or three-dimensional information of the object in the image,which is also an important part of industrial automation production.Traditional manual detection,due to its low efficiency and slow speed,has been difficult to meet the needs of industrial production and detection.With the development of science and technology,image detection technology get rapid development,because of its high detection efficiency,high accuracy and non-contact,improving the degree of industrial automation,it has received great attention in industry.This paper is based on machine vision and image processing technology,research on the high precision measurement method of circular and rectangular dimensions and the rapid detection method for the 3d shape of solder on PCB(Printed Circuit Board).This paper mainly research contents and work are as follows:(1)Proposed an improved Zernike moment subpixel edge detection method,changed the original algorithm's threshold condition and supressed the edge pixel bias,make the edge points are further refined and filtered,finally,based on the use of circular plate calibration pixel equivalent combine with the fitting of the least square method,measuring circular parameters.For the detection of rectangle,The paper mainly uses Canny operator and the improved Zernike moment subpixel edge detection method to detect the rectangle edge points,then,divides the detected image into blocks,calculates the number of edge points in every blocks,finally combines with the least square of linear fitting,gets the basic parameters of rectangle.After measuring the basic parameters of the circle and the rectangle,the cumulative error is compensated,and the influence of the camera exposure brightness,the camera and the measured object distance on the measurement results under natural light is discussed.The experimental results show that the improved method has strong environmental adaptability and the system is easy to build.In the case that the distance between the camera and the object is 50-60 cm,Circular measurement precision can reach 0.02 mm,the rectangular measurement precision can reach 0.035 mm.(2)Proposed an improved "Fast 2d phase-unwrapping algorithm based on sorting by reliability following a noncontinuous path" method to measure the 3d shape of solder on PCB.Based on the Miguel Arevallilo Herraez proposed "Fast 2d phase-unwrapping algorithm based on sorting by reliability following a noncontinuous path",The paperdesigns two masks and uses phase derivative variance map to replace the original second differences map,guidies phase-unwrapping to realize the accurate and stable measurement of the solder 3d shape on PCB.When the original algorithm measured the three-dimensional shape of the solder paste on the PCB,the results of the detection showed that the surrounding noise was large,the edge effect was strong,or the phase unwrapping was directly wrong,and no three-dimensional surface shape was detected.In order to improve the stability and accuracy of the detection,firstly,through iterative method,the original image is binarized to generate a binary edge template.Then,by the mathematical morphology extracting the edge noise of binary mask,producing a morphological mask,finally combine with the two templates and phase derivative variance map,limiting the area and isolate the outside noise of the solder paste to assist the method of phase-unwrapping based on sorting by reliability following a noncontinuous path,making the restored 3d shape is more stable and accurate.Experimental results show that the improved algorithm has less error and strong stability,the detection precision is about 12 um.