Research On Chip Component Detection System In Mounting Machine

In this paper, vision-based chip components detection technology inhigh-precision placement machine was researched deeply. Chip components are mostcommon used component in surface mounting technology productions, and they havesimple structure and other characteristics, these characteristics let them well suited tostudy the principles and detect the speed, efficiency of SMT mounting machine. Mainfunction of the component detection system of mounting machine is to obtain the poseof chip (pose of chip include center of the chip relative to the system coordinatesystem and the rotation angle of the chip), which can make up error in nozzle absorbchip components, and, therefore, directly affects the placement precision of chipcomponents of mounting machine. Therefore, the study on chip components detectionsystem of mounting machine has a very important significance.In this paper, three main aspects of chip components detection system of themounting machine were studied, they were as follows: links of the coarse positioningcomponents, links of identification of chip and link of precise positioning components.Described as follows:(1) Based on the characteristics of chip detection system of mounting machine, apractical components coarse positioning algorithms is proposed. Differing fromtraditional chip positioning algorithms, according to the fixed positional relationshipbetween the nozzle and chip components, this paper proposes a method of roughlylocation algorithms of chip components by positioning the nozzle. Experimentalresults show that the proposed algorithm can accurately detect the general location ofthe chip, but in the case of a lot of noise in the background, the interfering substancesin background may be mistaken as the nozzle tip.(2) For the error detection problem appeared in session one, a recognitionalgorithm of chip components is proposed. First of all, to eliminate the effect to chipimage recognition caused by interference point, a region filtering algorithm based onthe maximum connectivity is proposed. Then, three algorithms of chip detection areanalyzed and compared: squareness based component recognition algorithm,degree-of-polymerization-zone based component recognition algorithm andmoment-invariants based component recognition algorithm. By derivation andanalysis, we found that the degree of algorithms based on rectangular for rectangularchip and other shapes identified regions, based on regional density algorithm foridentifying different aspect ratio rectangle, based on moment invariants algorithmused to identify these two kinds of situations.(3) Finally, the article describes the elements precise positioning link. Thissession combines the edge tracking algorithm and the minimum bounding rectanglepositioning algorithm, there is no interference with the point on the binary image, withvery high precision. Experimental results show that the proposed chip components can accuratelydetecting element detects the center coordinates and rotation angle, and with highprecision, to achieve the design requirements.