The Research On Resolution Detecting Methods Of Camera Module

After the camera module finishes assembling different components, the distance between the camera lens and the image sensor is random, and often out of focus. This leads to failure of the object-image conjugate, which further causes blurry photos. Under the condition, it is necessary to adjust the focus. Resolution is an index to evaluate the camera module imaging quality, and a key element of quality assurance. In the industrial production line, the majority of small and medium-sized module manufacturers adopt manual focusing and visual evaluation of resolution, the focusing and the resolution test are usually synchronous. The operation process is simple, but it has defects. First, manual focusing usually results in poor precision, huge product differentiation and low productivity. Second, resolution test is faced with problems such as subjective difference and quick visual weariness of workers. The purpose of this research is to realize automatic focusing of the camera module and automatic test of resolution,simplify the module production process, and improve productivity.This paper adopts ISO 12233 as the theoretical basis, and 200W CCD Camera Module as the research object. An automatic focusing system based on resolution test and combining demands of module manufacturers is designed. This design consists of the equivalent visual resolution test target plate design, the focusing search algorithm design and the focusing institution model design. Through the design, automatic resolution test of the camera module is realized. Below is the specific research content of this paper:First, the automatic focusing technique based on image processing usually adopts the sharpness evaluate function as the focusing criterion. The hill climbing algorithm is employed to search the position with the highest image sharpness. However, the hill climbing algorithm usually comes up with the local optimum. Since the visual resolution value and the sharpness evaluate value show similar curve characteristics, and the visual resolution value of the camera module on the focus plane is fixed and can be acquired in advance. This paper thus proposes a theoretical viable search strategy, in which focusing is assisted by visual resolution value. To be specific, when the focusing mechanism stops and thinks it has achieved the right focus, the visual resolution value is introduced to judge whether the point overlaps with the sharp focus position. Since the sharpness evaluatie function is not universally applicable, its focusing performance varies in different scenarios. Considering the problem, this paper realizes 14 kinds of sharpness evaluation functions through simulation, and studies the focusing performance of different functions in the system by evaluating unimodality and unbiasedness of the curve,computing time, and robustness. It is found out that the Brenner function shows the highest applicability.Secondly, based on the imaging characteristics of the hyperbolic wedge line, a software which can automatically judge the visual resolution of camera module is developed. An experiment is designed to verify favorable consistency between the software test results and the workers'visual inspection test results,proved that the software can be applied to quality assurance of the production line. Firstly, the software finds the position of the hyperbolic wedge in the image by the target positioning technique. Secondly, through looking for the starting line(WSL), the limited distinguish line (LML) and the final distinguish line(WEL) of the hyperbolic wedge, and based on the relationship among the three, the visual resolution value is calculated and display.Finally, the principle and the realization process of e-SFR algorithm from ISO 12233 used to test spatial frequency response of the camera module are preliminarily studied. Factors influencing accuracy and stability of the algorithm are deeply analyzed from four aspects,including the contrast ratio, angle of inclination, aspect ratio and peak signal to noise ratio of the region of interest (ROI). The practical data are adopted to analyze and verify evaluation effects of the e-SFR algorithm and the wedge line algorithm. Results show that, under the condition of weak sharpening, the two algorithms can achieve consistent evaluation results.