Research On Fast Hybrid Focusing Algorithm In Zoom Camera

Autofocus technology is one of the key technologies of computer vision and various imaging systems.In the security monitoring industry,there are three significant problems caused by the use of high-magnification and intelligent high-definition network surveillance camera,slow focus speed,acquisition of object distance and measurement of high intelligent parameters.The traditional method uses a focus curve to speed up the focus speed,but it can not solve the other two new problems.In order to solve the above three problems at the same time,the reason of slow focus speed is explained by analyzing the structure of large magnification lens group.A hybrid fast focus algorithm based on focus surface and active focus and passive focus is formed.The focus surface is proposed to accelerate the focus speed in active focus.And a new passive focus is used to ensure the focus accuracy.This algorithm consists of focus surface and passive focus algorithm.The part of focus surface can also be called active autofocus.The active focus method is to drive the focus motor to reach the best focus position,which is calculated by the object distance measured by laser and the current magnification position.In this part,the structure of camera lens group is analyzed,the significance,measurement method and application method of focus surface are described,and other measurement methods of high intelligent parameters are given.Passive focus algorithm can also be called auto focus based on digital image processing.This part consists of focus window selection,clarity evaluation function and peak search algorithm.Due to the error of focus surface,it can not achieve accurate focus,so we use digital image processing to achieve accurate focus.Based on this,a passive focus algorithm is proposed,which is based on the strip window selection method,the improved Laplace clarity evaluation function and the twice local full search strategy.The proposed algorithm is experimented with a self-developed 33–fold 4K HD intelligent network surveillance camera,simulated by several pictures near the best position of focus in six different scenes.Using software to control the motor movement of the camera through the way of the upper computer controlling the lower computer.Experiments show that the algorithm has fast and accurate focus effect without noise or with low noise,while in the case of high noise at night,the focus effect takes a little longer,but it is faster than the traditional focus method.The algorithm has been successfully applied to the autofocus of self-developed 33-fold lens group and other high-magnification camera.