In order to increase the Quality of Monitoring (QoM) of Wireless Visual Sensor Networks (WVSNs), we revisit the barrier coverage problem with three factors, including importance of image, breadth of image and rotation capability. Without consideration of importance of image, camera sensors can capture images of the intruder crossing the barrier but cannot guarantee that the captured images are the important portion of the intruder. Without consideration of breadth of image, image identification may be difficult. If camera sensors have rotation capability, how to select effective camera sensors to reduce the number of camera sensors required for barrier construction is another important issue. In this paper, for WVSNs consisting of camera sensors with and without rotation capability, we will respectively propose an algorithm to find a barrier with encircled coverage capability and β breadth. The proposed algorithms ensure that if any intruder crosses the barrier, important portions of the intruder can be clearly captured. Finally, the success rate of the proposed algorithms in three camera sensor distribution settings, including Uniform distribution, Poisson distribution and Gaussian distribution, will be evaluated through simulations.