Industry News
    Home > News
  • Smart civilians - HD front-end development new trend

        Today the development of high-definition surveillance system , not only in high-definition video on a single scale is constantly expanding , such as hundreds of thousands of ordinary peace of the city roads, thousands of road even more, traditional HD cameras only used to collect , use Real-time monitoring is often difficult to effectively "care" to all the video , has been difficult to meet people , " good use " HD needs.
        HD IP camera is advancing rapidly, regardless of image quality, performance, features , price, replacement faster and faster, basically follow Moore´s Law , but eventually, all of the innovations are needed to enhance the focus on the needs of people , if the camera own image , sound, behavior can make intelligent judgments filter , both integrated platform for intelligent analysis of box functionality and avoid bottlenecks little ones , at the same time cost effective control smarter, see clearly , they also make the monitoring system becomes passive to active. In HD product homogeneity serious today , what technologies can be more real savings overall solution cost ? What can be more practical intelligence into clever front for the entire program with highlights ? Below these two areas as we expand the analysis:
        Cost savings technique
        With the continuous development of coding technology , and constantly improve the performance of the hardware platform , the low stream , to ensure quality , the bandwidth and storage of the original 1080P real- 6M ~ 8M, down to 2M ~ 4M, only the original half or even a quarter , reducing the overall cost of high-definition programs from the roots. Currently low stream has become standard on most manufacturers .
        Scalable video coding (SVC), reduces network jitter. Without compromising image quality , low weight dropped frames , stronger anti- packet loss , better network resilience. At the same time , PC can request a full frame rate of the main stream , the handheld device can request to the main stream of non- real time ; fast playback end time of storage , skipping decoding can reduce the pressure of the network transmission and decoding ; the same way streaming, real-time monitoring , non- real-time storage , saving network and storage costs.
        Region of interest (ROI), the designated area to improve the video quality. Under the same bandwidth , the limited resources on stream in the area of interest to the user. Ensuring low bandwidth , users are interested in areas of the image to be assured maximum support four regions simultaneously setting. Effectively alleviate the pressure on the network and storage .
        Corridor model to meet the needs of application scenarios in depth . Corridor particular scene mode , one device can achieve the effect of two cameras , a typical application scenarios such as a long corridor , roads and so on. Also as three yards stream 20 simultaneous access , ultra-low latency , are the hardware platform performance and ISP handling skills have higher requirements.
        HD IP camera to bring better visual experience , while the installation and commissioning personnel also have higher requirements. After the automatic focusing function , clear buttons and a key to integration via remote network control , can greatly reduce commissioning and ongoing maintenance costs.
        Practical and efficient intelligence analysis
        Fog : The current video surveillance system is very sensitive to weather conditions , when faced with low visibility fog , haze , etc., severely degraded scene image , which greatly limits the play outdoor video surveillance system utility . The current global climate is worsening , fog and haze often appears , Fog significant particularly important , according to the fog imaging scattering model , using dark channel priori principle, can achieve high performance through fog technology, weather monitoring system work .
        Through the fog algorithm characteristics
        1 , the atmosphere is more stable light intensity using statistical methods , making the high stability of the algorithm , for a variety of scenarios.
        2, the color reproduction is good: the traditional algorithm through the fog would dilute the color image , the algorithm restore the true color images .
        3 , good permeability : the algorithm to accurately estimate the fog density images , with good permeability.
        4 , consider the scene depth information to a great extent restored image detail: the algorithm to accurately estimate the depth of objects of different concentrations of mist , accurate through fog, mist that had previously been reserved and highlight hidden details ; conventional algorithms may close range through the fog over the good, but the vision will be left fog , when enhanced through the fog level to achieve the vision through fog , close-up detail will be lost .
        5 , the traditional method for improving image contrast , there will be cases of local overexposed or too dark , damage the image quality ; This algorithm without losing detail.
        6 , intelligent automatic through the fog : You can judge for yourself whether it is foggy weather , choose to open or close the Fog .
        Smart civilians - HD front-end development of a new trend of the future
        Technological breakthroughs :
        1 , the algorithm through the fog stronger , good reduction , compared to a single image to the fog , the algorithm complexity is very high , for this situation , the need for a series of improvements optimized to achieve real-time through the fog .
        2 , taking into account the correlation between the frames through improved control of the algorithm , to achieve a smooth transition to avoid flicker.
        3 , the development of intelligent automatic functions through the fog , the algorithm can judge for themselves whether the foggy weather , select the function on or off through the fog , which greatly facilitates users.
        Camera mounted on a roadside stand in marked encountered heavy vehicles through or windy weather , inevitable image dithering , this time video image stabilization technology can be put to great use . By estimating a motion vector of the current frame , and the estimated motion vector for motion compensation for the current frame , thereby achieving stable image sequence object . Can effectively reduce or even eliminate the image jitter.
        Face detection , face when someone appears in the monitored area , it will automatically capture the clearest pictures of facial and make optimization enhancements for query and forensics. Another virtual focus , scene changes , occlusion , etc. can be the first time an alarm message , avoid useless video.
        Sound anomaly detection . Detect such silent , no audio input , you can mute the alarm ; environmental noise filter to reduce noise input in public places ; audio mutation detection, such as screaming, sudden big noise decibel alarm can be generated . Audio requirements applicable to the place .
        Universal behavior analysis, such as tripwire invasion, climb fences , regional invasion, items left behind / disappears, illegal parking , fast running , and so on , all the smart features of the camera settings are completed .