Surveillance video technology has gone through the age of simulation and entered the digital age. In the process of digital monitoring, it is currently in the process of development from digitization, networking, and high-definition to intelligence. Among them, intelligent surveillance video analysis is one of the current key development directions. People hope to replace a large number of laborious manual observations, analysis, and summary of surveillance video content through intelligent analysis.
Intelligentization involves all kinds of processing for monitoring video images. From the technical level, it can be roughly divided into three levels: video image processing focusing on pixels or blocks, video image analysis focusing on image features and targets, and focusing on images. Intelligent analysis of content semantics. In popular terms, the goal of monitoring video image processing is to â€œseeâ€ the monitored scenes, including measures such as de-noising, de-blurring, de-jittering, de-sharpening, super-resolution reconstruction, and dynamic range expansion; monitoring of video image analysis The goal is to â€œseeâ€ the monitored objects, including license plate recognition, face recognition, and flow statistics; the goal of intelligent monitoring video analysis is to â€œunderstandâ€ the semantics of the monitoring scene content, ie, to give people the ability to understand Analyze the results, such as what happened here, what is there, etc. It contains a variety of intelligent analysis technologies, such as behavior identification, intrusion detection, legacy detection, group behavior identification, and illegal vehicle detection.
Of these three levels, it is clear that intelligent surveillance video analysis is our ultimate goal. It replaces the observation, analysis, and judgment of the content of video surveillance by humans. It is even more accurate and faster than humans in some aspects. However, to achieve this goal, we must provide as clear as possible video images to the intelligent analysis system. As far as possible accurate target division and recognition results, we call these technologies the basic technologies of intelligent video analysis, namely video image processing. And analysis. If the provided images do not meet such requirements, it is difficult to expect that the intelligent analysis system will obtain correct analysis results. Therefore, we say that intelligent video analysis is important, and the specific method of intelligent analysis is important, but the basic technology for providing good video images is particularly important in the initial stage of video analytics.
At present, from an application point of view, although smart video analysis has made great progress, there is still too much content to be researched and developed. Intelligent video analysis technology is still in the "initial stage" of video processing and video analysis. This part is the moment. Focus on practical applications. The basic monitoring video processing and analysis technologies include many aspects. In addition to the classic video processing technologies such as image enhancement and image denoising, the main concerns of everyone are image defogging, dynamic range expansion, and super resolution reconstruction. License plate recognition, face recognition, ..., and many other technologies. The following is an example of the video analysis instrument system, VAIS (Video Analysis Instrument System), which is well-known in the video surveillance industry. It focuses on the basic processing technologies of video image processing and video image analysis.
Video image processing
2.1 Fog antireflection treatment
Due to the effect of atmospheric scattering, the video images obtained under foggy weather conditions are relatively vague, which seriously affects the visual effects of the images. The main reason is that the target light is attenuated by the fog during the process of propagation, which results in the loss of image details and insufficiency of clarity. The participation of ambient light causes distortion of the image color and the tone is dull.
In order to improve such foggy images, video fog-refining technology (fog technology) can be used to make clear images due to fog, dust, etc., and to discover more information contained in images. Provides good conditions for the next step in the intelligent analysis of images.
The current methods for the penetration of fog are mainly divided into two types: fog image enhancement and fog image restoration. Foggy day image enhancement method is relatively simple, it does not consider the reasons for image degradation, only for image color processing, can effectively improve the contrast of fog images, highlight the details of the image, improve the visual effect of the image, but may Cause a certain amount of information loss.
Foggy image restoration aims at the mechanism of fog image degradation, establishes a fog image degradation model, and then uses image restoration to recover fog-degraded images to compensate for the distortion caused by the degenerative process and obtain an optimal estimate of fog-free images. , thus improving fog image quality. This method has strong pertinence, the natural defogging effect is obtained, and the loss of information is small. The key to processing is the estimation of parameters in the model.
2.2 Super-resolution reconstruction
As people's requirements for monitoring image quality become higher and higher, increasing the resolution of surveillance images has become an urgent requirement for the entire security industry. The most straightforward way to increase the image resolution is to increase the sensor density of the acquisition device (such as a camera). However, on the one hand, high-density image sensors are relatively expensive and difficult to handle in general applications. On the other hand, imaging systems are limited by the density of their sensor arrays.
In such a situation, an effective way to solve this problem is to improve the spatial resolution of the image by using a signal processing-based software method, namely Super-Resolution (SR) image reconstruction, which will be acquired. Low-resolution images are reconstructed into high-resolution images by certain algorithms. The core method of VAIS SR reconstruction is to use time bandwidth (acquire multi-frame image sequences of the same scene) in exchange for spatial resolution, and realize the conversion of temporal resolution to spatial resolution, so that the visual effect of the reconstructed image exceeds any one frame of low resolution. image.
2.3 Dynamic Range Expansion
In many video surveillance applications, images with color (brightness) dynamic range deviations often occur due to light, and overall or partial â€œexposureâ€ deficiency or â€œexposureâ€ is excessive. The processing of such video images is mainly to correct the deviation of the color dynamic range, including two aspects: the processing of the backlight image and the processing of the dark light (such as night scene) images.
(1) Processing of dark light images
Due to the lack of illumination intensity, the images collected in the night scene will cause the image brightness and contrast to decrease, lose details such as color and increase the noise of the image, thus reducing the image quality, and seriously affecting the use of the image and further intelligent analysis and processing. The currently popular method of dim image processing is the Retinex algorithm, which separates the luminance image and the reflection image from a given image. Under the condition of constant color, the enhancement is achieved by changing the ratio of the luminance image and the reflection image in the original image. The purpose of the dark light image. However, this method is very computationally expensive and difficult to apply to real-time processing. Therefore, after statistical analysis, the VAIS system finds that the inverted dark light image and the fog image have similar statistical characteristics. Based on this, a new fast algorithm for expanding the image dynamic range based on the image defogging method is designed. The reversed dark light image is processed, and the process is completed and then flipped to become an image with an expanded dynamic range.
(2) Backlight image processing
In the video capture process, if the camera is facing the light source, this part is very bright, and the target object in the scene is much darker than the light source. Due to the limitation of the dynamic range of the camera, the brightness of the target area is very dark, and details cannot be seen clearly. This is the so-called backlight phenomenon. Obviously, the key to backlight image compensation processing is to increase the brightness of the target area, reduce the brightness of the light source part, and at the same time ensure the smooth transition of the entire image. The VAIS system uses a method similar to dim image processing for backlight images, that is, a method based on image de-fogging to process the backlight image.
Third, video image analysis
3.1 Face Recognition
Face recognition technology based on image processing generally can be divided into two categories. One is cooperative face recognition, which is a common identification method for attendance, access control, conference signing and other occasions. At present, this type of technology has been compared. Mature; the other is non-cooperative face detection and recognition technology, the most typical is the face detection and recognition of the target person in the video surveillance, this type of technology is still not perfect, there are still many problems to be solved.
At present, in the application of surveillance video, most of them are non-cooperative face detection and recognition. Due to the inability to obtain the cooperation of target personnel and the influence of various adverse environmental factors, the accuracy of face images in video is not high. Faint, blurred faces, and large angles make it difficult to do high-precision face detection and face recognition.
The VASI system focuses on the non-cooperative face recognition technology and is dedicated to the optimization of algorithms in face recognition of small sample banks. Therefore, it has higher face detection and recognition rate. The VAIS system transmits all face images captured and detected in the video scene to the face recognition section of the VAIS. The WSR and LCR algorithms are used to reconstruct the blurred faces, and then they are compared with the face data of the VAIS system database. In the comparison, the closest one or more face data is selected for the user to confirm the target identity, or play a role of prompting and warning for the user, and remind the user to perform further identity verification. This kind of face recognition method is very useful in many occasions, such as preliminary identification and prompting of VIP or blacklist users at hotels, banks, airports, terminals, customs offices, etc.
3.2 License Plate Recognition
The license plate recognition technology mainly includes two types, one is a relatively stationary bayonet license plate recognition, and the other is a license plate recognition of a vehicle in a moving state.
For license plate recognition of parking lot type, due to the static or slow speed of the vehicle, coupled with the detection of the vehicle, it is relatively easy and the recognition rate is also high. The VAIS system recognizes such license plates. As long as the width of the license plate image captured by the camera reaches 30 pixels or more, the license plate number can be accurately identified and the accuracy rate is almost 100%.
For license plate recognition of vehicles in motion, such as on-road vehicles and even escaped vehicles, due to the existence of unfavorable factors such as complicated environment, high-speed motion, and occlusion of other vehicles, the technical difficulty of license plate segmentation and vehicle number identification is greatly increased. . In this case, the VAIS system uses the correlation and constraints existing between the targets in the monitoring scene to identify most of the fuzzy license plate numbers with a relatively high accuracy. At present, VAIS has successfully assisted many departments in quickly identifying the difficult cases of different degrees from the license plate recognition. In the future, with the continuous improvement of the VAIS fuzzy license plate recognition technology, the accuracy of license plate recognition will continue to increase.
FIG. 5 is an example of automatic identification of a license plate of a suspicious vehicle in a residential area. After passing the VAIS system to segment the blurred license plate, the automatic identification number is 624M4 after the operation identification algorithm is processed. Figure 6 shows an example of the automatic recognition of the license plate of a fast moving vehicle on the highway. The identified license plate number is 422AF.
In the process of rapid development of intelligent surveillance video analysis technology, when we encounter difficulties or errors, the source of the problem is often the quality of the images provided to the intelligent analysis system itself. For example, if the images are not clear, the target classification is inaccurate. Image content is dim and so on. Therefore, strengthening video image processing and preliminary analysis is one of the most critical and necessary conditions for obtaining accurate results of intelligent video analysis.
Above, we briefly introduced several important video image processing and analysis techniques, including image fogging, dark light image and backlight image equalization, face detection and recognition technology, license plate detection and recognition technology, etc. It is a basic function of the VAIS intelligent video analysis system and has achieved good results in practical applications. In addition to the above-mentioned several special video processing methods, there are many other pre-processing methods, such as the processing of the region of interest, anti-shake processing, dynamic target locking processing, tracking and escalation of the target.
Only by solidifying the foundation of video processing and analysis can it be possible to carry out intelligent video analysis on this basis. VAIS is working hard in this direction. Combining massive video data (big data) mining with artificial intelligence machines Learning methods, especially the deep learning method of artificial neural network, are improving the speed and accuracy of intelligent surveillance video analysis step by step, step by step to practical use.
Text / Guangdong Xuntong Technology Co., Ltd. Zhu Xiuchang
Grab Bucket Description
It is widely used to catch scrap metal,waste iron, ores, pig iron,sand,rubbish etc,and can be equipped on various cranes like Overhead Crane,gantry cranes,portal cranes
Mechanical Grab Bucket (Available in One, Two and Four Rope execution)
Scissor type grab bucket for bulk carrier loading and unloading:
ï‚·Suitable for port, storage yard application
ï‚·Design for coal, Mineral powder, Bulk chemical fertilizer handling
ï‚·High grab proportion can reach 1:1.5
Grab for wood handling
ï‚·Suitable for port, forestry centre, lumber storage yard application
ï‚·Design for wood handling
Manual operate one rope grab bucket
ï‚·Suitable for port, storage yard application
ï‚·Design coal, Mineral powder, Bulk chemical fertilizer handling
ï‚·Simple structure, unique connect structure design, easy and Convenient operation, high precision movement.
Grab bucket for Marine salvage
ï‚·Design for salvaging huge sunken ship under water
ï‚·Could separate the ship very easy under water and can finish the work with a high efficiency.
Marine salvage Orange Peel Grab Bucket
ï‚·Design for salvaging huge sunken ship under water
Could separate the ship very easy under water and can finish the work with a high efficiency.
Grab bucket for bulk material rope type
Suitable for port, hydropower plant, metallurgical , construction, and storage yard application
Design for coal, Mineral powder, Bulk chemical fertilizer, or other bulk material handling.
Leakproof rope type grab bucket
Suitable for port, hydropower plant, metallurgical, and storage yard application
Orange peel bulk material grab
ï‚·Suitable for port, hydropower plant, mining, construction, waste dump metallurgical, and storage yard application.
ï‚·Design for Pig iron, ore, slag or other bulk material handling.
Hydraulic Grab Bucket (Electro Hydraulic / Motorized)
Electric-hydraulic clamshell grab bucket
ï‚·Grab with hydraulic system, and all hydraulic unit adapt famous high quality brand in the world to make sure the high performance.
Electric Hydraulic orange peel grab bucket
ï‚·Grab with hydraulic system, and all hydraulic unit adapt famous high quality brand in the world to make sure the high performance.Every tine have a independent oil cylinder driven to make sure the material sift-proof.
Electric hydraulic rectangle type grab bucket
ï‚·Grab with hydraulic system, and all hydraulic unit adapt famous high quality brand in the world.
ï‚·Design for cleaning Garbage and debris in Reservoir, river, boxcar train.
Excavator Grab bucket
ï‚·Adapt the hydraulic system of Excavator to control the grab open, close and rotation.
ï‚·Design for Soil, sand, slag and other bulk material and construction waste handling in Construction yard.
How to choose
-which type do you want
The Arm width, pin size and pin center to center of the grab bucket
Grab Bucket,Hydraulic Grab Bucket,Orange Peel Grab Bucket,Wire Rope Grab Bucket
Jiangxi Mide Industrial Co., Ltd. , https://www.midecrane.com