The invention discloses a static target detection method based on two-background differencing. The method comprises the steps as follows: acquiring a video image; constructing a background model; detecting a moving target; establishing a pure background video image; detecting a foreground target; detecting a static target. Background modeling is performed by use of an improved Gaussian mixture model, and the method can be applied to accurate modeling of backgrounds having slow illumination change or repeated moving objects in complex scenes, particularly to conditions with slow illumination and weather change or higher moving object speed; by reducing the number of Gaussian distribution functions established for each pixel, the calculation amount is reduced, and the real-time performance is improved. Foreground object extraction is performed with a background differencing method, the principle and algorithm design are simple, the obtained result directly reflects the position, size and shape of a foreground object, and more accurate foreground object information can be obtained. A two-background model algorithm is adopted for static object detection, so that the method is lower in complexity and easy to implement.

Static target detection method based on two-background differencing

The present invention provides a human movement identification method through fusion of a deep neural network model and binary system Hash, belonging to the technical field of mode identification. Themethod comprises the steps of: performing preprocessing of a movement identification database, dividing the movement identification database into frame sequences, calculating an optical flow graph, employing an attitude estimation algorithm to calculate coordinates of human joint points, and employing result coordinates to extract video area frames; employing a pre-training VGC-16 network model to extract FC (Full-Convolutional) features of RGB flows and optical flows of the videos, selecting key frames from the video frame sequences, and obtaining a difference of the FC features corresponding to the key frames; performing binary processing of the difference; employing a binary-hashing method to obtain uniform feature expression of each video; employing a plurality of normalization methods such as L1 and L2 to obtain feature expressions of the videos after the fusion of the uniform feature expressions and the PCNN features; and finally, employing a support vector machine algorithm totrain a classifiers to identify the human movement videos. The human movement identification method through fusion of the deep neural network model and the binary system Hash has a high movement identification correct rate.

Human movement identification method through fusion of deep neural network model and binary system Hash

The invention discloses a moving object detection method based on symmetric inter-frame difference and background updating. The method comprises steps of firstly, improving a symmetric inter-frame difference method, differentiating every two of adjacent four frames of images and carrying out 'OR' operation on every two of obtained differential images to extract a moving object profile; using the first frame of image as an initial background, carrying out updating of the background images by combining the current frame of image and the previous frame of background image, and extracting a moving object profile through background differentiation; and finally, carrying out 'AND' operation on the two extracted moving object profiles, and carrying out connectivity analysis and morphological processing, precisely extracting a complete profile of the moving object. According to the invention, problems of a lengthened moving object and a fuzzy profile caused by the symmetric inter-frame difference method are solved; inaccurate extraction of the moving object caused by interference of light and background change in the smooth average background difference method is avoided; the complete profile of the moving object is effectively extracted; and excellent timeliness and robustness are achieved.

Moving object detection method based on symmetric inter-frame difference and background updating

The invention discloses an infrared thermal imager inhomogeneity two-point correction module and method. Low temperature blackbody images and high temperature blackbody images are acquired by a detector and received by the infrared thermal imager inhomogeneity two-point correction module and then weighted average operation is performed, and the operation result is stored in serial FLASH so that calibration is completed; and calibration images are firstly read from the serial FLASH to be stored in an external storage device in correction, correction parameters are calculated in real time after effective data come, and bad element positions are marked so that correction and bad element replacement can be completed. A multi-frame weighted average method is introduced in calibration so that accuracy of calibration is enhanced; the external storage device is introduced for arbitration control, calibration, correction and FLASH operation are integrated together, and the algorithm is highly modularized so that transplantability can be enhanced; and bad elements are positioned when correction is performed so that bad element replacement can be completed.

Infrared thermal imager inhomogeneity two-point correction module and method

The invention provides a motion detection method based on an image sequence. The method comprises successively acquiring one frame of image, down-sampling the image and updating the image into a cyclic image queue; periodically calculating an expected value of image gradient and acquiring an exposure time so as to determine a determination threshold of motion detection; acquiring two frames of image in the cyclic image queue, calculating the pixel grayscale difference mean value of the image and obtaining an overall grayscale change condition; and determining whether a current frame of image is moving according to the determination threshold, the pixel grayscale difference mean value, and the overall grayscale change condition, obtaining a motion detection result of continuous multiple frames of image in a preset time to output the motion state or the stationary state of the current frame of image. The present invention also provides a motion detection device based on the image sequence. The motion detection method and device based on the image sequence can well adapt to various scenes, and can quickly detect whether the current frame of image is in a stationary state or a motion state, so as to provide an effective strategy for a determination whether to activate anti-shake photography.

Motion detection method and device based on image sequence

The invention discloses an inter-frame noise reduction method based on motion detection According to the method, moving targets are extracted by a multi-Gaussian mixture background model method and an overlapping stationary area between two adjacent frames are found, then inter-frame accumulative filtering is performed on the area, and moving target areas and non-moving target areas in non-overlapping areas are replaced by background models established by an intra-frame filtering algorithm and the multi-Gaussian mixture background model method respectively Meanwhile, the algorithm can also self-adaptively adjust the number of stack frames and has a multistage adjustable function The innovative points reside in that moving target detection of images is performed firstly, then AND operation is performed on two successive frames of foreground images including the moving targets only, and the inter-frame filtering algorithm, the intra-frame filtering algorithm or a background model replacement algorithm is selected according to the result of AND operation so that the phenomena of edge virtual images, pseudo images and even lost of the moving targets caused by the conventional inter-frame filtering algorithm can be avoided, and the great noise reduction effect of the moving images can also be achieved

Inter-frame noise reduction method based on motion detection

The utility model discloses a two-point correction infrared thermal imaging system non uniformity module; a detector collects a low temperature blackbody image and a high temperature blackbody image; the images are received by the two-point correction infrared thermal imaging system non uniformity module for weight average calculation, and calculated result is stored in a serial FLASH, thus completing scaling; in correction, the scaling images are read out from the serial FLASH and stored in an external memory, and correction parameter can be calculated in real time after effectively data comes, and bad element position is marked, thus carrying out correction and bad element replacement. The two-point correction infrared thermal imaging system non uniformity module introduces multi-frame weight average in scaling, thus improving scaling accuracy; the external memory is arranged for arbitration control, thus integrating scaling, correction and FLASH operation, and algorithm is highly modularized, thus improving transplantability; the bad element can be positioned in correction, thus completing bad element replacement.

Two-point correction infrared thermal imaging system non uniformity module

The invention discloses a blocking perception Hash tracking method with shadow removing. The method comprises the steps of: determining shadow areas in an image according to the distribution characteristics of a shadow image in each channel grey-scale map of a CIELAB color space; then utilizing a color constancy theory to recover pixel points in the shadow areas to a non-shadow effect; combining blocking perception Hash coding values with color self-similarity for forming a similarity measure, and carrying out matching on tracked target sub-blocks of adjacent frames based on the similarity measure; and finally, combining the above sub-blocks to obtain the regional position of the tracked target in the current frame, and realizing the tracking of the tracked target in a video. The blocking perception Hash tracking method has the advantages that according to different moving ranges and deforming degrees of human body parts, a human body target is divided into eight sub-blocks, and on this basis, and the blocking perception Hash coding method is provided to solve the problem of an existing tracking algorithm that the tracking is unsuccessful when the human body is partially or totally shielded or partially rotated and when the illumination in the shadow areas and non-shadow areas of a natural scene changes suddenly.

Blocking perception Hash tracking method with shadow removing

The invention discloses a variable integral time non-uniformity correction method based on scenes. The method comprises calculating row projection vectors and column projection vectors of background images of a detector under different integral conditions firstly, storing the background row and column projection vectors, and then forming a lookup table. When the detector shoot an image at the different scenes, in order to determine best integral time of the detector, the row and column projection vectors of the image are calculated firstly, related operation respectively between the row and column projection vectors and the background row and column vectors in the lookup table are performed, the most relevant background of the image is found, and the best integral time of the detector is determined through the background. Innovation points of the method are in that the best integral time of the detector in the different scenes can be found, the image obtained under the best integral condition is in favor of subsequent image registration, registration precision is improved, and then the method has the good non-uniformity correction effect.

He Weiji

Papers

Inter-frame noise reduction method based on motion detection

Two-point correction infrared thermal imaging system non uniformity module

Blocking perception Hash tracking method with shadow removing

Variable integral time non-uniformity correction method based on scenes