CN104978547A - Object-detecting system - Google Patents
Object-detecting system Download PDFInfo
- Publication number
- CN104978547A CN104978547A CN201410129820.0A CN201410129820A CN104978547A CN 104978547 A CN104978547 A CN 104978547A CN 201410129820 A CN201410129820 A CN 201410129820A CN 104978547 A CN104978547 A CN 104978547A
- Authority
- CN
- China
- Prior art keywords
- image
- unique point
- described unique
- module
- time point
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Landscapes
- Image Analysis (AREA)
Abstract
The invention provides an object-detecting system comprising a memory unit, an image-acquiring module, a characteristic-searching module, a computing module, and a confirming module. The image-acquiring module reads a first image corresponding to a first time point, a second image corresponding to a second time point, and a third image corresponding to a third time point in a video stream from the memory unit, wherein a time difference between the first time point and the second time point is less than or equal to that between the second time point and the third time point. The characteristic-searching module searches characteristic points from the first image. When determining the characteristic points are located in close-distance detection area, the computing module computes the motion vectors of the characteristic points according to the first image and the second image, otherwise, computes the motion vectors of the characteristic points according to one of the first image and the second image, and the third image. The confirming module confirms whether a motion object exists according to the motion vectors of the characteristic points.
Description
Technical field
The invention relates to a kind of articles detection system, and relate to a kind of moving article detection system especially.
Background technology
Along with sale of automobile amount is constantly grown up, auto electronic industry is also flourish, wherein auto electronic industry contains six broad aspect such as automotive safety system, bodywork system, driver information system, suspention chassis system, engine transmission system, safety system, and it is the highest with the year multiple growth rate of automotive safety system.
But under automobile quantity is growing, the probability that road accident occurs also increases year by year.Show according to statistical data, as long as driver obtains early warning before 0.5 second of colliding, namely can avoid the car accident that knocks into the back of at least 60%, the head-on car accident of 30% and the road surface related accidents of 50%, and if have the pre-warning time in 1 second, then can avoid the accident of 90%.Therefore, the importance of automotive safety system has some idea of.
Significantly reduce along with vision facilities cost and computerized image identification technique is maked rapid progress, the vehicle based on computerized image and pedestrian warn technology becomes main direction of studying.This type of technology mainly installs the detection system based on computer vision onboard, continuously shot images in vehicle travels, again with the technology for detection road surfaces such as image procossing, computer vision and pattern recognition, roadside, other static and dynamic Status information such as middle vehicle, pedestrian of advancing, and then sense environmental aspect exactly, to provide driver more running information.
Detection system based on computer vision has diversity application, economy and uses the advantages such as elasticity.But, for the algorithm of various image procossing, its accuracy detected and usefulness difference to some extent.Therefore, a kind of high-accuracy how is provided and one of real subject under discussion be concerned about for those skilled in the art of dynamical detection system.
Summary of the invention
The invention provides a kind of articles detection system, it accurately and in real time can detect the moving article in video streaming.
The invention provides a kind of articles detection system, comprise storage unit, image collection module, feature search module, computing module and confirm module.Storage unit at least comprises the video streaming of multiple images in order to store.Image collection module, couple storage unit, in order to read the first image, the second image corresponding to the second time point and the 3rd image corresponding to the 3rd time point that correspond to very first time point in video streaming from storage unit, wherein very first time point and the mistiming of the second time point are less than or equal to the mistiming of the second time point and the 3rd time point.Feature search module in order to search at least one unique point in the first image.Computing module is in order to judge that described unique point is positioned at in-plant surveyed area or remote surveyed area.When described unique point is positioned at in-plant surveyed area, computing module, according to the first image and the second image, calculates the motion-vector of each described unique point; When described unique point is positioned at remote surveyed area, computing module, according to the first image and one of them person of the second image and the 3rd image, calculates the motion-vector of each described unique point.Confirm that module is in order to the motion-vector according to each described unique point, confirm whether moving article exists.
In one embodiment of this invention, the above-mentioned feature search module described unique point of searching is formed by a 1x1 picture element or multiple picture element.
In one embodiment of this invention, above-mentioned computing module is the motion-vector utilizing multiple pixel characteristic Comparison Method or optical flow method to calculate each described unique point.
In one embodiment of this invention, above-mentioned computing module is according to the motion-vector of each described unique point, calculate the displacement of each described unique point, thus calculate the translational speed of each described unique point, and confirm that module judges whether the translational speed of each described unique point is greater than the speed threshold value with each unique point distance dependent.When confirming that module judges that the translational speed of at least one described unique point is greater than speed threshold value, confirm that module confirms that moving article exists.
The invention provides another kind of articles detection system, comprise storage unit, image collection module, feature search module, computing module and confirm module.Storage unit at least comprises the video streaming of multiple images in order to store.Image collection module, couple storage unit, in order to read the first image, the second image corresponding to the second time point, the 3rd image corresponding to the 3rd time point and the 4th image corresponding to the 4th time point that correspond to very first time point in video streaming from storage unit, wherein very first time point and the mistiming of the second time point are less than or equal to the mistiming of the 3rd time point and the 4th time point.Feature search module in order to search at least one unique point in the first image.Computing module is in order to judge that described unique point is positioned at in-plant surveyed area or remote surveyed area.When described unique point is positioned at in-plant surveyed area, computing module, according to the first image and the second image, calculates the motion-vector of each described unique point; When described unique point is positioned at remote surveyed area, computing module, according to the 3rd image and the 4th image, calculates the motion-vector of each described unique point.Confirm that module is in order to the motion-vector according to each described unique point, confirm whether moving article exists.
In one embodiment of this invention, the above-mentioned feature search module described unique point of searching is formed by a 1x1 picture element or multiple picture element.
In one embodiment of this invention, above-mentioned computing module is the motion-vector utilizing multiple pixel characteristic Comparison Method or optical flow method to calculate each described unique point.
In one embodiment of this invention, above-mentioned computing module is according to the motion-vector of each described unique point, calculate the displacement of each described unique point, thus calculate the translational speed of each described unique point, and confirm that module judges whether the translational speed of each described unique point is greater than the speed threshold value with each unique point distance dependent.When confirming that module judges that the translational speed of at least one described unique point is greater than speed threshold value, confirm that module confirms that moving article exists.
Based on above-mentioned, articles detection system provided by the present invention, it utilizes images different in video streaming to calculate the motion-vector of unique point.When unique point is positioned at the surveyed area of far and near distance, two images utilizing the mistiming less to calculate the motion-vector of unique point, with shorten object detect time; When unique point is positioned at remote surveyed area, two images utilizing the mistiming larger to calculate the motion-vector of unique point, with improve object detect accuracy rate.In addition, the present invention, also by the motion-vector of unique point, calculates the displacement of unique point, thus calculates the translational speed of unique point, to confirm the existence of moving article.Base this, articles detection system of the present invention can grasp the information of moving article more accurately, and can reach the usefulness of scan picture, to apply on the consumption electronic products of low cost, strengthens the present invention applicability in actual applications.
For above-mentioned feature and advantage of the present invention can be become apparent, special embodiment below, and coordinate accompanying drawing to be described in detail below.
Accompanying drawing explanation
Fig. 1 is the calcspar of the articles detection system shown by one embodiment of the invention;
Fig. 2 is the process flow diagram according to the object detection method shown by first embodiment of the invention;
Fig. 3 A is the schematic diagram of the first image shown by one embodiment of the invention;
Fig. 3 B is the schematic diagram of step S209 in the object detection method shown by one embodiment of the invention;
Fig. 4 is the process flow diagram according to the object detection method shown by second embodiment of the invention.
Description of reference numerals:
100: articles detection system;
110: storage unit;
120: image collection module;
130: feature search module;
140: computing module;
150: confirm module;
S201 ~ S211, S401 ~ S411: the flow process of object detection method;
30: object;
310: the first images;
320: the second images;
330: the three images;
312,332: surveyed area at a distance;
314,324: in-plant surveyed area.
Embodiment
Generally speaking, when detecting the mobile message of moving article in dynamic image, no matter the distance of moving article position, utilizes one group of identical image to detect often.In the case, when moving article is positioned at surveyed area far away, then the mobile message estimated likely can produce great error with actual mobile message.
For example, the example of table one be utilize identical one group of image to detect to lay respectively at closely with the speed estimation error of remote moving article:
Table one
Often opening in image, the actual displacement of real space (actual space) corresponding to the actual displacement picture element of each picture element is different.In this example, suppose closely time, each actual displacement picture element corresponding to actual displacement picture element is 50cm; When remote, each actual displacement picture element corresponding to actual displacement picture element is 100cm.When estimating displacement picture element and there is 1 pixel comparison error (One-Pixel MatchingError), for closely with speed (5,5) velocity error that the moving article of cm/ms movement is estimated is 20%, and for the same velocity error estimated with the moving article of speed (5,5) cm/ms movement at a distance then for 40%.Therefore, if carry out speed estimation with identical one group of image to moving article, be positioned at remote moving article and can produce larger speed estimation error.Base this, concept of the present invention is the distance first judging moving article position, and readaptability ground utilizes different group image to obtain the mobile message of moving article.
Next section Example of the present invention will coordinate accompanying drawing to describe in detail, the component symbol that following description is quoted, when different accompanying drawing occurs that identical component symbol will be considered as same or analogous element.These embodiments are a part of the present invention, and unexposed all embodiments of the present invention.More precisely, these embodiments are the example of the apparatus and method in patent claim of the present invention.
Fig. 1 is the calcspar of the articles detection system shown by one embodiment of the invention, but this is only for convenience of description, not in order to limit the present invention.First Fig. 1 first introduces all components and the configuration relation of articles detection system, and detailed functions will coordinate Fig. 2 open in the lump.
Please refer to Fig. 1, articles detection system 100 comprises storage unit 110, image collection module 120, feature search module 130, computing module 140 and confirms module 150.Articles detection system 100 can be built-in or be external in the electronic installation that PC, notebook computer, digital camera, digital code camera, network camera, smart mobile phone, panel computer, drive recorder, vehicle video-audio system etc. have shoot function, and the present invention is not as limit.
Storage unit 110 can be such as the fixed of arbitrary form or packaged type random access memory (Random Access Memory, be called for short RAM), ROM (read-only memory) (Read-Only Memory, be called for short ROM), the combination of flash memory (Flash memory), hard disk or other similar devices or these devices, in order to store view data.
Image collection module 120, feature search module 130, computing module 140 and confirmation module 150 can be realized by software program, hardware circuit or its combination, in order to perform the function that object detects.Software is such as source code, application software, driver or specially in order to realize the software module of specific function or function etc.Hardware is such as CPU (central processing unit) (Central Processing Unit, be called for short CPU), programmable controller, digital signal processor (Digital Signal Processor, be called for short DSP), or the microprocessor of the general service of other programmables or specific use (Microprocessor) etc.For example, image collection module 120, feature search module 130, computing module 140 and confirmation module 150 are such as computer program or instruction, and it is by the processor of articles detection system 100, thus performs the function of object detection.
Fig. 2 is the process flow diagram according to the object detection method shown by first embodiment of the invention.Please refer to Fig. 2, the method for the present embodiment is applicable to the articles detection system 100 of Fig. 1, the detailed step of the every component description object detection method in articles detection system 100 of namely arranging in pairs or groups below.In the present embodiment, driving recorder will be coupled to be described for articles detection system 100.Therefore, before execution object detection method, the storage unit 110 of articles detection system 100 has stored the video streaming comprising multiple images that bicycle register in advance obtains, image before wherein the content of video streaming is captured by driving recorder car.
Referring to Fig. 1 and Fig. 2, image collection module 120 reads the first image, the second image corresponding to the second time point and the 3rd image corresponding to the 3rd time point that correspond to very first time point in video streaming from storage unit 110, suppose very first time point t
1early than the second time point t
2, the second time point t
2early than the 3rd time point t
3, then very first time point and the mistiming of the second time point are less than the mistiming (step S201) of very first time point and the 3rd time point.In other words, very first time point t
1, the second time point t
2and the 3rd time point t
3the relation of inequality (1) will be met:
| t
1-t
2|≤| t
1-t
3| inequality (1)
Then, feature search module 130 searches at least one unique point (step S203) in the first image.In detail, feature search module 130 searches the unique point in the first image according to the calculation method (FeatureDetection Algorithm) of existing feature point detection, with the shape of the object occurred in identification first image or structure.And the described unique point that feature search module 130 is searched is formed by a 1x1 picture element or multiple picture element.For example, in the present embodiment, the object feature that feature search module 130 will be searched can be vehicles or pedestrians, and described unique point is edge contour, angle point or the block etc. that form vehicles or pedestrians, and the present invention does not limit at this.
In the present embodiment, feature search module 130 utilization orientation histogram of gradients (Histogram ofOriented Gradient, be called for short HOG) can search unique point in the first image.First Iamge Segmentation is first multiple pane location (cell) by feature search module 130, and carrys out the foundation as identification unique point by the Grad of pixel each in pane location and with gradient direction.This mode is comparatively strong for the object identification capability with obvious border, and the impact that partial modifications or small-sized foreign matter can be avoided to cause for image, be applicable in the application such as pedestrian detection.
In another embodiment, the calculation mode of feature point detection can be such as rim detection (EdgeDetection) algorithm, by the change of GTG to search the frontier point of the first objects in images and background, or utilize angle to detect (Corner Detection) algorithm, by local bright spot or dim spot, line segment terminal, curve having the point corresponding to maximum curvature, to search the angle point of the first objects in images.Whether the profile that all frontier points of feature search module 130 identification or angle point are formed is be such as the feature of vehicle or pedestrian.Whether in another embodiment, feature search module 130 also can detect (BlobDetection) algorithm according to block, utilizes color and GTG to detect the discrepant block with surrounding, be such as the feature of vehicle or pedestrian with this block of identification.The present invention does not limit at the calculation method of feature point detection.
When feature search module 130 cannot search unique point in the first image, articles detection system 100 will terminate the object testing process in the first image.And when feature search module 130 hunts out at least one unique point in the first image, computing module 140 judges that described unique point is positioned at in-plant surveyed area or remote surveyed area (step S205).In the region that this in-plant surveyed area is nearer apart from the position (i.e. driving recorder) of shooting the first image, and surveyed area is the region far away apart from the position of shooting the first image at a distance.
For example, Fig. 3 A is the schematic diagram of the first image shown by one embodiment of the invention.
Please refer to Fig. 3 A, articles detection system 100 is after storage unit 110 obtains the first image 310, and feature search module 130 can search unique point from the first image 310.In the present embodiment, described unique point is and forms the feature of object 30, and this object 30 is forward and the vehicle that comes.Namely computing module 140 can judge that the described unique point forming object 30 is positioned at remote surveyed area 314 or in-plant surveyed area 312.And in first image 310 of Fig. 3 A, described unique point is then be positioned at remote surveyed area 314.
When computing module 140 judges that described unique point is positioned at in-plant surveyed area, according to the first image and the second image, calculate the motion-vector (step S207) of each described unique point; When computing module 140 judges that described unique point is positioned at remote surveyed area, according to the first image and one of them person of the second image and the 3rd image, calculate the motion-vector (step S209) of each described unique point.In detail, when described unique point is positioned at in-plant surveyed area, the profile that unique point is formed or the block area shared by each image is larger.At one time and under translational speed, the displacement of described unique point is comparatively remarkable, therefore computing module 140 first image that the mistiming can be utilized less and the second image are to calculate the motion-vector of each described unique point, to reduce the search time of unique point and to reach effect of real-time process (Real-time Processing).On the other hand, when described unique point is positioned at remote surveyed area, the profile that unique point is formed or the block area shared by each image is less.At one time and under translational speed, the displacement of described unique point is more not remarkable, therefore computing module 140 first image that the mistiming can be utilized larger and the 3rd image are to calculate the motion-vector of each described unique point, to calculate the motion-vector being positioned at unique point described in a distant place more accurately.
In another embodiment, as very first time point t
1, the second time point t
2and the 3rd time point t
3when meeting the relation of inequality (1) and following inequality (1.5):
| t
1-t
2|≤| t
2-t
3| inequality (1.5)
When computing module 140 judges that described unique point is positioned at in-plant surveyed area, according to the first image and the second image, calculate the motion-vector of each described unique point; When computing module 140 judges that described unique point is positioned at remote surveyed area, then can according to mistiming larger the first image and the 3rd image, or the second image and the 3rd image, calculate the motion-vector of each described unique point.
Fig. 3 B is the schematic diagram of step S209 in the object detection method shown by one embodiment of the invention.
Please refer to Fig. 3 B, suppose that image collection module 120 reads in video streaming from storage unit 110 and correspond to very first time point t
1the first image 310, corresponding to the second time point t
2the second image 320 and correspond to the 3rd time point t
3the 3rd image 330.As Fig. 3 A say, the first image 310 comprises in-plant surveyed area 314 and remote surveyed area 312.When computing module 140 judges that preceding feature point is positioned at in-plant surveyed area 314, computing module 140 according to the in-plant surveyed area 324 of the in-plant surveyed area 314 of the first image 310 and the second image 320, can calculate the motion-vector of each described unique point.
On the other hand, when computing module 140 judges that preceding feature point is positioned at remote surveyed area 312, in the present embodiment, computing module 140 according to the remote surveyed area 332 of the remote surveyed area 312 of the first image 310 and the 3rd image 330, can calculate the motion-vector of each described unique point.In another embodiment, computing module 140 also according to the remote surveyed area 332 of the remote surveyed area (not shown) of the second image 320 and the 3rd image 330, can calculate the motion-vector of each described unique point.
The example of table two be utilize two groups of different images to detect respectively to be positioned at closely with the speed estimation error of remote moving article:
Table two
Often opening in image, the actual displacement of real space corresponding to the actual displacement picture element of each picture element is different.In this example, suppose closely time, each actual displacement picture element corresponding to actual displacement picture element is 50cm; When remote, each actual displacement picture element corresponding to actual displacement picture element is 100cm.When described unique point is positioned at remote surveyed area, when same 1 pixel comparison error, the present invention utilizes the first image and the 3rd image for remote with speed (5cm/ms, the velocity error that moving article 5cm/ms) is estimated is 20%, be 40% to compare with the velocity error of table one, improve 2 times.
In addition, computing module 140 can utilize pixel characteristic Comparison Method (Pixel-matching Algorithm) or optical flow method (Optical Flow Algorithm) to calculate the motion-vector of each described unique point.
In detail, suppose that feature search module 130 is after the unique point that the first image searches is positioned at in-plant surveyed area, computing module 140 utilizes pixel characteristic Comparison Method, to find out the unique point similar in appearance to the first image 310 in the second image 320.In the present embodiment, computing module 140 can correspond to centered by the position of the unique point of the first image 310 in the second image 320, utilizes the features such as the color of the unique point of the first image 310 to be positioned at the unique point of the second image 320 from the picture element search that surrounding is contiguous.Then, computing module 140 can utilize the position of the unique point in the first image 310 and the second image 320 to calculate the motion-vector of each described unique point in continuous time.
In another embodiment, suppose that feature search module 130 is positioned at in-plant surveyed area from the unique point that the first image searches, and after computing module 140 searches characteristic of correspondence point in the second image, the unchangeability of each described unique point brightness value in different images in optical flow method can be utilized, calculate its motion-vector.
Afterwards, confirm that module 150 is according to the motion-vector of each described unique point, confirm whether moving article exists (step S211).In detail, computing module is after the motion-vector obtaining each described unique point, translational speed, frame per second (Frame Rate) and motion-vector during shooting image can be utilized, calculate and respectively state the displacement of unique point, and thus calculate the translational speed of each described unique point.Confirm that module 150 will judge whether the translational speed of each described unique point is greater than a speed threshold value.Translational speed when this speed threshold value then must consider shooting image.In other words, confirm the difference between translational speed when module 150 can judge the translational speed of each described unique point and take image, whether be greater than speed threshold value.In one embodiment, the unique point being positioned at in-plant surveyed area can be different from the speed threshold value corresponding to the unique point being positioned at remote surveyed area, judge more accurately to reach.
When confirming that module 150 judges that the translational speed of at least one described unique point is greater than speed threshold value, representing the feature that described unique point is such as other moving article such as driving vehicle or pedestrian, confirming that module 150 will confirm that moving article exists.Otherwise representing described unique point is be such as the feature of the static objects such as road, buildings or roadside parked vehicles, confirms that confirmation moving article does not exist by module 150.
In one embodiment, articles detection system 100 also comprises alarm module (not shown), when confirming that module 150 confirms that moving article exists, alarm module can provide such as prompt text, the sound or light one or a combination set of, to remind the existence of driving moving article.
Fig. 4 is the process flow diagram according to the object detection method shown by second embodiment of the invention.Please refer to Fig. 4, the method for the present embodiment is also applicable to the articles detection system 100 of Fig. 1.In the present embodiment, similarly driving recorder will be coupled to be described for articles detection system 100.Therefore, before execution object detection method, the storage unit 110 of articles detection system 100 has stored the video streaming comprising multiple images that bicycle register in advance obtains, image before wherein the content of video streaming is captured by driving recorder car.
Referring to Fig. 1 and Fig. 4, image collection module 120 reads the first image, the second image corresponding to the second time point, the 3rd image corresponding to the 3rd time point and the 4th image corresponding to the 4th time point that correspond to very first time point in video streaming from storage unit 110, suppose very first time point t
1early than the second time point t
2, the second time point t
2early than the 3rd time point t
3, the 3rd time point t
3early than the 4th time point t
4, and very first time point and mistiming of the second time point be less than or equal to the mistiming (step S401) of the 3rd time point and the 4th time point.In other words, very first time point t
1, the second time point t
2, the 3rd time point t
3and the 4th time point t
4the relation of inequality (2) will be met:
| t
1-t
2|≤| t
3-t
4| inequality (2)
Then, feature search module 130 searches at least one unique point (step S403) in the first image.When feature search module 130 cannot search unique point in the first image, articles detection system 100 will terminate the object testing process in the first image.And when feature search module 130 hunts out at least one unique point in the first image, computing module 140 judges that described unique point is positioned at in-plant surveyed area or remote surveyed area (step S405).When computing module 140 judges that described unique point is positioned at in-plant surveyed area, according to the first image and the second image, calculate the motion-vector (step S407) of each described unique point; When computing module 140 judges that described unique point is positioned at remote surveyed area, according to the 3rd image and the 4th image, calculate the motion-vector (step S409) of each described unique point.Afterwards, confirm that module 150 is according to the motion-vector of each described unique point, confirm whether moving article exists (step S411).
The difference of the embodiment of Fig. 4 and the embodiment of Fig. 2 is the amount of images for calculating needed for motion-vector.But the concept that both embodiments are used is similar, when unique point is positioned at in-plant surveyed area, two images that computing module 140 utilizes the mistiming less are to calculate the motion-vector of unique point; When unique point is positioned at remote surveyed area, two images that computing module 140 utilizes the mistiming larger are to calculate the motion-vector of unique point.The detailed content of above-mentioned steps S403 ~ step S411, those skilled in the art can refer to the related description of Fig. 1 to Fig. 3 B and analogize it, do not repeat them here.
In sum, articles detection system provided by the present invention, it utilizes images different in video streaming to calculate the motion-vector of unique point.When unique point is positioned at remote surveyed area, two images utilizing the mistiming larger to calculate the motion-vector of unique point, with shorten object detect time; When unique point is positioned at remote surveyed area, two images utilizing the mistiming larger to calculate the motion-vector of unique point, with improve object detect accuracy rate.In addition, the present invention, also by the motion-vector of unique point, calculates the displacement of unique point, thus calculates the translational speed of unique point, to confirm the existence of moving article.Therefore, articles detection system of the present invention can grasp the information of moving article more accurately, and can reach the usefulness of scan picture, to apply on the consumption electronic products of low cost, strengthens the present invention's applicability in actual applications.
Last it is noted that above each embodiment is only in order to illustrate technical scheme of the present invention, be not intended to limit; Although with reference to foregoing embodiments to invention has been detailed description, those of ordinary skill in the art is to be understood that: it still can be modified to the technical scheme described in foregoing embodiments, or carries out equivalent replacement to wherein some or all of technical characteristic; And these amendments or replacement, do not make the essence of appropriate technical solution depart from the scope of various embodiments of the present invention technical scheme.
Claims (8)
1. an articles detection system, is characterized in that, comprising:
Storage unit, at least stores the video streaming comprising multiple images;
Image collection module, couple this storage unit, responsible reading is stored in the first image, the second image corresponding to the second time point and the 3rd image corresponding to the 3rd time point that correspond to very first time point in this video streaming in this storage unit, and wherein this very first time point and the mistiming of this second time point are less than or equal to the mistiming of this second time point and the 3rd time point;
Feature search module, searches at least one unique point in this first image;
Computing module, judge that described unique point is positioned at in-plant surveyed area or remote surveyed area, when described unique point is positioned at this in-plant surveyed area, according to this first image and this second image, calculate the motion-vector of each described unique point, when described unique point is positioned at this remote surveyed area, according to this first image and one of them person of this second image and the 3rd image, calculate the motion-vector of each described unique point; And
Confirm module, according to the motion-vector of each described unique point, confirm whether this moving article exists.
2. articles detection system according to claim 1, is characterized in that, this feature search module described unique point of searching is formed by a 1x1 picture element or multiple picture element.
3. articles detection system according to claim 1, is characterized in that, this computing module is the motion-vector utilizing multiple pixel characteristic Comparison Method or optical flow method to calculate each described unique point.
4. articles detection system according to claim 1, it is characterized in that, this computing module is according to the motion-vector of each described unique point, calculate the displacement of each described unique point, thus calculate the translational speed of each described unique point, and this confirmation module judges whether the translational speed of each described unique point is greater than the speed threshold value of each unique point distance dependent with this, when this confirmation module judges that the translational speed of at least one described unique point is greater than this speed threshold value, this confirmation module confirms that this moving article exists.
5. an articles detection system, is characterized in that, comprising:
Storage unit, at least stores the video streaming comprising multiple images;
Image collection module, couple this storage unit, responsible reading is stored in the first image, the second image corresponding to the second time point, the 3rd image corresponding to the 3rd time point and the 4th image corresponding to the 4th time point that correspond to very first time point in this video streaming in this storage unit, and wherein this very first time point and the mistiming of this second time point are less than or equal to the mistiming of the 3rd time point and the 4th time point;
Feature search module, searches multiple unique point in this first image;
Computing module, judge that described unique point is positioned at in-plant surveyed area or remote surveyed area, when described unique point is positioned at this in-plant surveyed area, according to this first image and this second image, calculate the motion-vector of each described unique point, when described unique point is positioned at this remote surveyed area, according to the 3rd image and the 4th image, calculate the motion-vector of each described unique point; And
Confirm module, according to the motion-vector of each described unique point, confirm whether this moving article exists.
6. articles detection system according to claim 5, is characterized in that, this feature search module described unique point of searching is formed by a 1x1 picture element or multiple picture element.
7. articles detection system according to claim 5, is characterized in that, the calculation method that this computing module calculates the motion-vector of each described unique point utilizes multiple pixel characteristic Comparison Method or optical flow method.
8. articles detection system according to claim 5, it is characterized in that, this computing module is according to the motion-vector of each described unique point, obtain the displacement of each described unique point, thus calculate the translational speed of each described unique point, and this confirmation module judges whether the translational speed of each described unique point is greater than the speed threshold value of each unique point distance dependent with this, when this confirmation module judges that the translational speed of at least one described unique point is greater than this speed threshold value, this confirmation module confirms that this moving article exists.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410129820.0A CN104978547B (en) | 2014-04-01 | 2014-04-01 | Articles detection system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410129820.0A CN104978547B (en) | 2014-04-01 | 2014-04-01 | Articles detection system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104978547A true CN104978547A (en) | 2015-10-14 |
| CN104978547B CN104978547B (en) | 2019-06-18 |
Family
ID=54275037
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410129820.0A Expired - Fee Related CN104978547B (en) | 2014-04-01 | 2014-04-01 | Articles detection system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104978547B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2001126184A (en) * | 1999-10-29 | 2001-05-11 | Matsushita Electric Ind Co Ltd | Automatic license plate recognizing device and vehicle speed measuring method |
| CN103413325A (en) * | 2013-08-12 | 2013-11-27 | 大连理工大学 | Vehicle speed identification method based on vehicle body feature point positioning |
-
2014
- 2014-04-01 CN CN201410129820.0A patent/CN104978547B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2001126184A (en) * | 1999-10-29 | 2001-05-11 | Matsushita Electric Ind Co Ltd | Automatic license plate recognizing device and vehicle speed measuring method |
| CN103413325A (en) * | 2013-08-12 | 2013-11-27 | 大连理工大学 | Vehicle speed identification method based on vehicle body feature point positioning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104978547B (en) | 2019-06-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11643076B2 (en) | Forward collision control method and apparatus, electronic device, program, and medium | |
| TWI595450B (en) | Object detection system | |
| CN107038723B (en) | Method and system for estimating rod-shaped pixels | |
| Wu et al. | Lane-mark extraction for automobiles under complex conditions | |
| US7366325B2 (en) | Moving object detection using low illumination depth capable computer vision | |
| Liu et al. | Pedestrian detection using stereo night vision | |
| US8005266B2 (en) | Vehicle surroundings monitoring apparatus | |
| US20130286205A1 (en) | Approaching object detection device and method for detecting approaching objects | |
| CN108021856B (en) | Vehicle tail lamp identification method and device and vehicle | |
| JP5480917B2 (en) | Vehicle periphery monitoring device | |
| US10878259B2 (en) | Vehicle detecting method, nighttime vehicle detecting method based on dynamic light intensity and system thereof | |
| Lin et al. | Lane departure and front collision warning using a single camera | |
| US11473921B2 (en) | Method of following a vehicle | |
| JP2007310706A (en) | Vehicle periphery monitoring device | |
| CN111881832A (en) | Lane target detection method, device, equipment and computer readable storage medium | |
| CN112528807A (en) | Method and device for predicting driving track, electronic equipment and storage medium | |
| Tanaka et al. | Vehicle Detection Based on Perspective Transformation Using Rear‐View Camera | |
| CN111144415B (en) | Detection method for tiny pedestrian target | |
| US20220318456A1 (en) | Simulation method based on three-dimensional contour, storage medium, computer equipment | |
| KR20070019347A (en) | device and method for detecting mobile in automated driving system | |
| Dai et al. | A driving assistance system with vision based vehicle detection techniques | |
| US9183448B2 (en) | Approaching-object detector, approaching object detecting method, and recording medium storing its program | |
| US20150178577A1 (en) | Image processing apparatus, and image processing method | |
| CN104978547A (en) | Object-detecting system | |
| CN111192290B (en) | Blocking processing method for pedestrian image detection |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20190618 Termination date: 20200401 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |