Summary of the invention
In view of above-mentioned the deficiencies in the prior art, the object of the present invention is to provide a kind of panoramic parking assist system and its implementation, be intended to solve the safety of existing automobile parking process existence and the problem of stability.
Technical scheme of the present invention is as follows:
A kind of panoramic parking assist system, wherein, comprise be arranged on car body all around for gathering the wide-angle camera of peripheral vehicle image, for proofreading and correct and splicing the image of wide-angle camera collection with the panoramic picture generation module of generating panorama image, for described panoramic picture being carried out to the detection module of obstacle recognition and for according to the reminding module of the testing result output detections information of detection module, described wide-angle camera, panoramic picture generation module, detection module and reminding module are electrically connected successively.
Described panoramic parking assist system, wherein, is provided with image pattern storehouse in described detection module, store the positive sample image that comprises obstacle information in panoramic picture and do not contain the negative sample image of obstacle information in described image pattern storehouse.
Described panoramic parking assist system, wherein, is also provided with target classification device in described detection module, described target classification device carries out obstacle target detect for positive sample image and negative sample image that image pattern storehouse is stored, and then cognitive disorders thing.
Described panoramic parking assist system, wherein, described detection information is exported with image, word or speech form.
Described panoramic parking assist system, wherein, described reminding module comprises the display device for mounting on vehicle for showing detection information.
Described panoramic parking assist system, is characterized in that, described display device for mounting on vehicle is one or more combinations in vehicular LCD display screen, OLED screen, LED screen, video input display equipment.
Described panoramic parking assist system, wherein, described reminding module also comprises the voice playing module for playing detection information.
Described panoramic parking assist system, wherein, described obstacle comprises people, animal or Rigid Bodies.
An implementation method for described panoramic parking assist system, wherein, comprises the following steps:
A, described wide-angle camera gather the image of peripheral vehicle;
B, described panoramic picture generation module are to described correct image, splicing generating panorama image;
C, described detection module carry out obstacle recognition to described panoramic picture and obtain testing result;
D, described reminding module are according to testing result output detections information.
Beneficial effect: panoramic parking assist system of the present invention and its implementation, gather the image of peripheral vehicle by wide-angle camera; Panoramic picture generation module is to described correct image, splicing generating panorama image; Detection module carries out obstacle recognition to described panoramic picture and obtains testing result; Reminding module is according to testing result output detections information, the detection information accurate and effective that the present invention adopts multiple technologies to obtain, can effectively judge the obstacle information of peripheral vehicle, the operations such as driving of parking of chaufeur are facilitated, improve the safety of driving, cost of the present invention is low, device is simple, practical.
Detailed description of the invention
The invention provides a kind of panoramic parking assist system and its implementation, for making object of the present invention, technical scheme and effect clearer, clear and definite, below the present invention is described in more detail.Should be appreciated that specific embodiment described herein, only in order to explain the present invention, is not intended to limit the present invention.
Panoramic parking assist system of the present invention, as described in Figure 1, comprise be arranged on car body all around for gathering the wide-angle camera 100 of peripheral vehicle image, for proofreading and correct and splice image that wide-angle camera 100 the gathers panoramic picture generation module 200 with generating panorama image, for described panoramic picture being carried out to the detection module 300 of obstacle recognition and for according to the reminding module 400 of the testing result output detections information of detection module, described wide-angle camera 100, panoramic picture generation module 200, detection module 300 and reminding module 400 are electrically connected successively.Below the each module in the present invention is described in detail.
Wide-angle camera 100 is arranged on car body all around in the present invention, thereby apply four wide-angle cameras 100 and obtain the image that peripheral vehicle 360 is spent whole orientation, specifically wide-angle camera 100 can be separately positioned on to grid position and the left side of Vehicular body front, in the outside rear-view mirror of right side, be on back door, to obtain best shooting effect, certainly if necessary, some wide-angle cameras also can be set more, in case run into shooting blind angle amount, 8 wide-angle cameras are for example set, cover the image of all field ranges of peripheral vehicle, allow chaufeur is clear checks whether vehicle-surroundings exists obstacle and understand relative azimuth and the distance of obstacle, help the light parked vehicle of chaufeur.
Described panoramic picture generation module 200 for correct image that wide-angle camera 100 is gathered and splicing to generate the panoramic picture of peripheral vehicle, specifically, this wide-angle camera 100 can be input to the vision signal that comprises image in panoramic picture generation module 200, then panoramic picture generation module 200 carries out visual angle conversion and synthetic fusion treatment to image, 4 road images are synthesized to a birds-eye view directly over complete car body, as for specifically how correct image, splicing can, with reference to prior art, being repeated no more.
Core part of the present invention is that panoramic picture is carried out to obstacle recognition, thereby judge whether vehicle-surroundings has obstacle, and will detect information indicating to chaufeur, facilitate chaufeur directly to obtain detection information, without carrying out unnecessary work, alleviate the burden of chaufeur.Specifically describe how panoramic picture is carried out to obstacle recognition and output detections information below.
In detection module 300, be provided with image pattern storehouse, in this image pattern storehouse, store the positive sample image that contains obstacle information, and do not contain the negative sample image of obstacle information.In detection module, be also provided with target classification device, this target classification device is to have utilized obstacle feature in image pattern storehouse to train to obtain, after training completes, target classification device can carry out recognition detection to for example obstacle of interested region in panoramic picture, when obstacle detected, target classification device is output as 1, otherwise is output as 0; In order to detect view picture panoramic picture, can be in panoramic picture mobile search window, in order to detect the obstacles of different sizes, can carry out size change to target classification device, testing result will be truer like this.
Target classification device in the present invention is that training generates target classification device based on embedded platform, and the sorter model drawing by this training method can reach the object that reduces calculated amount.
In the present invention, described detection information is exported with image, word or speech form.Obstacle can be the Rigid Bodies of people, animal or for example vehicle.
Described reminding module 400 comprises the display device for mounting on vehicle for showing detection information, when this detection information is presented at display device for mounting on vehicle with graphicform, obstacle can be identified, utilize marking frame to identify obstacle, thus user friendly checking.Described display device for mounting on vehicle can be one or more combinations in vehicular LCD display screen, OLED screen, LED screen, video input display equipment, facilitates user according to use habit and need to carry out free selection.
Described reminding module 400 also comprises the voice playing module for playing detection information, when this detection information is while exporting with speech form, can in detection information, add the azimuth information of obstacle, by voice playing module, thereby detection information is play to prompting driver again and note where is it obstacle, this prompting mode has alleviated driver's burden, also convenient for users to use.Certainly, also above-mentioned detection information can be exported in one or more modes that combine of image, word or voice, to reach the object of hommization more.
Based on said system, the present invention also provides a kind of implementation method of described panoramic parking assist system, as shown in Figure 2, comprises the following steps:
S101, described wide-angle camera gather the image of peripheral vehicle;
S102, described panoramic picture generation module are to described correct image, splicing generating panorama image;
S103, described detection module carry out obstacle recognition to described panoramic picture and obtain testing result;
S104, described reminding module are according to testing result output detections information.
In sum, panoramic parking assist system of the present invention and its implementation, gather the image of peripheral vehicle by wide-angle camera; Panoramic picture generation module is to described correct image, splicing generating panorama image; Detection module carries out obstacle recognition to described panoramic picture and obtains testing result; Reminding module is according to testing result output detections information, the detection information accurate and effective that the present invention adopts multiple technologies to obtain, can effectively judge the obstacle information of peripheral vehicle, the operations such as driving of parking of chaufeur are facilitated, improve the safety of driving, cost of the present invention is low, device is simple, practical.
Should be understood that, application of the present invention is not limited to above-mentioned giving an example, and for those of ordinary skills, can be improved according to the above description or convert, and all these improvement and conversion all should belong to the protection domain of claims of the present invention.