A kind of optimized algorithm based on images match indoor positioningTechnical field
The present invention relates to a kind of optimized algorithm based on images match indoor positioning.
Background technology
At present, the indoor positioning of existence has wifi three-point fix and images match location.
Wherein, wifi intensity is located, and adopts the vector of the signal intensity from each access point AP receiving different model mobile terminal and detect at each location point indoor; According to fuzzy clustering algorithm, cluster is carried out to the signal intensity vector received; According to cluster result, each model mobile terminal is divided into stable AP and problem AP at the AP that each location point indoor is corresponding; Cross the signal intensity cluster intermediate value that filtering algorithm calculates the different capacity signal that different model mobile terminal is launched for many times at the same stable AP that indoor same position point detects, and forming many relation curves about mobile terminal model, every bar relation curve identifies the signal intensity cluster intermediate value corresponding to same mobile terminal model different transmission power; Receive the model of the signal intensity from each AP that mobile terminal to be measured detects in indoor and mobile terminal to be measured; By the model of described mobile terminal to be measured and described relation curve, mobile terminal to be measured is converted to corresponding signal intensity with reference to mobile terminal for the signal intensity of each stable AP; After based on described relation curve the signal intensity that mobile terminal to be measured receives being converted to the signal intensity received with reference to mobile terminal, changed reference the mobile terminal signal intensity received and the reference mobile terminal stored in fingerprint base are mated for the signal intensity of each stable AP at each location point indoor, to know the location point at the current place of mobile terminal to be measured.There is the defect of jitter in Wifi signal intensity location, there is deviation when reality is used.
Another kind of, images match is located, accurately can locate the position of people, the Visual Map of image acquisition is used to set up the position of each reference point of mode record, the characteristic point information that the rotation matrix of camera and SURF algorithm draw, to the photograph extract minutiae captured by the user mobile phone needing to locate, its unique point with image each in database is mated, select positional information and the rotation matrix of the highest some width images of matching degree and image corresponding thereto, use the Mismatching point in RANSAC algorithm rejecting matching image afterwards, finally use Epipolar geometry completes the estimation to customer location.
But due in megastore or large-scale room area, when the feature image traveling through every shop or sample is mated, the time of needs is long, but user is very harsh the time response to location.Generally more than 5 seconds, also do not complete location, user just can select to close this position application software.
Summary of the invention
The present invention is to solve above-mentioned deficiency, provides a kind of optimized algorithm based on images match indoor positioning.
Above-mentioned purpose of the present invention is realized by following technical scheme: a kind of optimized algorithm based on images match indoor positioning, is characterized in that: comprise the following steps:
(1) get shop sample: Wifi coverage radius is 10 ~ 15 meters, get several shop samples of this wifi coverage;
(2) take pictures in shop: three reference base picture are taken in each shop, comprise left view, front elevation, right view;
(3) shop numbering: the reference base picture in every shop is numbered;
(4) wifiMAC address and shop picture is bound: the picture number in several shops covered in each wifiMAC address and this wifi is bound;
(5) building database: building database (identification information storehouse), storage wifi information, picture sequence numbers, pictorial information that wifi is corresponding;
(6) wifi that the number of winning the confidence is the strongest: utilize mobile phone to detect wifi signal intensity, the wifi that the number of winning the confidence is the strongest, the object that the shop feature image of its correspondence travels through as images match;
(7) images match traversal is carried out.
The present invention's advantage is compared with prior art: the present invention is by detecting wifi intensity, and reduce the matching range of image, thus reduce unnecessary image matching operations, save time cost.Meanwhile, wifi intensity provides auxiliary, avoids the labile factor of wifi intensity on the impact of location.Save the time cost of location, and improve the stability of location, having can in the effective practical feature of mobile platform section.
This invention also has the following advantages:
1, in conjunction with wifi and images match two kinds of technology, a kind of algorithm is efficiently proposed;
2, need not server end be operated in, reduce expenses;
3, the accuracy rate of images match is improved;
4, avoid wifi signal intensity labile factor on the impact of location.
Accompanying drawing explanation
Fig. 1 is positioning flow figure of the present invention.
Fig. 2 is identification information storehouse of the present invention schematic diagram.
Fig. 3 is system flow schematic diagram of the present invention.
Embodiment
Below in conjunction with accompanying drawing, the present invention is described in further detail.
As shown in Figure 1, a kind of optimized algorithm based on images match indoor positioning, is characterized in that: comprise the following steps:
(1) get shop sample: Wifi coverage radius is 10 ~ 15 meters, get 6 shop samples of this wifi coverage;
(2) take pictures in shop: three reference base picture are taken in each shop, comprise left view, front elevation, right view;
(3) shop numbering: the reference base picture in every shop is numbered;
(4) wifiMAC address and shop picture is bound: 6 shop 18 pictures covered in each wifiMAC address and this wifi are numbered and bound;
(5) building database: building database (identification information storehouse, as shown in Figure 2), storage wifi information, picture sequence numbers, pictorial information that wifi is corresponding;
(6) wifi that the number of winning the confidence is the strongest: utilize mobile phone to detect wifi signal intensity, the wifi that the number of winning the confidence is the strongest, the object that the shop feature image of its correspondence travels through as images match;
(7) images match traversal is carried out.
As shown in Figure 1, Figure 2 and Figure 3, in concrete utilization optimized algorithm of the present invention, concrete positioning flow is: step 1: set up the StoreFront picture library that in market, all retail shops StoreFront is corresponding; Step 2: obtain the shop front picture that user takes present position at that time; Step 3: the corresponding MAC Address of wifi obtaining user's current position signal again; Step 4: pre-service is carried out to user's pictures taken; Step 5: Sift characteristic vector pickup is carried out to pre-service picture; Step 6: in identification information storehouse, each picture that this wifi binds mates; Step 7: according to matching result, the positional information in current map system determines the retail shop present position of described shooting, and presents on map.
The foregoing is only embodiments of the invention; not thereby the scope of the claims of the present invention is limited; every utilize instructions of the present invention and accompanying drawing content to do equivalent structure or equivalent flow process conversion; or be directly or indirectly used in other relevant technical fields, be all in like manner included in scope of patent protection of the present invention.