TWI627433B - Indoor light positioning method and system - Google Patents
Indoor light positioning method and system Download PDFInfo
- Publication number
- TWI627433B TWI627433B TW106136257A TW106136257A TWI627433B TW I627433 B TWI627433 B TW I627433B TW 106136257 A TW106136257 A TW 106136257A TW 106136257 A TW106136257 A TW 106136257A TW I627433 B TWI627433 B TW I627433B
- Authority
- TW
- Taiwan
- Prior art keywords
- positioning
- light emitting
- emitting devices
- mobile device
- server
- Prior art date
Links
Abstract
本發明係關於一種光室內定位方法及系統,包含設置複數個光發射裝置於室內空間,複數個光發射裝置對應產生複數個訊號值,複數個訊號值彼此相異,以供識別光發射裝置;設置伺服器於室內空間,伺服器包含複數個參考訊號向量以及對應的複數個參考位置座標;設置行動裝置於室內空間之定位位置,行動裝置連接至伺服器,並感測複數個光發射裝置,於伺服器內對應產生複數個定位訊號向量;以三次方權重方式,得出k個最近鄰居演算法之最佳k值,再透過權重平均處理,得出定位位置座標值。The present invention relates to a method and system for locating a light indoor, comprising: arranging a plurality of light emitting devices in an indoor space; the plurality of light emitting devices correspondingly generating a plurality of signal values, wherein the plurality of signal values are different from each other for identifying the light emitting device; Setting a server in the indoor space, the server includes a plurality of reference signal vectors and corresponding plurality of reference position coordinates; setting a position of the mobile device in the indoor space, the mobile device is connected to the server, and sensing the plurality of light emitting devices, Correspondingly, a plurality of positioning signal vectors are generated in the server; in the cubic weighting manner, the optimal k values of the k nearest neighbor algorithms are obtained, and then the weighted average processing is performed to obtain the positioning position coordinate values.
Description
本發明是關於一種光室內定位方法及系統,尤其指透過動態k-NN機制,提高定位精確度,甚至是結合動態k-NN機制及接收信號角度定位法(AOA),進一步縮短資料搜集時間的光室內定位方法及系統。The invention relates to an optical indoor positioning method and system, in particular to improving positioning accuracy through a dynamic k-NN mechanism, and even combining dynamic k-NN mechanism and received signal angular positioning method (AOA) to further shorten data collection time. Optical indoor positioning method and system.
近年來,室內定位服務之應用與日俱增,可以廣泛的應用於許多基於位置的服務場所,例如:家中、超市、購物中心、貨倉以及有如迷宮的室內場所。精確的室內定位可以提供更精準的室內導航服務,以購物中心為例,這將有助於顧客方便找尋他們想要的商品,避免因為找不到商品而使得商品未能售出,此外,亦可以給予顧客各個商品貨架的廣告資訊,當顧客走到某一貨架前,系統會給予顧客相對應的廣告訊息,讓顧客能夠隨時隨地得到商品最新資訊,另一方面,系統可以追蹤每位顧客的位置,並且可以分析出他們的瀏覽習慣,來得知商場較受歡迎之區域,甚至可以針對不同顧客的習慣給予客製化的商品廣告。In recent years, the application of indoor positioning services has been increasing, and it can be widely applied to many location-based service places, such as homes, supermarkets, shopping centers, warehouses, and indoor spaces like labyrinths. Precise indoor positioning can provide more accurate indoor navigation services. Take the shopping center as an example. This will help customers to find the products they want and avoid the failure to sell the goods because they can't find them. The customer can give the advertisement information of each product shelf. When the customer walks to a certain shelf, the system will give the customer corresponding advertising message, so that the customer can get the latest information of the product anytime and anywhere. On the other hand, the system can track each customer's Location, and can analyze their browsing habits to learn about the more popular areas of the mall, and even to customize the merchandise advertisements for different customers' habits.
許多室內定位系統藉由使用環境中之無線射頻(Radio Frequency, RF)訊號作為定位計算依據,例如:ZigBee、Wi-Fi、Bluetooth…等無線射頻訊號,但無線射頻訊號容易受到環境干擾而降低定位精確度。此外,Wi-Fi和其他基於無線射頻的定位方法僅提供以公尺為單位的精準度,這將會使得這些方法很難適用於賣場導航和貨品廣告等許多應用。Many indoor positioning systems use radio frequency (RF) signals in the environment as positioning calculations, such as ZigBee, Wi-Fi, Bluetooth, etc., but wireless RF signals are susceptible to environmental interference and reduce positioning. Accuracy. In addition, Wi-Fi and other radio-based positioning methods only provide accuracy in meters, which makes these methods difficult to apply to many applications such as store navigation and goods advertising.
有鑑於上述習知技術的問題,本發明之目的就是在提供一種能夠具有高精確度,且不易受干擾的光室內定位方法及系統。In view of the above problems of the prior art, it is an object of the present invention to provide an optical indoor positioning method and system that can have high accuracy and is less susceptible to interference.
根據本發明之一目的,提出一種光室內定位方法,包含以下步驟:設置複數個光發射裝置於室內空間,且複數個光發射裝置對應產生複數個訊號值,複數個訊號值彼此相異,以供識別光發射裝置;於室內空間設置行動裝置,其連接至伺服器,伺服器包含複數個參考訊號向量以及對應的複數個參考位置座標;設置行動裝置於定位位置,並感測複數個光發射裝置,於伺服器內對應產生複數個定位訊號向量;計算複數個定位訊號向量以及複數個參考訊號向量間的複數個歐基里德距離,以得出對應前N筆最小歐基里德距離之N個參考位置座標;以三次方權重方式計算上一步驟所得之N筆歐基里德距離對定位位置的影響強度,將權重值大於10%之取樣數作為k個最近鄰居演算法之最佳k值;計算最佳k值所對應之複數個參考位置座標對定位位置的影響強度,得到複數個參考點權重值;以及透過權重平均處理,將參考點權重值乘上對應之參考位置座標,得出定位位置之定位位置座標值。According to an aspect of the present invention, an optical indoor positioning method is provided, which includes the steps of: setting a plurality of light emitting devices in an indoor space, and a plurality of light emitting devices correspondingly generating a plurality of signal values, wherein the plurality of signal values are different from each other, For identifying the light emitting device; setting a mobile device in the indoor space, the server is connected to the server, the server includes a plurality of reference signal vectors and corresponding plurality of reference position coordinates; setting the mobile device at the positioning position, and sensing the plurality of light emission The device generates a plurality of positioning signal vectors in the server, and calculates a plurality of Occidental distances between the plurality of positioning signal vectors and the plurality of reference signal vectors to obtain a minimum Euclidean distance corresponding to the first N strokes. N reference position coordinates; calculate the influence of the N-Euclide distance of the previous step on the positioning position by the cubic weight method, and use the number of samples with the weight value greater than 10% as the best of the k nearest neighbor algorithms. k value; calculate the influence of the plurality of reference position coordinates corresponding to the optimal k value on the positioning position, and obtain a plurality of Test sites weighting value; and a weighted average process through the reference point multiplied by the weight value corresponding to the reference position coordinates, the position of the positioning of the positioning obtained positional coordinates.
較佳地,光室內定位方法進一步結合接收信號角度定位法(AOA),運用凸透鏡成像原理,得到不同高度(z軸座標)下之定位位置座標值。Preferably, the optical indoor positioning method is further combined with the received signal angular positioning method (AOA), and the convex lens imaging principle is used to obtain the coordinate values of the positioning positions under different heights (z-axis coordinates).
較佳地,複數個定位訊號向量以及複數個參考訊號向量對應於相同的光發射裝置。Preferably, the plurality of positioning signal vectors and the plurality of reference signal vectors correspond to the same light emitting device.
較佳地,複數個光發射裝置使用開關鍵控(on-off keying)調變。Preferably, the plurality of light emitting devices use on-off keying modulation.
較佳地,行動裝置具有捲簾快門。Preferably, the mobile device has a rolling shutter.
根據本發明之另一目的,提出一種光室內定位系統,用於對室內空間進行定位,光室內定位系統包含複數個光發射裝置、行動裝置以及伺服器。複數個光發射裝置係設置於室內空間內,且複數個光發射裝置對應產生複數個訊號值,其中,複數個訊號值彼此相異,以供識別光發射裝置。行動裝置,設置於室內空間內,並偵測複數個光發射裝置。伺服器,設置於室內空間內,並與行動裝置建立通訊連結。其中,伺服器包含複數個參考訊號向量以及對應的複數個參考位置座標,透過行動裝置感測複數個光發射裝置,於伺服器內對應產生複數個定位訊號向量,並以三次方權重方式獲得k個最近鄰居演算法之最佳k值,進一步透過權重平均處理得出設置行動裝置之定位位置座標值。According to another object of the present invention, an optical indoor positioning system is provided for positioning an indoor space, and the optical indoor positioning system includes a plurality of light emitting devices, mobile devices, and servers. A plurality of light emitting devices are disposed in the indoor space, and the plurality of light emitting devices correspondingly generate a plurality of signal values, wherein the plurality of signal values are different from each other for identifying the light emitting device. The mobile device is disposed in the indoor space and detects a plurality of light emitting devices. The server is installed in the indoor space and establishes a communication link with the mobile device. The server includes a plurality of reference signal vectors and a corresponding plurality of reference position coordinates, and the plurality of light emitting devices are sensed by the mobile device, and a plurality of positioning signal vectors are correspondingly generated in the server, and obtained by the cubic weighting method. The optimal k value of the nearest neighbor algorithm is further processed by the weight averaging process to obtain the positioning position coordinate value of the set mobile device.
較佳地,伺服器具有執行接收信號角度定位法(AOA)之功能。Preferably, the server has the function of performing a received signal angular positioning method (AOA).
較佳地,複數個定位訊號向量以及複數個參考訊號向量對應於相同的光發射裝置。Preferably, the plurality of positioning signal vectors and the plurality of reference signal vectors correspond to the same light emitting device.
較佳地,複數個光發射裝置使用開關鍵控(on-off keying)調變。Preferably, the plurality of light emitting devices use on-off keying modulation.
較佳地,行動裝置具有捲簾快門。Preferably, the mobile device has a rolling shutter.
以下將參照相關圖式,說明依本發明之光室內定位方法及系統之實施例,為使便於了解,下述實施例中之相同元件以相同之符號標示來表示。The embodiments of the optical indoor positioning method and system according to the present invention will be described below with reference to the related drawings. For ease of understanding, the same elements in the following embodiments are denoted by the same reference numerals.
參照第1圖,其為根據本發明一實施例之光室內定位方法流程圖,根據本發明示例性實施例之光室內定位方法包含三個階段,第一階段為光室內定位系統的設置(S110)、第二階段為訓練階段(S120)以及第三階段為定位階段(S130)。Referring to FIG. 1 , which is a flow chart of a method for positioning an optical room according to an embodiment of the present invention, an optical indoor positioning method according to an exemplary embodiment of the present invention includes three stages, and the first stage is a setting of an optical indoor positioning system (S110). The second phase is the training phase (S120) and the third phase is the positioning phase (S130).
關於上述第一階段之光室內定位系統的設置,請參照第2圖,其為根據本發明一實施例之光室內定位系統1000方塊圖。根據本發明示例性實施例之光室內定位系統,包含光發射系統100、行動裝置200以及伺服器300,其中,行動裝置200通訊連結伺服器300。With regard to the arrangement of the first stage optical indoor positioning system, please refer to FIG. 2, which is a block diagram of an optical indoor positioning system 1000 according to an embodiment of the present invention. An optical indoor positioning system according to an exemplary embodiment of the present invention includes a light emitting system 100, a mobile device 200, and a server 300, wherein the mobile device 200 is communicatively coupled to the server 300.
更進一步地,光發射系統100包含複數個光發射裝置101以及微控制器102。參照第3圖,其為根據本發明一實施例之開關鍵控(on-off keying)調變示意圖,複數個光發射裝置101係設置於室內空間中座標已知之位置,透過微控制器102對複數個光發射裝置101進行OOK(on-off keying) 調變,使複數個光發射裝置101交替明滅而形成閃爍現象。其中,OOK(on-off keying) 調變係定義於可見光無線通訊協議的IEEE 802.15.7(參考文獻:IEEE Standard Association, “IEEE Std. for local and metropolitan area networks - part 15.7: Short-rang wireless optical communication using visible light,” IEEE Computer Society, 2011.) 。Further, the light emitting system 100 includes a plurality of light emitting devices 101 and a microcontroller 102. Referring to FIG. 3, which is a schematic diagram of on-off keying modulation according to an embodiment of the present invention, a plurality of light-emitting devices 101 are disposed at known positions in the indoor space, and are transmitted through the microcontroller 102. The plurality of light-emitting devices 101 perform OOK (on-off keying) modulation, and the plurality of light-emitting devices 101 are alternately turned off to form a flicker phenomenon. Among them, OOK (on-off keying) modulation is defined in IEEE 802.15.7 of visible light wireless communication protocol (Reference: IEEE Standard Association, "IEEE Std. for local and metropolitan area networks - part 15.7: Short-rang wireless optical Communication using visible light,” IEEE Computer Society, 2011.).
在一實施例中,光發射裝置101可以發出可見光,提供室內照明,其閃爍頻率高於人眼所能感知的頻率,以避免對人眼造成不舒服的感覺。具體地,光發射裝置101可以是 LED,因為LED的反應時間非常短,可以實現非常高的閃爍頻率(例如1000 Hz),使得人眼無法察覺。In one embodiment, the light emitting device 101 can emit visible light, providing room illumination with a frequency that is higher than the frequency perceived by the human eye to avoid an uncomfortable feeling to the human eye. Specifically, the light-emitting device 101 may be an LED because the reaction time of the LED is very short, and a very high flicker frequency (for example, 1000 Hz) can be realized, making it impossible for the human eye to perceive.
在本實施例中,行動裝置200可設置影像調整模組201、影像擷取模組202、第一通訊模組203及CMOS感測元件204。其中,行動裝置200設置於室內空間,用以對複數個光發射裝置101進行感測及拍攝。CMOS感測元件204具有捲簾快門(rolling shutter)。影像調整模組201可調整影像擷取參數,例如曝光時間和光感度(ISO),來獲得較佳的影像品質。影像擷取模組202則可利用捲簾快門的垂直或水平掃描特性,讓行動裝置200上的像素點逐行曝光。In this embodiment, the mobile device 200 can be provided with an image adjustment module 201, an image capture module 202, a first communication module 203, and a CMOS sensing component 204. The mobile device 200 is disposed in the indoor space for sensing and capturing a plurality of light emitting devices 101. The CMOS sensing element 204 has a rolling shutter. The image adjustment module 201 can adjust image capture parameters such as exposure time and light sensitivity (ISO) to obtain better image quality. The image capture module 202 can utilize the vertical or horizontal scanning characteristics of the rolling shutter to expose the pixel points on the mobile device 200 row by row.
根據本發明一實施例,參照第4圖,其為根據本發明一實施例之開關鍵控以及捲簾快門示意圖,將光發射裝置101進行OOK(on-off keying) 調變,並且以具有捲簾快門的行動裝置200對複數個光發射裝置101進行感測及拍攝,並透過影像調整模組201適當地調整曝光時間及ISO值,則行動裝置200拍攝的影像呈現明暗間隔的條紋(如第4圖所示)。參照第5圖的(a)部分及第5圖的(b)部分,其為根據本發明一實施例之光發射裝置設置照片,具有捲簾快門的行動裝置200在如第5圖的(a)部分之環境下,可拍攝到如第5圖的(b)部分之影像。並且透過第一通訊模組203,進一步將上述條紋影像由行動裝置200傳送至伺服器300。According to an embodiment of the present invention, reference is made to FIG. 4, which is a schematic diagram of an opening key control and a rolling shutter according to an embodiment of the present invention, and the light emitting device 101 is subjected to OOK (on-off keying) modulation, and has a volume. The action device 200 of the curtain shutter senses and photographs a plurality of light emitting devices 101, and appropriately adjusts the exposure time and the ISO value through the image adjusting module 201, and the image captured by the mobile device 200 presents stripes of light and dark intervals (eg, Figure 4). Referring to part (a) of Fig. 5 and part (b) of Fig. 5, which is a photograph of a light-emitting device according to an embodiment of the present invention, a mobile device 200 having a rolling shutter is as shown in Fig. 5 (a) In part of the environment, images of part (b) of Figure 5 can be captured. And the stripe image is further transmitted from the mobile device 200 to the server 300 through the first communication module 203.
伺服器300內建了影像處理模組301、影像訊息擷取模組302、計算模組303、資料庫304和第二通訊模組305。伺服器300透過影像處理模組301以及影像訊息擷取模組302來取得該影像所包含的訊息。進一步來說,透過影像處理模組301以及影像訊息擷取模組302,上述條紋影像被以傅立葉轉換(Fourier transform)運算以得其閃爍頻率值。根據本發明另一實施例,每一個光發射裝置101的閃爍頻率可以是彼此相異的,透過行動裝置200拍攝的影像中的每一個光發射裝置101所呈現的明暗條紋間距也不同,也就是說,透過傅立葉轉換(Fourier transform)運算後所得之閃爍頻率值不同。因此,得以用這些相異的閃爍頻率值來識別每一個光發射裝置101,並對已知位置的每一個光發射裝置101進行編號。The server 300 has an image processing module 301, an image message capturing module 302, a computing module 303, a database 304, and a second communication module 305. The server 300 obtains the information included in the image through the image processing module 301 and the image message capturing module 302. Further, through the image processing module 301 and the image message capturing module 302, the fringe image is computed by Fourier transform to obtain a flicker frequency value. According to another embodiment of the present invention, the flicker frequencies of each of the light emitting devices 101 may be different from each other, and each of the light emitting devices 101 imaged by the mobile device 200 may have different light and dark stripe pitches, that is, It is said that the value of the flicker frequency obtained by the Fourier transform operation is different. Therefore, it is possible to identify each of the light-emitting devices 101 with these different flicker frequency values, and number each of the light-emitting devices 101 at known positions.
具體地,行動裝置200可以是諸如智慧型手機或平板電腦等手持設備,其具有CMOS感測元件。In particular, the mobile device 200 can be a handheld device such as a smart phone or tablet that has CMOS sensing elements.
根據本發明一實施例,光發射裝置101在上述影像中的數量若少於3個,則該影像將不進行下面將描述的定位演算法的分析。換言之,只有當影像中具有3個或大於3個光發射裝置101時,該影像才屬於可用影像。According to an embodiment of the present invention, if the number of the light emitting devices 101 in the image is less than three, the image will not be analyzed by the positioning algorithm to be described below. In other words, the image belongs to the available image only when there are 3 or more light emitting devices 101 in the image.
進一步地,為了以更精確的資訊做有效的定位,影像處裡模組301以及影像訊息擷取模組302從影像中擷取每一個光發射裝置101的輪廓,對輪廓所圍的區域算出重心位置,作為該光發射裝置101在影像中的影像座標,並計算影像中的每個光發射裝置101與影像中心的距離。Further, in order to perform effective positioning with more accurate information, the image in-camera module 301 and the image information capturing module 302 extract the contour of each light-emitting device 101 from the image, and calculate the center of gravity of the region surrounded by the contour. The position is used as an image coordinate of the light emitting device 101 in the image, and the distance between each of the light emitting devices 101 in the image and the center of the image is calculated.
參照第6圖,其為根據本發明一實施例之參考影像示意圖,行動裝置200拍攝到的光發射裝置101數量為5,由於光發射裝置101數量大於3,故第6圖代表一可用影像,且每一個光發射裝置101影像之重心位置代表其影像座標。Referring to FIG. 6 , which is a schematic diagram of a reference image according to an embodiment of the present invention, the number of light emitting devices 101 captured by the mobile device 200 is 5. Since the number of light emitting devices 101 is greater than 3, FIG. 6 represents an available image. And the position of the center of gravity of each image of the light emitting device 101 represents its image coordinates.
參照第7圖,其為根據本發明一實施例之參考點設置平面圖,在垂直座標z=197之水平平面上,以10 cm為間距設置了複數個參考點,複數個參考點之座標 F j 為已知,其中,j為參考點編號。例如:在第7圖中,箭頭所指該參考點位於座標 F j , F j =(10,10,197)。 Referring to FIG 7, which is in accordance with a coordinate F j reference point of the reference point Example of an embodiment of the present invention is provided a plan view of the vertical coordinate z = the level 197 of the plane, to 10 cm provided with a plurality of reference points for the pitches, the complex It is known that j is a reference point number. For example, in Figure 7, the reference point indicated by the arrow is located at coordinates F j , F j = (10, 10, 197).
重新參照第1圖,根據本發明之光室內定位方法,其第二階段為訓練階段,係在如第7圖所示之每一個參考點上以行動裝置200對複數個光發射裝置101進行感測及拍攝,對於拍攝到的光發射裝置101數量大於或等於3個的影像,才屬於可用影像。應當注意,每個參考點間距離10 cm僅作為示例,不應侷限於此。應當注意,根據此一示例性實施例設置參考點的場地長度為1.2米,寬度為1米,此場地大小僅為示例,不應侷限於此。Referring back to FIG. 1 , the second stage of the optical indoor positioning method according to the present invention is a training stage in which a plurality of light emitting devices 101 are sensed by the mobile device 200 at each reference point as shown in FIG. For measurement and shooting, the number of images of the light emitting device 101 that are captured is greater than or equal to three, which are available images. It should be noted that the distance between each reference point of 10 cm is merely an example and should not be limited thereto. It should be noted that the field length at which the reference point is set according to this exemplary embodiment is 1.2 meters and the width is 1 meter. This site size is merely an example and should not be limited thereto.
在訓練階段拍攝到的可用影像將透過影像處理模組301以及影像訊息擷取模組302進行分析,並將分析所得資料儲存於資料庫304。The available images captured during the training phase are analyzed by the image processing module 301 and the image message capture module 302, and the analyzed data is stored in the database 304.
接下來,為了更清楚理解上述透過影像處理模組301以及影像訊息擷取模組302進行分析的細節,作為示例,將進一步對第6圖所代表影像進行分析:Next, in order to more clearly understand the details of the analysis performed by the image processing module 301 and the image message capturing module 302, as an example, the image represented by FIG. 6 will be further analyzed:
參照第6圖,以影像中心作為影像座標原點(x=0, y=0),得到複數個光發射裝置101的影像重心所位於的影像座標,並對應得出複數個參考訊號向量 ji (x i ,y i ),其中,i為複數個光發射裝置101的編號,j為參考點編號。 Referring to FIG. 6 , the image center is used as the image coordinate origin (x=0, y=0), and the image coordinates of the image center of gravity of the plurality of light emitting devices 101 are obtained, and a plurality of reference signal vectors are obtained correspondingly. Ji (x i , y i ), where i is the number of the plurality of light-emitting devices 101, and j is the reference point number.
具體地,第6圖為當行動裝置200設置於j=5號參考點時,對該參考點上方複數個光發射裝置101進行感測及拍攝所得之參考影像示意圖,以參考影像中心作為影像座標原點(x=0, y=0),對於i=1之光發射裝置101而言,其在此參考影像中之影像座標(重心座標)為(x=307, y=384),對應之參考訊號向量 ji(x i,y i)= 51(x 1,y 1)=(307,384);對於i=3之光發射裝置101而言,其在此參考影像中之影像座標(重心座標)為(x=-402, y=-512),對應之參考訊號向量 ji(x i,y i)= 53(x 3,y 3)=( -402, -512)。 Specifically, FIG. 6 is a schematic diagram of a reference image obtained by sensing and photographing a plurality of light emitting devices 101 above the reference point when the mobile device 200 is disposed at the j=5 reference point, and the reference image center is used as the image coordinate. The origin (x=0, y=0), for the light emitting device 101 with i=1, the image coordinates (center of gravity coordinates) in the reference image is (x=307, y=384), corresponding to Reference signal vector Ji (x i ,y i )= 51 (x 1 , y 1 )=(307,384); for the light emitting device 101 with i=3, the image coordinates (center of gravity coordinates) in the reference image are (x=-402, y=-512). Corresponding reference signal vector Ji (x i ,y i )= 53 (x 3 , y 3 )=( -402, -512).
在訓練階段,行動裝置200在每一個參考點上對複數個光發射裝置101進行感測和拍攝,並透過影像處理模組301和影像訊息擷取模組302對影像進行分析,再將所得之複數個參考訊號向量 ji (x i ,y i )儲存於資料庫304。 During the training phase, the mobile device 200 senses and captures a plurality of light emitting devices 101 at each reference point, and analyzes the images through the image processing module 301 and the image message capturing module 302, and then obtains the image. Multiple reference signal vectors Ji (x i , y i ) is stored in the database 304.
影像處理模組301以及影像訊息擷取模組302更進一步計算複數個參考訊號向量 ji (x i ,y i )的長度 : The image processing module 301 and the image message capturing module 302 further calculate a plurality of reference signal vectors. Length of ji (x i , y i ) :
具體地,在上述參照第6圖之實施例中,對於i=1之光發射裝置101而言,其 R 51 即為座標(307,384)與座標原點 (00)之距離。 Specifically, in the above-described embodiment with reference to Fig. 6, for the light emitting device 101 of i = 1, R 51 is the distance between the coordinates (307, 384) and the coordinate origin (00).
在複數個參考點上重複上述的感測及分析,並將複數個參考位置座標 F j 、所對應之參考影像、所對應之複數個參考訊號向量 ji (x i ,y i )和複數個參考訊號長度 儲存於資料庫304,完成本發明之光室內定位方法之第二階段。 Repeating the above sensing and analysis on a plurality of reference points, and combining a plurality of reference position coordinates F j , corresponding reference images, and corresponding plurality of reference signal vectors Ji (x i , y i ) and a plurality of reference signal lengths It is stored in the database 304 to complete the second stage of the optical indoor positioning method of the present invention.
參照第1圖,根據本發明之光室內定位方法之第三階段為定位階段(S810~ S840),參照第8圖所示,其為根據本發明一實施例之光室內定位方法之定位階段流程圖。定位階段分為兩個主要步驟,第一個步驟為本發明技術特徵之一之動態k-NN機制(S820),第二個步驟為權重計算(S830)。Referring to FIG. 1 , the third stage of the optical indoor positioning method according to the present invention is a positioning stage (S810 to S840), which is shown in FIG. 8 , which is a positioning stage flow of the optical indoor positioning method according to an embodiment of the present invention. Figure. The positioning phase is divided into two main steps, the first step is a dynamic k-NN mechanism (S820) which is one of the technical features of the present invention, and the second step is a weight calculation (S830).
在動態k-NN機制中,又包含三個子步驟: (1) 訊號過濾及比對(S821); (2) 計算歐基里德距離(S822);以及 (3) 決定最佳k值(S823)。In the dynamic k-NN mechanism, there are three sub-steps: (1) signal filtering and comparison (S821); (2) calculating the Euclid distance (S822); and (3) determining the optimal k value (S823) ).
接下來將依序對上述三個子步驟進行詳細說明:The following three sub-steps will be described in detail:
(1) 訊號過濾及比對:進行定位時,首先將行動裝置200設置於所欲定位之定位位置,對複數個光發射裝置101進行感測及拍攝,以獲得可用的定位影像(拍攝到的光發射裝置101數量大於或等於3個)。定位影像同樣地在影像處理模組301以及影像訊息擷取模組302進行分析:參照第9圖,其為根據本發明一實施例之定位影像示意圖,以定位影像中心作為影像座標原點(x=0, y=0),得到複數個光發射裝置101的影像重心所位於的影像座標,以及對應的複數個定位訊號向量 i (x i ,y i )和複數個定位訊號長度 M i 。接下來,根據定位影像所包含的資訊,將定位影像與包含了定位影像所拍攝到的複數個光發射裝置101中的每一個的參考影像進行資料比對。例如,當定位影像包含了編號1(i=1)、編號2(i=2)以及編號3(i=3)的光發射裝置101,參考影像包含了編號1(i=1)、編號2(i=2) 、編號3(i=3)以及編號4(i=4)的光發射裝置101,則此參考影像將與定位影像進行比對,也就是說,此參考影像屬於有效參考影像;若定位影像包含了編號1(i=1)、編號2(i=2) 以及編號3(i=3)的光發射裝置101,參考影像包含了編號1(i=1)、編號2(i=2) 、編號4(i=4)以及編號5(i=5)的光發射裝置101,由於此參考影像缺漏了編號3(i=3)的光發射裝置101,因此,將不取樣此參考影像與定位影像進行比對,也就是說,此參考影像不屬於有效參考影像。 (1) Signal filtering and comparison: When positioning is performed, the mobile device 200 is first set at the positioning position to be positioned, and a plurality of light emitting devices 101 are sensed and photographed to obtain a usable positioning image (photographed The number of light emitting devices 101 is greater than or equal to three). The positioning image is similarly analyzed by the image processing module 301 and the image message capturing module 302. Referring to FIG. 9, a schematic image of the positioning image according to an embodiment of the present invention is used to locate the image center as the image coordinate origin (x). =0, y=0), the image coordinates of the image center of gravity of the plurality of light emitting devices 101 are obtained, and the corresponding plurality of positioning signal vectors are obtained. i (x i , y i ) and a plurality of positioning signal lengths M i . Next, based on the information included in the positioning image, the positioning image is compared with the reference image of each of the plurality of light emitting devices 101 captured by the positioning image. For example, when the positioning image includes the light emitting device 101 of number 1 (i=1), number 2 (i=2), and number 3 (i=3), the reference image includes number 1 (i=1), number 2 (i=2), number 3 (i=3), and number 4 (i=4) of the light emitting device 101, the reference image will be compared with the positioning image, that is, the reference image belongs to the effective reference image. If the positioning image includes the light emitting device 101 numbered 1 (i=1), number 2 (i=2), and number 3 (i=3), the reference image contains the number 1 (i=1), number 2 ( i=2), number 4 (i=4), and number 5 (i=5) of the light-emitting device 101, since the reference image lacks the light-emitting device 101 of number 3 (i=3), therefore, no sampling is performed. This reference image is compared with the positioning image, that is, the reference image is not a valid reference image.
(2) 計算歐基里德距離:在計算模組303中,計算有效參考影像中的參考訊號向量 ji (x i ,y i )與定位影像中的定位訊號向量 i (x i ,y i )的歐基里德距離 : 其中, M i 為定位訊號向量 i (x i ,y i )的長度, R ji 為參考訊號向量 ji (x i ,y i )的長度,i為定位影像所包含的光發射裝置101的編號。 (2) Calculating the Euclidean distance: In the calculation module 303, calculating the reference signal vector in the effective reference image Ji (x i , y i ) and the positioning signal vector in the positioning image Euclid distance of i (x i , y i ) : Where M i is a positioning signal vector The length of i (x i , y i ), R ji is the reference signal vector The length of ji (x i , y i ), i is the number of the light-emitting device 101 included in the positioning image.
具體地,例如,當定位影像包含了編號1(i=1)、編號2(i=2)以及編號3(i=3)的光發射裝置101,參考影像包含了編號1(i=1)、編號2(i=2) 、編號3(i=3)以及編號4(i=4)的光發射裝置101,則上述歐基里德距離公式將計算i=1、i=2以及i=3之貢獻。Specifically, for example, when the positioning image includes the light emitting device 101 of the number 1 (i=1), the number 2 (i=2), and the number 3 (i=3), the reference image includes the number 1 (i=1). , the number 2 (i=2), the number 3 (i=3), and the number 4 (i=4) of the light-emitting device 101, the above-mentioned Euclid distance formula will calculate i=1, i=2, and i= 3 contribution.
具體地,若對應定位影像之有效參考影像數量為5,則會得到5比歐基里德距離 。例如,當定位影像包含了編號1(i=1)、編號2(i=2)、編號3(i=3)以及編號5(i=5)的光發射裝置101,且有效參考影像數量為5,即是代表有五張參考影像皆包含了編號1(i=1)、編號2(i=2)、編號3(i=3)以及編號5(i=5)的光發射裝置101,上述五張參考影像代表在五個不同參考點上之拍攝,且上述五張參考影像將各別與定位影像進行如上述歐基里德距離 分析。也就是說,具有至少5個參考點接近該定位位置。例如,有效參考影像之j=1、j=2、 j=4、j=6以及 j=7,代表該定位位置接近1號參考點、2號參考點、4號參考點、6號參考點以及7號參考點。 Specifically, if the number of effective reference images corresponding to the positioned image is 5, a 5-by-Euclidean distance is obtained. . For example, when the positioning image includes the light emitting device 101 numbered 1 (i=1), number 2 (i=2), number 3 (i=3), and number 5 (i=5), and the number of valid reference images is 5, that is, the light-emitting device 101 is represented by five reference images including number 1 (i=1), number 2 (i=2), number 3 (i=3), and number 5 (i=5). The above five reference images represent shots at five different reference points, and the above five reference images will be individually and positioned to perform the above-mentioned Euclid distance. analysis. That is, there are at least 5 reference points close to the positioning position. For example, j=1, j=2, j=4, j=6, and j=7 of the effective reference image, indicating that the positioning position is close to the reference point No. 1, the reference point No. 2, the reference point No. 4, and the reference point No. 6 And reference point No. 7.
(3) 決定最佳k值:將步驟(2)所得之歐基里德距離 由小至大排序,並將前N筆歐基里德距離以下列三次方權重方式計算,用以拉大各個影響強度的對比: 其中,j為前N筆歐基里德距離所對應的參考點編號, 為由步驟(2)所得的歐基里德距離, 為j號參考點對定位位置的影響強度百分比。 (3) Determine the best k value: the Euclid distance obtained in step (2) Sort from small to large, and calculate the former N-Euclide distance in the following cubic weights to increase the contrast of each impact intensity: Where j is the reference point number corresponding to the top N Euclidean distance, For the Euclid distance obtained by step (2), The percentage of the intensity of the influence of the reference point of j on the position of the position.
較佳地,N值為3,更佳地,N值為4或5。 取樣 大於10 %的資料,捨棄 小於10 %的資料,以上述取樣數作為本發明提出之動態k-NN機制的最佳k值。 Preferably, the N value is 3, and more preferably, the N value is 4 or 5. sampling More than 10% of the data, discard For less than 10% of the data, the above sample number is used as the optimal k value of the dynamic k-NN mechanism proposed by the present invention.
具體地,例如,由步驟(2)所得之歐基里德距離 分別為 以及 (由小至大排列),當N=4,代表只取 進行上述三次方權重計算,因此上述三次方權重公式中,將只計算j=1、j=4、j=2以及j=6之權重比例。 Specifically, for example, the Euclid distance obtained by the step (2) Separately as well as (from small to large), when N=4, it means only The above-described cubic weight calculation is performed, so in the above-described cubic weight formula, only the weight ratios of j=1, j=4, j=2, and j=6 will be calculated.
具體地,參照表1及表2,根據本發明之一示例性實施例,提供對上述動態k-NN機制更詳細的理解。在表1中,其N=4,但只有第1個參考點的影響強度比例大於10 %,故此實施例之動態k-NN的k值為1;以及在表2中,其N=4,但四個參考點的影響強度比例皆大於10 %,故此實施例之動態k-NN的k值為4。 <TABLE border="1" borderColor="#000000" width="85%"><TBODY><tr><td> 參考位置座標F<sub>j</sub></td><td> 歐基里德距離E<sub>j</sub></td><td> S<sub>j</sub>排序 </td><td> S<sub>j</sub> % </td></tr><tr><td> F<sub>1</sub>= (10, 10, 197) </td><td> 10.38457 </td><td> 1 </td><td> 99.23921 </td></tr><tr><td> F<sub>2</sub>= (10, 0, 197) </td><td> 57.80492 </td><td> 2 </td><td> 0.57538 </td></tr><tr><td> F<sub>3</sub>= (0, 10, 197) </td><td> 105.19949 </td><td> 3 </td><td> 0.095457 </td></tr><tr><td> F<sub>4</sub>= (20, 10, 197) </td><td> 107.30282 </td><td> 4 </td><td> 0.089953 </td></tr></TBODY></TABLE>表1 <TABLE border="1" borderColor="#000000" width="85%"><TBODY><tr><td> 參考位置座標F<sub>j</sub></td><td> 歐基里德距離E<sub>j</sub></td><td> S<sub>j</sub>排序 </td><td> S<sub>j</sub> % </td></tr><tr><td> F<sub>2</sub> =(10, 0, 197) </td><td> 74.1151 </td><td> 1 </td><td> 37.16849 </td></tr><tr><td> F<sub>5</sub> = (0, 0, 197) </td><td> 87.44512 </td><td> 2 </td><td> 22.6302 </td></tr><tr><td> F<sub>3</sub> = (0, 10, 197) </td><td> 87.9542 </td><td> 3 </td><td> 22.23951 </td></tr><tr><td> F<sub>1</sub> = (10, 10, 197) </td><td> 94.4456 </td><td> 4 </td><td> 17.9618 </td></tr></TBODY></TABLE>表2 Specifically, with reference to Tables 1 and 2, a more detailed understanding of the dynamic k-NN mechanism described above is provided in accordance with an exemplary embodiment of the present invention. In Table 1, N=4, but only the influence intensity ratio of the first reference point is greater than 10%, so the k value of the dynamic k-NN of this embodiment is 1; and in Table 2, N=4, However, the intensity ratio of the four reference points is greater than 10%, so the k value of the dynamic k-NN of this embodiment is 4. <TABLE border="1" borderColor="#000000" width="85%"><TBODY><tr><td> Reference position coordinates F<sub>j</sub></td><td> Reed distance E<sub>j</sub></td><td> S<sub>j</sub> sorting</td><td> S<sub>j</sub> % </td> </tr><tr><td> F<sub>1</sub>= (10, 10, 197) </td><td> 10.38457 </td><td> 1 </td><td> 99.23921 </td></tr><tr><td> F<sub>2</sub>= (10, 0, 197) </td><td> 57.80492 </td><td> 2 </ Td><td> 0.57538 </td></tr><tr><td> F<sub>3</sub>= (0, 10, 197) </td><td> 105.19949 </td>< Td> 3 </td><td> 0.095457 </td></tr><tr><td> F<sub>4</sub>= (20, 10, 197) </td><td> 107.30282 </td><td> 4 </td><td> 0.089953 </td></tr></TBODY></TABLE> Table 1 <TABLE border="1" borderColor="#000000" width="85%"><TBODY><tr><td> Reference position coordinates F<sub>j</sub></td><td> Reed distance E<sub>j</sub></td><td> S<sub>j</sub> sorting</td><td> S<sub>j</sub> % </td> </tr><tr><td> F<sub>2</sub> =(10, 0, 197) </td><td> 74.1151 </td><td> 1 </td><td> 37.16849 </td></tr><tr><td> F<sub>5</sub> = (0, 0, 197) </td><td> 87.44512 </td><td> 2 </ Td><td> 22.6302 </td></tr><tr><td> F<sub>3</sub> = (0, 10, 197) </td><td> 87.9542 </td>< Td> 3 </td><td> 22.23951 </td></tr><tr><td> F<sub>1</sub> = (10, 10, 197) </td><td> 94.4456 </td><td> 4 </td><td> 17.9618 </td></tr></TBODY></TABLE> Table 2
接下來,請重新參照第8圖,對定位階段的權重計算部分做說明。Next, please refer to Figure 8 again to explain the weight calculation part of the positioning stage.
權重計算部分同樣包括兩個子步驟: (1) 計算權重值(S831);以及 (2) 計算定位位置座標P(S832)。The weight calculation section also includes two sub-steps: (1) calculating the weight value (S831); and (2) calculating the positional position coordinate P (S832).
接下來將依序對上述兩個子步驟進行詳細說明:The following two substeps will be described in detail:
(1) 計算權重值:對上述動態k-NN的k值所對應之j號參考點進行權重計算,根據以下的公式得出參考點權重值 : 其中,i為動態k-NN的k值所對應之每一個參考點編號, 為由動態k-NN機制之子步驟(2)所得的歐基里德距離, 為j號參考點對定位位置的影響權重。 (1) Calculating the weight value: performing weight calculation on the reference point j corresponding to the k value of the dynamic k-NN, and obtaining the reference point weight value according to the following formula : Where i is the number of each reference point corresponding to the k value of the dynamic k-NN, The Euclid distance obtained by substep (2) of the dynamic k-NN mechanism, The weight of the influence of the reference point of j on the positioning position.
(2) 計算定位位置座標P:計算參考位置座標 F j 及對應的權重,得出定位位置座標P,完成定位: (2) Calculate the positioning position coordinate P: Calculate the reference position coordinate F j and the corresponding weight, and obtain the positioning position coordinate P to complete the positioning:
具體地,參照表3及表4,提供對上述權重運算更詳細的理解。在表3中(對應於表1),其動態k-NN的k值為1,也就是說,步驟(1)參考點權重公式之j=1,因此,第1個參考點的權重為100 %,故計算所得之定位位置座標P即為第1個參考點的座標,P=(10, 10, 197);以及在表4中(對應於表2,行動裝置200設置於(5, 5, 197)),其動態k-NN的k值為4,也就是說,步驟(1)參考點權重公式之j=2、5、3以及1,故定位位置座標P由四個參考點的權重貢獻決定, P=(5.136, 4.683, 197)。 <TABLE border="1" borderColor="#000000" width="85%"><TBODY><tr><td> 參考位置座標F<sub>j</sub></td><td> 歐基里德距離E<sub>j</sub></td><td><i>W</i>排序 </td><td><i>W</i></td></tr><tr><td> F<sub>1</sub>= (10, 10, 197) </td><td> 10.38457 </td><td> 1 </td><td> 1 </td></tr><tr><td> F<sub>2</sub>= (10, 0, 197) </td><td> 57.80492 </td><td> 2 </td><td> - </td></tr><tr><td> F<sub>3</sub>= (0, 10, 197) </td><td> 105.19949 </td><td> 3 </td><td> - </td></tr><tr><td> F<sub>4</sub>= (20, 10, 197) </td><td> 107.30282 </td><td> 4 </td><td> - </td></tr></TBODY></TABLE>表3 <TABLE border="1" borderColor="#000000" width="85%"><TBODY><tr><td> 參考位置座標F<sub>j</sub></td><td> 歐基里德距離E<sub>j</sub></td><td><i>W</i>排序 </td><td><i>W</i></td></tr><tr><td> F<sub>2</sub> =(10, 0, 197) </td><td> 74.1151 </td><td> 1 </td><td> 0.28777 </td></tr><tr><td> F<sub>5</sub> = (0, 0, 197) </td><td> 87.44512 </td><td> 2 </td><td> 0.24391 </td></tr><tr><td> F<sub>3</sub> = (0, 10, 197) </td><td> 87.9542 </td><td> 3 </td><td> 0.24249 </td></tr><tr><td> F<sub>1</sub> = (10, 10, 197) </td><td> 94.4456 </td><td> 4 </td><td> 0.22583 </td></tr></TBODY></TABLE>表4 Specifically, a more detailed understanding of the weight calculation described above is provided with reference to Tables 3 and 4. In Table 3 (corresponding to Table 1), the k value of the dynamic k-NN is 1, that is, the reference point weight formula of step (1) is j=1, therefore, the weight of the first reference point is 100. %, so the calculated position position coordinate P is the coordinate of the first reference point, P = (10, 10, 197); and in Table 4 (corresponding to Table 2, the mobile device 200 is set at (5, 5) , 197)), its dynamic k-NN has a k value of 4, that is, step (1) reference point weight formula j=2, 5, 3, and 1, so the positional position coordinate P is composed of four reference points. The weight contribution decision, P = (5.136, 4.683, 197). <TABLE border="1" borderColor="#000000" width="85%"><TBODY><tr><td> Reference position coordinates F<sub>j</sub></td><td> Reed is sorted by E<sub>j</sub></td><td><i>W</i></td><td><i>W</i></td></tr ><tr><td> F<sub>1</sub>= (10, 10, 197) </td><td> 10.38457 </td><td> 1 </td><td> 1 </ Td></tr><tr><td> F<sub>2</sub>= (10, 0, 197) </td><td> 57.80492 </td><td> 2 </td>< Td> - </td></tr><tr><td> F<sub>3</sub>= (0, 10, 197) </td><td> 105.19949 </td><td> 3 </td><td> - </td></tr><tr><td> F<sub>4</sub>= (20, 10, 197) </td><td> 107.30282 </td ><td> 4 </td><td> - </td></tr></TBODY></TABLE>Table 3 <TABLE border="1" borderColor="#000000" width="85%"><TBODY><tr><td> Reference position coordinates F<sub>j</sub></td><td> Reed is sorted by E<sub>j</sub></td><td><i>W</i></td><td><i>W</i></td></tr ><tr><td> F<sub>2</sub> =(10, 0, 197) </td><td> 74.1151 </td><td> 1 </td><td> 0.28777 </ Td></tr><tr><td> F<sub>5</sub> = (0, 0, 197) </td><td> 87.44512 </td><td> 2 </td>< Td> 0.24391 </td></tr><tr><td> F<sub>3</sub> = (0, 10, 197) </td><td> 87.9542 </td><td> 3 </td><td> 0.24249 </td></tr><tr><td> F<sub>1</sub> = (10, 10, 197) </td><td> 94.4456 </td ><td> 4 </td><td> 0.22583 </td></tr></TBODY></TABLE> Table 4
在上述的定位方法中,在計算模組303中,計算歐基里德距離、動態k-NN機制、權重計算以及計算定位位置座標P,最後再透過第二通訊模組305將所得之定位位置座標P傳回行動裝置200,顯示於行動裝置200。In the above positioning method, in the calculation module 303, the Euclid distance, the dynamic k-NN mechanism, the weight calculation, and the calculation of the positioning position coordinate P are calculated, and finally the obtained positioning position is obtained through the second communication module 305. The coordinate P is transmitted back to the mobile device 200 and displayed on the mobile device 200.
在上述的定位方法中,行動裝置200在訓練階段以及定位階段的z軸方向高度是固定的,即屬於高度固定下的三維定位方法。In the above positioning method, the height of the mobile device 200 in the z-axis direction during the training phase and the positioning phase is fixed, that is, the three-dimensional positioning method under the fixed height.
較佳地,本發明之光室內定位方法進一步結合動態k-NN機制及接收信號角度定位法(AOA),對於高度不固定的三維定位需求同樣提供了精確的定位。Preferably, the optical indoor positioning method of the present invention further combines the dynamic k-NN mechanism and the received signal angular positioning method (AOA) to provide accurate positioning for highly unfixed three-dimensional positioning requirements.
參照第10圖,其為根據習知技術之凸透鏡成像示意圖,當物距( H R 及 H M )改變,成像大小( 及 M i )存在一比例關係,其比例參數為 Q: Referring to Fig. 10, it is a schematic diagram of convex lens imaging according to the prior art, when the object distance ( H R and H M ) is changed, the imaging size ( And M i ) has a proportional relationship, and the proportional parameter is Q :
具體地,行動裝置200在物距(行動裝置200與光發射裝置101的垂直距離)為 H R 之水平平面進行訓練階段之參考影像拍攝,並分析參考訊號向量 ji (x i ,y i )和參考訊號長度 ;在定位階段,行動裝置200的物距為 H M ( H M 可以與 H R 相異),定位訊號向量 i (x i ,y i )之定位訊號長度為 M i ,透過上述比例關係之比例參數 Q,可運算得出在物距 H R 的定位訊號長度 ,便能夠計算定位訊號長度 與在物距 H R 拍攝及分析所得之參考訊號長度 間的歐基里德距離,進行如上所述之動態k-NN機制下的定位。 Specifically, the mobile device 200 performs reference imaging of the training phase in the horizontal plane of the object distance (the vertical distance of the mobile device 200 from the light emitting device 101) is H R , and analyzes the reference signal vector. Ji (x i , y i ) and reference signal length In the positioning phase, the object distance of the mobile device 200 is H M ( H M can be different from H R ), and the positioning signal vector The position signal length of i (x i , y i ) is M i , and the length of the positioning signal at the object distance H R can be calculated by using the proportional parameter Q of the above proportional relationship. , can calculate the length of the positioning signal Reference signal length obtained by shooting and analyzing at object distance H R The inter-Euclidean distance is located under the dynamic k-NN mechanism as described above.
根據本發明另一示例性實施例,參照表5,提供對上述結合動態k-NN機制及接收信號角度定位法(AOA)更詳細的理解。在表5中,行動裝置200設置於(10, 10, 185),透過AOA(參考資料:Ye-Sheng Kuo, Pat Pannuto, Ko-Jen Hsiao, and Prabal Dutta. 2014. Luxapose: indoor positioning with mobile phones and visible light. In Proceedings of the 20th annual international conference on Mobile computing and networking (MobiCom '14). ACM, New York, NY, USA, 447-458.),可得其z=184.64184= H M ,對於參考訊號向量 ji (x i ,y i )資料集之 H R 為197來說,其比例關係為: <TABLE border="1" borderColor="#000000" width="85%"><TBODY><tr><td> i </td><td> 1 </td><td> 2 </td><td> 3 </td><td> 4 </td><td> 5 </td></tr><tr><td><i>M<sub>i</sub></i></td><td> 525.02381 </td><td> 628.68514 </td><td> 703.2951 </td><td> 616.40409 </td><td> 95.67131 </td></tr><tr><td><img wi="27" he="22" file="TWI627433B_D0036.tif" img-format="jpg"></img></td><td> 492.08802 </td><td> 589.24646 </td><td> 659.17599 </td><td> 577.73583 </td><td> 89.66966 </td></tr><tr><td> R<sub>ji</sub></td><td> 499.76895 </td><td> 579.10448 </td><td> 651.87806 </td><td> 587.13286 </td><td> 82.03658 </td></tr><tr><td> E<sub>j</sub>=19.018;k=1 </td></tr></TBODY></TABLE>表5 In accordance with another exemplary embodiment of the present invention, with reference to Table 5, a more detailed understanding of the above-described combined dynamic k-NN mechanism and received signal angular localization (AOA) is provided. In Table 5, the mobile device 200 is placed at (10, 10, 185) through the AOA (Reference: Ye-Sheng Kuo, Pat Pannuto, Ko-Jen Hsiao, and Prabal Dutta. 2014. Luxapose: indoor positioning with mobile phones And visible light. In Proceedings of the 20th annual international conference on Mobile computing and networking (MobiCom '14). ACM, New York, NY, USA, 447-458.), which gives z = 184.64184 = H M , for reference Signal vector The ji (x i , y i ) data set has an H R of 197, and the proportional relationship is: <TABLE border="1"borderColor="#000000"width="85%"><TBODY><tr><td> i </td><td> 1 </td><td> 2 </td><td> 3 </td><td> 4 </td><td> 5 </td></tr><tr><td><i>M<sub>i</sub></i></td><td> 525.02381 </td><td> 628.68514 </td><td> 703.2951 </td><td> 616.40409 </td><td> 95.67131 </td></tr><tr ><td><img wi="27"he="22"file="TWI627433B_D0036.tif"img-format="jpg"></img></td><td> 492.08802 </td><td> 589.24646 </td><td> 659.17599 </td><td> 577.73583 </td><td> 89.66966 </td></tr><tr><td>R<sub>ji</sub></Td><td> 499.76895 </td><td> 579.10448 </td><td> 651.87806 </td><td> 587.13286 </td><td> 82.03658 </td></tr><tr><Td>E<sub>j</sub>=19.018;k=1</td></tr></TBODY></TABLE>Table 5
由表五所得之定位位置座標P=(10, 10, 184.64184),可見在動態k-NN機制結合AOA的定位方法中,即使只有物距為 H R 之水平平面上的參考訊號,對於在物距為 H M ( H M 可以與 H R 相異)之定位需求,亦有相當優異的定位表現。 From the positioning position coordinates P=(10, 10, 184.64184) obtained in Table 5, it can be seen that in the dynamic k-NN mechanism combined with the AOA positioning method, even if only the reference signal on the horizontal plane with the object distance is H R , for the object The positioning requirement of H M ( H M can be different from H R ) also has a very good positioning performance.
參照表6,其為根據本發明一示例性實施例之實驗結果;並同時參照第11圖,其為根據同一實施例之CDF圖,其中,實驗1表示以AOA進行固定高度的定位、實驗2表示以k-NN機制且固定k值為4所進行固定高度的定位、以及實驗3表示以根據本發明技術特徵之一之動態k-NN機制所進行固定高度的定位。 <TABLE border="1" borderColor="#000000" width="85%"><TBODY><tr><td> </td><td> 實驗1 </td><td> 實驗2 </td><td> 實驗3 </td></tr><tr><td> 平均誤差(公分) </td><td> 11.43168 </td><td> 1.47079 </td><td> 0.15133 </td></tr><tr><td> 精確度(%) </td><td> 99.96579 </td><td> 99.99943 </td><td> 99.99999 </td></tr><tr><td> 最小誤差(公分) </td><td> 1.141859 </td><td> 0.170071 </td><td> 0 </td></tr><tr><td> 最大誤差(公分) </td><td> 65.60551 </td><td> 5.43629 </td><td> 4.32988 </td></tr><tr><td> CDF(50%)(公分) </td><td> 9.02908 </td><td> 1.20194 </td><td> 0 </td></tr><tr><td> CDF(70%) (公分) </td><td> 11.67671 </td><td> 1.75007 </td><td> 0 </td></tr><tr><td> CDF(90%) (公分) </td><td> 23.21883 </td><td> 2.95723 </td><td> 0 </td></tr></TBODY></TABLE>表6 Referring to Table 6, which is an experimental result according to an exemplary embodiment of the present invention; and also refers to FIG. 11, which is a CDF diagram according to the same embodiment, wherein Experiment 1 indicates positioning with a fixed height of AOA, Experiment 2 Positioning indicating a fixed height with a k-NN mechanism and a fixed k value of 4, and Experiment 3 representing a fixed height positioning with a dynamic k-NN mechanism according to one of the technical features of the present invention. <TABLE border="1" borderColor="#000000" width="85%"><TBODY><tr><td> </td><td> Experiment 1 </td><td> Experiment 2 </td ><td> Experiment 3 </td></tr><tr><td> Average Error (cm) </td><td> 11.43168 </td><td> 1.47079 </td><td> 0.15133 < /td></tr><tr><td> Accuracy (%) </td><td> 99.96579 </td><td> 99.99943 </td><td> 99.99999 </td></tr> <tr><td> Minimum error (cm) </td><td> 1.141859 </td><td> 0.170071 </td><td> 0 </td></tr><tr><td> Maximum Error (cm) </td><td> 65.60551 </td><td> 5.43629 </td><td> 4.32988 </td></tr><tr><td> CDF(50%)(cm) </td><td> 9.02908 </td><td> 1.20194 </td><td> 0 </td></tr><tr><td> CDF(70%) (cm) </td> <td> 11.67671 </td><td> 1.75007 </td><td> 0 </td></tr><tr><td> CDF(90%) (cm) </td><td> 23.21883 </td><td> 2.95723 </td><td> 0 </td></tr></TBODY></TABLE>Table 6
參照表7,其為根據本發明一示例性實施例之實驗結果;並同時參照第12圖,其為根據同一實施例之CDF圖,其中,實驗4表示以AOA進行高度不固定的二維(水平座標)定位、實驗5表示以根據本發明技術特徵之一之動態k-NN機制所進行高度不固定的二維(水平座標)定位、以及實驗6表示以根據本發明技術特徵之一之動態k-NN機制結合AOA所進行高度不固定的二維(水平座標)定位。 <TABLE border="1" borderColor="#000000" width="85%"><TBODY><tr><td> </td><td> 實驗4 </td><td> 實驗5 </td><td> 實驗6 </td></tr><tr><td> 平均誤差(公分) </td><td> 11.10302 </td><td> 2.74168 </td><td> 2.43021 </td></tr><tr><td> 精確度(%) </td><td> 99.96772 </td><td> 99.99803 </td><td> 99.99845 </td></tr><tr><td> 最小誤差(公分) </td><td> 0.139768 </td><td> 0 </td><td> 0 </td></tr><tr><td> 最大誤差(公分) </td><td> 72.59017 </td><td> 17.53132 </td><td> 13.71892 </td></tr><tr><td> CDF(50%)(公分) </td><td> 8.57792 </td><td> 3.382318 </td><td> 0 </td></tr><tr><td> CDF(70%) (公分) </td><td> 11.80708 </td><td> 3.982418 </td><td> 4.4489 </td></tr><tr><td> CDF(90%) (公分) </td><td> 22.71855 </td><td> 7.11323 </td><td> 6.63262 </td></tr></TBODY></TABLE>表7 Referring to Table 7, which is an experimental result according to an exemplary embodiment of the present invention; and at the same time, reference is made to Fig. 12, which is a CDF diagram according to the same embodiment, wherein Experiment 4 represents a two-dimensionally unfixed height with AOA ( Horizontal coordinate), experiment 5 represents a highly unfixed two-dimensional (horizontal coordinate) positioning with a dynamic k-NN mechanism according to one of the features of the present invention, and Experiment 6 represents dynamics according to one of the features of the present invention. The k-NN mechanism combines AOA with highly non-fixed two-dimensional (horizontal coordinates) positioning. <TABLE border="1" borderColor="#000000" width="85%"><TBODY><tr><td> </td><td> Experiment 4 </td><td> Experiment 5 </td ><td> Experiment 6 </td></tr><tr><td> Average Error (cm) </td><td> 11.10302 </td><td> 2.74168 </td><td> 2.43021 < /td></tr><tr><td> Accuracy (%) </td><td> 99.96772 </td><td> 99.99803 </td><td> 99.99845 </td></tr> <tr><td> Minimum error (cm) </td><td> 0.139768 </td><td> 0 </td><td> 0 </td></tr><tr><td> Maximum Error (cm) </td><td> 72.59017 </td><td> 17.53132 </td><td> 13.71892 </td></tr><tr><td> CDF(50%)(cm) </td><td> 8.57792 </td><td> 3.382318 </td><td> 0 </td></tr><tr><td> CDF(70%) (cm) </td> <td> 11.80708 </td><td> 3.982418 </td><td> 4.4489 </td></tr><tr><td> CDF(90%) (cm) </td><td> 22.71855 </td><td> 7.11323 </td><td> 6.63262 </td></tr></TBODY></TABLE> Table 7
參照表8,其為根據本發明一示例性實施例之實驗結果;並同時參照第13圖,其為根據同一實施例之CDF圖,其中,實驗7表示以AOA進行高度不固定的三維定位、實驗8表示以根據本發明技術特徵之一之動態k-NN機制所進行高度不固定的三維定位、以及實驗9表示以根據本發明技術特徵之一之動態k-NN機制結合AOA所進行高度不固定的三維定位。 <TABLE border="1" borderColor="#000000" width="85%"><TBODY><tr><td> </td><td> 實驗7 </td><td> 實驗8 </td><td> 實驗9 </td></tr><tr><td> 平均誤差(公分) </td><td> 11.61865 </td><td> 4.1497 </td><td> 2.82779 </td></tr><tr><td> 精確度(%) </td><td> 99.9562 </td><td> 99.98957 </td><td> 99.99551 </td></tr><tr><td> 最小誤差(公分) </td><td> 1.84437 </td><td> 0 </td><td> 0.01465 </td></tr><tr><td> 最大誤差(公分) </td><td> 72.61268 </td><td> 28.72281 </td><td> 13.96489 </td></tr><tr><td> CDF(50%)(公分) </td><td> 9.08855 </td><td> 3.2423 </td><td> 3.97023 </td></tr><tr><td> CDF(70%) (公分) </td><td> 12.01519 </td><td> 5.46295 </td><td> 5.2382 </td></tr><tr><td> CDF(90%) (公分) </td><td> 23.57759 </td><td> 11.17699 </td><td> 7.49193 </td></tr></TBODY></TABLE>表8 Reference is made to Table 8, which is an experimental result according to an exemplary embodiment of the present invention; and also refers to Figure 13, which is a CDF diagram according to the same embodiment, wherein Experiment 7 represents a three-dimensional positioning with a high degree of unfixed AOA. Experiment 8 shows a highly unfixed three-dimensional positioning with a dynamic k-NN mechanism according to one of the features of the present invention, and Experiment 9 shows that the height is not combined with the AOA according to the dynamic k-NN mechanism according to one of the technical features of the present invention. Fixed three-dimensional positioning. <TABLE border="1" borderColor="#000000" width="85%"><TBODY><tr><td> </td><td> Experiment 7 </td><td> Experiment 8 </td ><td> Experiment 9 </td></tr><tr><td> Average Error (cm) </td><td> 11.61865 </td><td> 4.1497 </td><td> 2.82779 < /td></tr><tr><td> Accuracy (%) </td><td> 99.9562 </td><td> 99.98957 </td><td> 99.99551 </td></tr> <tr><td> Minimum error (cm) </td><td> 1.84437 </td><td> 0 </td><td> 0.01465 </td></tr><tr><td> Maximum Error (cm) </td><td> 72.61268 </td><td> 28.72281 </td><td> 13.96489 </td></tr><tr><td> CDF(50%)(cm) </td><td> 9.08855 </td><td> 3.2423 </td><td> 3.97023 </td></tr><tr><td> CDF(70%) (cm) </td> <td> 12.01519 </td><td> 5.46295 </td><td> 5.2382 </td></tr><tr><td> CDF(90%) (cm) </td><td> 23.57759 </td><td> 11.17699 </td><td> 7.49193 </td></tr></TBODY></TABLE> Table 8
參考表6至表8,透過實驗驗證本發明所提出的定位方法的效能比以往AOA定位機制有更高的定位精準度,在二維(水平座標)室內定位方面只有2.43公分誤差,在三維室內定位方面的精確度只有2.83公分的誤差。由實驗結果得知,本研究提出之定位機制可提供較高之定位準確度。Referring to Tables 6 to 8, it is verified by experiments that the performance of the positioning method proposed by the present invention has higher positioning accuracy than the previous AOA positioning mechanism, and only 2.43 cm error in two-dimensional (horizontal coordinate) indoor positioning in a three-dimensional interior. The accuracy of positioning is only 2.83 cm. It is known from the experimental results that the positioning mechanism proposed in this study can provide higher positioning accuracy.
以上所述僅為舉例性,而非為限制性者。任何未脫離本創作之精神與範疇,而對其進行之等效修改或變更,均應包含於後附之申請專利範圍中。The above is intended to be illustrative only and not limiting. Any equivalent modifications or alterations to the spirit and scope of this creation shall be included in the scope of the appended patent application.
100‧‧‧光發射系統
101‧‧‧光發射裝置
102‧‧‧微控制器
200‧‧‧行動裝置
201‧‧‧影像調整模組
202‧‧‧影像擷取模組
203‧‧‧第一通訊模組
204‧‧‧CMOS感測元件
300‧‧‧伺服器
301‧‧‧影像處理模組
302‧‧‧影像訊息擷取模組
303‧‧‧計算模組
304‧‧‧資料庫
305‧‧‧第二通訊模組
1000‧‧‧光室內定位系統
Fj ‧‧‧參考點座標
HR 、HM ‧‧‧物距
i (x i ,y i )‧‧‧定位訊號向量
Mi 、‧‧‧定位訊號長度
ji (x i ,y i )‧‧‧參考訊號向量100‧‧‧Light launch system
101‧‧‧Light emitting device
102‧‧‧Microcontroller
200‧‧‧ mobile device
201‧‧‧Image Adjustment Module
202‧‧‧Image capture module
203‧‧‧First communication module
204‧‧‧CMOS sensing components
300‧‧‧Server
301‧‧‧Image Processing Module
302‧‧‧Image message capture module
303‧‧‧ Calculation Module
304‧‧‧Database
305‧‧‧Second communication module
1000‧‧‧Light indoor positioning system
F j ‧‧‧ reference point coordinates
H R , H M ‧‧‧
i (x i , y i ) ‧ ‧ positioning signal vector
M i , ‧‧‧Location signal length
Ji (x i , y i ) ‧ ‧ reference signal vector
第1圖是根據本發明一實施例之光室內定位方法流程圖;1 is a flow chart of a method for positioning an optical room according to an embodiment of the present invention;
第2圖是根據本發明一實施例之光室內定位系統方塊圖;2 is a block diagram of an optical indoor positioning system according to an embodiment of the present invention;
第3圖是根據本發明一實施例之開關鍵控(on-off keying)調變示意圖;3 is a schematic diagram of on-off keying modulation according to an embodiment of the present invention;
第4圖是根據本發明一實施例之開關鍵控以及捲簾快門示意圖;Figure 4 is a schematic view of the opening key control and the rolling shutter according to an embodiment of the present invention;
第5圖是根據本發明一實施例之光發射裝置設置照片;Figure 5 is a photograph showing a light emitting device according to an embodiment of the present invention;
第6圖是根據本發明一實施例之參考影像示意圖;Figure 6 is a schematic diagram of a reference image according to an embodiment of the invention;
第7圖是根據本發明一實施例之參考點設置平面圖;Figure 7 is a plan view showing a reference point according to an embodiment of the present invention;
第8圖是根據本發明一實施例之光室內定位方法之定位階段流程圖;Figure 8 is a flow chart showing the positioning phase of the optical indoor positioning method according to an embodiment of the present invention;
第9圖是根據本發明一實施例之定位影像示意圖;FIG. 9 is a schematic diagram of a positioning image according to an embodiment of the invention; FIG.
第10圖是根據習知技術之凸透鏡成像示意圖;Figure 10 is a schematic view showing the convex lens imaging according to the prior art;
第11圖是根據本發明一實施例之CDF圖;Figure 11 is a CDF diagram in accordance with an embodiment of the present invention;
第12圖是根據本發明另一實施例之CDF圖;以及Figure 12 is a CDF diagram in accordance with another embodiment of the present invention;
第13圖是根據本發明再一實施例之CDF圖。Figure 13 is a CDF diagram in accordance with still another embodiment of the present invention.
Claims (10)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW106136257A TWI627433B (en) | 2017-10-20 | 2017-10-20 | Indoor light positioning method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW106136257A TWI627433B (en) | 2017-10-20 | 2017-10-20 | Indoor light positioning method and system |
Publications (2)
Publication Number | Publication Date |
---|---|
TWI627433B true TWI627433B (en) | 2018-06-21 |
TW201917407A TW201917407A (en) | 2019-05-01 |
Family
ID=63256161
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
TW106136257A TWI627433B (en) | 2017-10-20 | 2017-10-20 | Indoor light positioning method and system |
Country Status (1)
Country | Link |
---|---|
TW (1) | TWI627433B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111751784A (en) * | 2020-06-23 | 2020-10-09 | 上海申核能源工程技术有限公司 | Three-dimensional optical positioning system of nuclear power station |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
TWI746417B (en) * | 2020-04-14 | 2021-11-11 | 鑫行動股份有限公司 | Computer vision positioning method and device |
TWI745932B (en) * | 2020-04-14 | 2021-11-11 | 鑫行動股份有限公司 | Computer vision positioning method and device |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4136955A (en) * | 1975-09-02 | 1979-01-30 | Kern & Co. Ag | Apparatus for correcting theodolites |
US20100322635A1 (en) * | 2009-06-18 | 2010-12-23 | Sony Ericsson Mobile Communications Ab | Using ambient led light for broadcasting info and navigation |
CN104048661A (en) * | 2014-06-26 | 2014-09-17 | 中国科学技术大学 | Visible light-based indoor positioning method and system |
TW201604570A (en) * | 2014-07-31 | 2016-02-01 | 國立雲林科技大學 | A visible light positioning method and a system thereof |
CN206541028U (en) * | 2017-01-20 | 2017-10-03 | 百色学院 | The single image sensor indoor visible light alignment system measured based on non-angled |
-
2017
- 2017-10-20 TW TW106136257A patent/TWI627433B/en active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4136955A (en) * | 1975-09-02 | 1979-01-30 | Kern & Co. Ag | Apparatus for correcting theodolites |
US20100322635A1 (en) * | 2009-06-18 | 2010-12-23 | Sony Ericsson Mobile Communications Ab | Using ambient led light for broadcasting info and navigation |
CN104048661A (en) * | 2014-06-26 | 2014-09-17 | 中国科学技术大学 | Visible light-based indoor positioning method and system |
TW201604570A (en) * | 2014-07-31 | 2016-02-01 | 國立雲林科技大學 | A visible light positioning method and a system thereof |
CN206541028U (en) * | 2017-01-20 | 2017-10-03 | 百色学院 | The single image sensor indoor visible light alignment system measured based on non-angled |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111751784A (en) * | 2020-06-23 | 2020-10-09 | 上海申核能源工程技术有限公司 | Three-dimensional optical positioning system of nuclear power station |
CN111751784B (en) * | 2020-06-23 | 2023-11-21 | 上海申核能源工程技术有限公司 | Three-dimensional light positioning system of nuclear power station |
Also Published As
Publication number | Publication date |
---|---|
TW201917407A (en) | 2019-05-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Zhang et al. | Image sensor based visible light positioning system with improved positioning algorithm | |
CN107256377B (en) | Method, device and system for detecting object in video | |
TWI627433B (en) | Indoor light positioning method and system | |
CN108875577A (en) | Object detection method, device and computer readable storage medium | |
US20160042515A1 (en) | Method and device for camera calibration | |
TW201229962A (en) | Augmenting image data based on related 3D point cloud data | |
US10261164B2 (en) | Active person positioning device and activity data acquisition device | |
KR101953626B1 (en) | Method of tracking an object based on multiple histograms and system using the method | |
CN107110949A (en) | Camera parameter is changed based on wireless signal information | |
US10438405B2 (en) | Detection of planar surfaces for use in scene modeling of a captured scene | |
CN103593641B (en) | Object detecting method and device based on stereo camera | |
CN109691092A (en) | System and method for improved depth sense | |
CN108091288A (en) | Test method, terminal and the computer readable storage medium of display screen uniformity | |
CN110619807B (en) | Method and device for generating global thermodynamic diagram | |
US20190191277A1 (en) | System and method for information enhancement in a mobile environment | |
CN106462244A (en) | Methods and systems for calibrating sensors using recognized objects | |
US20150170331A1 (en) | Method and Device for Transforming an Image | |
CN106845378B (en) | A method of human body target is identified from opposite image | |
JP5901370B2 (en) | Image processing apparatus, image processing method, and image processing program | |
CN107003385A (en) | Vision data is marked with wireless signal information | |
Ifthekhar et al. | Radiometric and geometric camera model for optical camera communications | |
Quan et al. | Indoor positioning modeling by visible light communication and imaging | |
Zhao et al. | CNN-based indoor occupant localization via active scene illumination | |
JP2004046464A (en) | Apparatus and method for estimating three-dimensional position of mobile object, program, and recording medium thereof | |
KR102316875B1 (en) | A method for measuring fine dust concentration using a terminal having a camera and a light emitting unit, a method for sharing measured fine dust concentration information, and a server for the method |