TWI465961B - Intelligent seat passenger image sensing device - Google Patents
Intelligent seat passenger image sensing device Download PDFInfo
- Publication number
- TWI465961B TWI465961B TW100142057A TW100142057A TWI465961B TW I465961 B TWI465961 B TW I465961B TW 100142057 A TW100142057 A TW 100142057A TW 100142057 A TW100142057 A TW 100142057A TW I465961 B TWI465961 B TW I465961B
- Authority
- TW
- Taiwan
- Prior art keywords
- passenger
- image
- seat
- sensing device
- probability
- Prior art date
Links
- 238000012545 processing Methods 0.000 claims description 15
- 238000003860 storage Methods 0.000 claims description 9
- 238000010187 selection method Methods 0.000 claims description 3
- 238000012805 post-processing Methods 0.000 claims 1
- 238000007781 pre-processing Methods 0.000 claims 1
- 238000005516 engineering process Methods 0.000 description 5
- 238000013461 design Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000005422 blasting Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000000034 method Methods 0.000 description 2
- 238000012827 research and development Methods 0.000 description 2
- 238000012356 Product development Methods 0.000 description 1
- 206010039203 Road traffic accident Diseases 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
- Air Bags (AREA)
Description
本發明是有關於一種智慧型座椅乘客影像感測裝置，尤指一種可擷取多種乘客狀態之共同特徵及引入影像特徵空間脈絡關係，以有效克服光線劇烈變化及座椅變動時之影像辨識，而達到高辨識率之功效者。The invention relates to a smart seat passenger image sensing device, in particular to a common feature of capturing a plurality of passenger states and introducing a spatial relationship of the image feature space, so as to effectively overcome the image change when the light changes drastically and the seat changes. And achieve the effect of high recognition rate.
按，安全氣囊的設計與發明，確實有效降低了交通意外發生時的傷害，其設計原理系當車輛遭受強烈撞擊時，立即引爆藥包並快速充氣，以達到保護乘客之目的。但由於瞬間充氣力道強烈，安全氣囊爆破在某些狀態下，例如：乘客為幼兒或兒童，本身卻也對使用者(例如乘車駕駛、乘客)造成一定傷害。早期安全氣囊並未針對不同的保護對象而設計不同的爆破壓制，往往造成爆破傷害或保護不足的結果。有鑑於此，目前許多產業界以及學術機構皆著手在研發感測器用以偵測駕駛乘坐狀態，以正確控制安全氣囊爆，避免爆破傷害。According to the design and invention of the airbag, it effectively reduces the damage when the traffic accident occurs. The design principle is that when the vehicle is subjected to a strong impact, the drug pack is immediately detonated and quickly inflated to protect the passenger. However, due to the strong inflating power, the airbag blasting in certain states, for example, the passenger is a child or a child, but it also causes certain damage to the user (such as driving, passengers). Early airbags did not design different bursts of compression for different objects of protection, often resulting in blast damage or inadequate protection. In view of this, many industry and academic institutions are currently developing sensors to detect the driving state, in order to properly control the airbag burst and avoid blast damage.
目前常用之感測裝置主要以重量感測器為主，其主要原理係以壓電裝置感測並估算座椅上之乘客重量，作為控制安全氣曩爆破程度之主要判定依據，但重量感測器常因乘客乘坐姿態差異，或行駛過程之震動因素，導致重量判定上誤差；另一方面，重量感測器製造成本昂貴且安裝不易。有鑑於此，本發明感測方式主要採用影像裝置，作為乘客辨識技術之主要依據。At present, the commonly used sensing devices are mainly weight sensors. The main principle is to sense and estimate the passenger weight on the seat by using the piezoelectric device as the main basis for controlling the degree of safety air blasting, but the weight sensing. The device often causes errors in weight determination due to differences in passenger seating posture or vibration factors during driving; on the other hand, the weight sensor is expensive to manufacture and difficult to install. In view of this, the sensing method of the present invention mainly uses an image device as a main basis for passenger identification technology.
以電腦影像為基礎之乘客辨識技術，根據使用相機數量之不同，大致上可區分為單眼視覺辨識技術(Monocular Vision) 以及立體視覺辨識技術(Stereo Vision)兩大類，此兩類最大差異在於如何獲得有效及不受光線強弱變化影響之乘客影像特徵，以做為乘客種類辨識依據。單眼視覺辨識技術所選用之特徵主要包含，不受光線影像之邊緣影像，乘客輪廓外觀影像，亦或與背景相減後之差異影像作為特徵。而立體視覺辨識技術所採用特徵，為透過兩臺相機以模仿人類視覺成像原理，進而感測三維空間物體與相機間之相對距離影像，稱之為視差圖(Disparity Map)。Computer image-based passenger identification technology can be roughly classified into Monocular Vision according to the number of cameras used. And stereoscopic identification technology (Stereo Vision) two categories, the biggest difference between the two types is how to obtain the image characteristics of passengers that are effective and not affected by changes in light intensity, as a basis for passenger identification. The features selected by the monocular vision recognition technology mainly include, not being affected by the edge image of the light image, the appearance image of the passenger contour, or the difference image after subtracting from the background. The feature adopted by the stereoscopic vision recognition technology is to use the two cameras to imitate the principle of human visual imaging, and then to sense the relative distance between the three-dimensional object and the camera, which is called a disparity map.
上述之習用技術，無論採用單眼或立體視覺，針對所採用之特徵進行分類，以達到乘客辨識目的，而所採用的分類演算法包含：最相近分類法(Nearest Neighbor)、k-最相近分類法(k Nearest Neighbor)、非線性鑑別分析(Non-linear Discriminate Analysis)或支援向量機(Support Vector Machine)。然先前技藝能夠有效克服場景光線變化以及乘客外觀差異，主要基於相機固定與座椅不動之假設下，但此假設並不適用於實際產品開發與應用。The above-mentioned conventional techniques, whether using monocular or stereoscopic vision, classify the features used to achieve passenger identification purposes, and the classification algorithms used include: Nearest Neighbor, k-closest classification (k Nearest Neighbor), Non-linear Discriminate Analysis or Support Vector Machine. However, the previous techniques can effectively overcome scene light changes and passenger appearance differences, mainly based on the assumption that the camera is fixed and the seat is not moving, but this assumption is not applicable to actual product development and application.
有鑑於此，本案之發明人特針對前述習用發明問題深入探討，並藉由多年從事相關產業之研發與製造經驗，積極尋求解決之道，經過長期努力之研究與發展，終於成功的開發出本發明「智慧型座椅乘客影像感測裝置」，藉以改善習用之種種問題。In view of this, the inventors of this case have intensively discussed the above-mentioned problems of conventional inventions, and actively pursued solutions through years of experience in R&D and manufacturing of related industries. After long-term efforts in research and development, they finally succeeded in developing this book. Invented the "smart seat passenger image sensing device" to improve various problems in the past.
本發明之主要目的係在於，可擷取多種乘客狀態之共同特徵及引入影像特徵空間脈絡關係，以有效克服光線劇烈變化及 座椅變動時之影像辨識，而達到高辨識率之功效。The main purpose of the present invention is to capture the common features of various passenger states and introduce spatial relationship between image features to effectively overcome the dramatic changes in light and Image recognition when the seat changes, and achieves high recognition rate.
為達上述之目的，本發明係一種智慧型座椅乘客影像感測裝置其包含有：一擷取車內座椅上乘客狀態影像之影像擷取機構；一與影像擷取機構連接之影像資料處理機構，係經由學習與辨識估算影像之整體特徵以及整體特徵所對應之乘客狀態機率，並整合該乘客狀態機率，而選取出機率最大之乘客種類，以正確辨識出乘客之各種狀態；以及一與影像資料處理機構連接之儲存機構，可儲存處理前及處理後之影像資料。In order to achieve the above object, the present invention is a smart seat passenger image sensing device comprising: an image capturing mechanism for capturing a passenger state image on a seat in the vehicle; and an image data connected to the image capturing mechanism The processing mechanism estimates the overall characteristics of the image and the passenger state probability corresponding to the overall feature by learning and identifying, and integrates the passenger state probability, and selects the passenger type with the highest probability to correctly identify the various states of the passenger; A storage mechanism connected to the image data processing mechanism can store image data before and after processing.
於本發明之一實施例中，該影像擷取機構係可為一攝影機。In an embodiment of the invention, the image capturing mechanism can be a camera.
於本發明之一實施例中，該影像資料處理機構係包含有一學習單元及一辨識單元。In an embodiment of the invention, the image data processing mechanism includes a learning unit and an identification unit.
於本發明之一實施例中，該學習單元主要先透過建構多種座椅狀態之無人乘客背景模型，乘客外觀特徵主要利用影像與背景模型之差異量(Difference Measure)來描述，而乘客模型之學習，則是透過共享特徵選取方式來達成，其主要為所選取之特徵區塊可同時對應至一個或多個乘客種類。In an embodiment of the present invention, the learning unit mainly constructs an unmanned passenger background model of various seat states, and the appearance characteristics of the passenger are mainly described by using a difference between the image and the background model (Difference Measure), and the passenger model is learned. It is achieved by a shared feature selection method, which mainly means that the selected feature block can simultaneously correspond to one or more passenger categories.
於本發明之一實施例中，該辨識單元除估算其於所習之乘客模型之可能性外，並同時引入乘客外觀與背景間之特徵空間脈絡關係，來降低因座椅變動之影響。In an embodiment of the present invention, the identification unit not only estimates the possibility of the passenger model, but also introduces a characteristic spatial relationship between the appearance of the passenger and the background to reduce the influence of the seat variation.
於本發明之一實施例中，該儲存機構係可為快閃記憶體。In an embodiment of the invention, the storage mechanism can be a flash memory.
請參閱『第1及第2圖』所示，係分別為本發明之組成架構示意圖及本發明之使用狀態流程示意圖。如圖所示：本發明 係一種智慧型座椅乘客影像感測裝置，其至少包含有一影像擷取機構1、一影像資料處理機構2以及一儲存機構3所構成。Please refer to the "1st and 2nd drawings" for a schematic diagram of the composition of the present invention and a schematic diagram of the state of use of the present invention. As shown: the present invention The invention relates to a smart seat passenger image sensing device, which comprises at least one image capturing mechanism 1, an image data processing mechanism 2 and a storage mechanism 3.
上述所提之影像擷取機構1係可擷取車內座椅上之乘客狀態影像，而該影像擷取機構1係可為一攝影機。The image capturing mechanism 1 mentioned above can capture the image of the passenger state on the seat in the vehicle, and the image capturing mechanism 1 can be a camera.
該影像資料處理機構2係與影像擷取機構1連接，經由學習與辨識估算影像之整體特徵以及整體特徵所對應之乘客狀態機率，並整合該乘客狀態機率，而選取出機率最大之乘客種類，以正確辨識出乘客之各種狀態，其中該影像資料處理機構2係包含有一學習單元21及一辨識單元22，該學習單元21主要先透過建構多種座椅狀態之無人乘客背景模型，乘客外觀特徵主要利用影像與背景模型之差異量(Difference Measure)來描述，而乘客模型之學習，則是透過共享特徵選取方式來達成，其主要為所選取之特徵區塊可同時對應至一個或多個乘客種類；而該辨識單元22除估算其於所習之乘客模型之可能性外，並同時引入乘客外觀與背景間之特徵空間脈絡關係，來降低因座椅變動之影響。The image data processing mechanism 2 is connected to the image capturing mechanism 1 to estimate the passenger state probability corresponding to the overall feature of the image and the overall feature by learning and recognizing, and integrating the passenger state probability to select the passenger type with the highest probability. In order to correctly identify the various states of the passengers, the image data processing mechanism 2 includes a learning unit 21 and an identification unit 22, which firstly constructs an unmanned passenger background model of various seat states, and the passenger appearance characteristics are mainly The difference between the image and the background model is described by the Difference Measure, and the learning of the passenger model is achieved by the shared feature selection method, which mainly means that the selected feature block can simultaneously correspond to one or more passenger types. And the identification unit 22 not only estimates the possibility of the passenger model, but also introduces the characteristic spatial relationship between the appearance of the passenger and the background to reduce the influence of the seat variation.
該儲存機構3係與影像資料處理機構2連接，可儲存處理前及處理後之影像資料，而該儲存機構3係可為快閃記憶體。如是，藉由上述之設計構成一全新之智慧型座椅乘客影像感測裝置。The storage mechanism 3 is connected to the image data processing unit 2, and can store image data before and after processing, and the storage mechanism 3 can be a flash memory. If so, a new smart seat passenger image sensing device is constructed by the above design.
當本發明於使用時，係以影像擷取機構1擷取一目標區域之一狀態影像s101(如：車內座椅上之乘客)，並以影像資料處理機構2根據所學習之乘客模型區塊與背景模型，計算區塊之向量特徵s102，且估算前述計算出之特徵向量，對應於乘客狀態之機率s103，之後再估算整張影像之整體特徵影像，並將 其對映至維度較小之特徵向量s104，且引入特徵空間脈絡關係，計算前述整體特徵向量所對應之乘客狀態機率s105，最後再將前述兩者之乘客狀態機率統整s106，並選取出機率最大之乘客種類；藉此，以達到正確辨識出人員的各種狀態。When the present invention is used, the image capturing mechanism 1 captures a state image s101 of a target area (eg, a passenger on a seat in the vehicle), and uses the image data processing mechanism 2 to learn the passenger model area. a block and background model, calculating a vector feature s102 of the block, and estimating the calculated feature vector, corresponding to the probability s103 of the passenger state, and then estimating the overall feature image of the entire image, and It is mapped to the feature vector s104 with a small dimension, and the feature space relationship is introduced, the passenger state probability s105 corresponding to the overall feature vector is calculated, and finally the passenger state probability of the two is integrated s106, and the probability is selected. The largest type of passenger; thereby, in order to correctly identify the various states of the person.
綜上所述，本發明智慧型座椅乘客影像感測裝置可有效改善習用之種種缺點，可擷取多種乘客狀態之共同特徵及引入影像特徵空間脈絡關係，以有效克服光線劇烈變化及座椅變動時之影像辨識，而達到高辨識率之功效；進而使本發明之產生能更進步、更實用、更符合消費者使用之所須，確已符合發明專利申請之要件，爰依法提出專利申請。In summary, the intelligent seat passenger image sensing device of the present invention can effectively improve various shortcomings of the conventional use, and can acquire the common features of various passenger states and introduce spatial relationship of image features to effectively overcome the dramatic changes in light and the seat. The image recognition during the change, and the effect of high recognition rate is achieved; thus, the invention can be made more progressive, more practical, and more suitable for the use of the consumer, and has indeed met the requirements of the invention patent application, and the patent application is filed according to law. .
惟以上所述者，僅為本發明之較佳實施例而已，當不能以此限定本發明實施之範圍；故，凡依本發明申請專利範圍及發明說明書內容所作之簡單的等效變化與修飾，皆應仍屬本發明專利涵蓋之範圍內。However, the above is only the preferred embodiment of the present invention, and the scope of the present invention is not limited thereto; therefore, the simple equivalent changes and modifications made in accordance with the scope of the present invention and the contents of the invention are modified. All should remain within the scope of the invention patent.
1‧‧‧影像擷取機構1‧‧‧Image capture agency
2‧‧‧影像資料處理機構2‧‧‧Image data processing agency
21‧‧‧學習單元21‧‧‧Learning unit
22‧‧‧辨識單元22‧‧‧ Identification unit
3‧‧‧儲存機構3‧‧‧Storage agency
s101~s106‧‧‧步驟S101~s106‧‧‧Steps
第1圖，係本發明之組成架構示意圖。Figure 1 is a schematic diagram of the composition of the present invention.
第2圖，係本發明之使用狀態流程示意圖。Fig. 2 is a schematic flow chart showing the state of use of the present invention.
1‧‧‧影像擷取機構1‧‧‧Image capture agency
2‧‧‧影像資料處理機構2‧‧‧Image data processing agency
21‧‧‧學習單元21‧‧‧Learning unit
22‧‧‧辨識單元22‧‧‧ Identification unit
3‧‧‧儲存機構3‧‧‧Storage agency
Claims (3)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| TW100142057A TWI465961B (en) | 2011-11-17 | 2011-11-17 | Intelligent seat passenger image sensing device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| TW100142057A TWI465961B (en) | 2011-11-17 | 2011-11-17 | Intelligent seat passenger image sensing device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| TW201322046A TW201322046A (en) | 2013-06-01 |
| TWI465961B true TWI465961B (en) | 2014-12-21 |
Family
ID=49032366
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| TW100142057A TWI465961B (en) | 2011-11-17 | 2011-11-17 | Intelligent seat passenger image sensing device |
Country Status (1)
| Country | Link |
|---|---|
| TW (1) | TWI465961B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TW201044008A (en) * | 2009-06-12 | 2010-12-16 | Chung Shan Inst Of Science | Night time pedestrian detection technology |
| JP2011207564A (en) * | 2010-03-29 | 2011-10-20 | Fuji Xerox Co Ltd | Image forming system |
| TW201140502A (en) * | 2010-05-13 | 2011-11-16 | Hon Hai Prec Ind Co Ltd | System and method for monitoring objects |
-
2011
- 2011-11-17 TW TW100142057A patent/TWI465961B/en active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TW201044008A (en) * | 2009-06-12 | 2010-12-16 | Chung Shan Inst Of Science | Night time pedestrian detection technology |
| JP2011207564A (en) * | 2010-03-29 | 2011-10-20 | Fuji Xerox Co Ltd | Image forming system |
| TW201140502A (en) * | 2010-05-13 | 2011-11-16 | Hon Hai Prec Ind Co Ltd | System and method for monitoring objects |
Non-Patent Citations (1)
| Title |
|---|
| Michael E. Farmer and Anil K. Jain，"Occupant Classification System for Automotive Airbag Suppression"，Proceedings of the 2003 IEEE Computer Society Conference on Computer Vision and Pattern Recognition * |
Also Published As
| Publication number | Publication date |
|---|---|
| TW201322046A (en) | 2013-06-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP6942712B2 (en) | Detection of partially obstructed objects using context and depth order | |
| CN107852533B (en) | Three-dimensional content generating device and method for generating three-dimensional content | |
| JP2012059030A (en) | Human body identification method and human body identification apparatus using range image camera | |
| EP4099221A1 (en) | Face recognition method and apparatus | |
| EP2884460A1 (en) | Image capturing apparatus, monitoring system, image processing apparatus, image capturing method, and non-transitory computer readable recording medium | |
| CN107346422A (en) | A kind of living body faces recognition methods based on blink detection | |
| CN113361452A (en) | Driver fatigue driving real-time detection method and system based on deep learning | |
| CN105590090B (en) | Object detection device and method for vehicle | |
| KR101551576B1 (en) | Robot cleaner, apparatus and method for recognizing gesture | |
| JP2009150900A (en) | Database system for use in real-time head pose detection | |
| JP2013045433A5 (en) | ||
| CN103606149B (en) | Method and apparatus for calibration of binocular camera and binocular camera | |
| CN106997457B (en) | Human body recognition method, human body recognition device and electronic device | |
| CN109508576B (en) | A kind of abnormal driving behavior detection method, device and electronic device | |
| CN111832373A (en) | A method of vehicle driving attitude detection based on multi-eye vision | |
| CN102812474A (en) | Head Recognition Method | |
| CN110268416B (en) | Image processing apparatus and method | |
| WO2011086803A1 (en) | Image matching system, image matching method, computer program | |
| CN110192224A (en) | Device for estimating state of driver and estimating state of driver method | |
| TWI465961B (en) | Intelligent seat passenger image sensing device | |
| TWI499999B (en) | The 3D ring car image system based on probability calculation and its obtaining method | |
| CN104228688B (en) | Method for adjusting rearview mirror and electronic device using the method | |
| JP2020061139A (en) | CNN learning method, testing method, learning device, and testing device for blind spot monitoring | |
| CN114097005B (en) | Electronic device for recognizing object by distorting image and control method thereof | |
| TWI442326B (en) | Image classification method and image classification system |