TW582174B - Method for image compression - Google Patents
Method for image compression Download PDFInfo
- Publication number
- TW582174B TW582174B TW91118121A TW91118121A TW582174B TW 582174 B TW582174 B TW 582174B TW 91118121 A TW91118121 A TW 91118121A TW 91118121 A TW91118121 A TW 91118121A TW 582174 B TW582174 B TW 582174B
- Authority
- TW
- Taiwan
- Prior art keywords
- image
- image data
- data
- encoding
- patent application
- Prior art date
Links
Landscapes
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
Description
582174 A7 B7 經 濟 部 智 慧 財 產 局 員 工 消 費 合 社 印 製 五、發明説明( 發明領域: 本發明係有關於一種影像壓縮的方法,特別是有關於 種具有依影像資料之差異,調整壓縮方式的影像壓縮方 法 發明背景: 影像壓縮技術在遠端控制系統上是十分的重要。當遠 端控制系統進行遠端電腦控制時,可利用影像的直接傳 輸’使在遠方的客戶電腦可透過網路等等資訊傳輸媒體, 進行電腦的控制。JPEG是一種影像壓縮的演算法,係由國 際才示 準 組織(International Organization for582174 A7 B7 Printed by the Consumers ’Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs 5. Field of the Invention (Field of the Invention: The present invention relates to a method of image compression, and in particular to an image having a compression method that is adjusted according to differences in image data. Background of the invention of the compression method: Image compression technology is very important in the remote control system. When the remote control system performs remote computer control, the direct transmission of the image can be used to make the remote client computer through the network, etc. Information transmission media for computer control. JPEG is an image compression algorithm developed by the International Organization for
Standardization ; ISO)和國際電話電報諮詢委員會 (International Telegraph and Telephone Consultative Committee ; CCITT)所建立的一個數位影像壓縮標準,主 要疋用於靜態影像壓縮方面。當使用JPEG的演算法進行影 像壓縮時,數位影像將依據JPEG演算法的要求,先將影像 區分為複數個區塊(Block),再進行每個區塊的壓縮與傳 送。一般而言,為要求較佳的影像品質,即使每一個區塊 中僅有部分的晝素變更,但是整個區塊均會被要求壓縮及 傳送。 第一圖係為習知的遠端控制系統,利用影像壓縮的技 術,使客戶端電腦可由遠端控制其他電腦。如圖中所示, ::K......-·裝訂.........參 (、請先閲讀背面之注意事項再填寫本頁)Standardization; ISO) and the International Telegraph and Telephone Consultative Committee (CCITT) established a digital image compression standard, mainly used for still image compression. When using the JPEG algorithm for image compression, the digital image will be divided into multiple blocks according to the requirements of the JPEG algorithm, and then each block will be compressed and transmitted. In general, in order to require better image quality, even if only a part of the diurnal changes in each block, the entire block will be required to be compressed and transmitted. The first picture is a conventional remote control system, which uses image compression technology to enable client computers to remotely control other computers. As shown in the picture, :: K ......- · Binding ......... Refer to (Please read the notes on the back before filling this page)
丨7公釐) 582174 A7 B7 五、發明説明() 此遠端控制系統有控制器120 ,電腦切換器 (Keyboard-Video-Mouse Switch ; KVM Switch)104,與四台 (請先閲讀背面之注意事項再填寫本頁) 由電腦切換器104所控制的電腦1〜4 ι〇2。此遠端控制器 120透過網路122與客戶端電腦124聯絡。而此遠端控制 器120包含,A/D轉換器106、先進先出控制器(First w Out; FIFO) 108、輸出輸入模組11〇、中央處理單元(Central Processing Unit; CPU) 112、記憶體控制器 114、記憶體 116 及網路卡1 1 8。 若A/D轉換器106進行影像資料的取樣後,產生2〇〇 個畫素(Pixe”改變’且其分配纟2〇〇個區塊中,而每個區 塊有16*16個畫素的情況時,$ 了壓縮及傳輸這些影像資 料,系統必須具有強大的壓縮與讀取能力,以能在適當的 時間間隔内,完成這些影像資料的處理。 經濟部智慧財產局員工消費合作社印製 而在這些影像資料完成壓縮之後,若欲進_步使用網 路以進行壓縮後影像資料的傳輸’則在網路有限的可用頻 寬下,即使已經壓縮完成之影像資料,欲進行如此魔大的 影像資料的即時傳輸,仍將造成網路極大的負擔。一般而 言’利用壓縮技術可有效抒解網路中資料傳輸形成:瓶 頸’但是由於目前影像資料大量在網路中傳輪,現有網路 在承受大量的影像資料,在有限的頻寬下,欲進行即時影 像傳輸及電腦控制,將受限於網路的傳輸能力。 如何有效的壓縮連續而魔大的影像資料,以利用有限 ^網路頻寬’進行影像即時傳輪,並進行遠端電腦的控制, 實為網路及伺服器管理者所共同之目標。 本、’’氏張尺度適用中國gj A標準(CNS)A4規格(2ι〇><297公楚) 經濟部智慧財產局員工消費合作社印製 582174 五、發明説明() 發明目的及概述: ϋ於上述之發明背景中, 你 已壓縮—成 右 、、、而龐大的影像資料即使 但利用有限網路頻寬及傳輸能力 二 續的影像即時傳輸與遠端電腦人進仃連 .^ ^ W觉限於網路僂給合匕 力的限制。如何有效壓縮連續 ' ^ ^ 1豕貝枓以減輕網路的备 擔’實為使用者之共同企盼。 、 本發明的目的之一,係利用 ^ ^ 4 m 種經由網路有效的值鈐 連續壓縮的影像資料的方法。 剧 本發明的又一目的,係利用乂 宁刊用比較目前的影像區塊與 前的影像區塊,計算其中不同的晝素數量,自動計算所需 使用的壓縮影像資料的演算法,進行影像資料的壓縮與傳 輸。 〃本發明的再一目的,係利用合適的壓縮影像資料的演 算法,根據不同影像情況所需,進行不同的壓縮演算,有 效的降低網路頻寬與傳輸速率的需求,使經由網路傳輸即 時影像與控制電腦設備,更為容易且時間延誤因而降低。 根據以上所述之目的,本發明係一種影像資料編碼方 法,使用於連續影像傳輸,可減少編碼後影像資料大小降 低網路頻寬的需求之功能。特別適合電腦切換器 (Keyboard-Video-Mouse Switch; KVM Switch)經由網路, 管理遠端電腦時之影像傳輸上所需。此方法包含,讀取連 續影像,使用模糊比較法,比較連續影像之前影像與後影 ......蜂裝:::·----可......... ί請先閲讀背面之注意事項再填寫本頁) ___丨 7mm) 582174 A7 B7 V. Description of the invention () This remote control system has a controller 120, a computer switch (Keyboard-Video-Mouse Switch; KVM Switch) 104, and four (please read the note on the back first) (Please fill in this page again for details.) Computers 1 to 4 controlled by computer switcher 104. The remote controller 120 communicates with the client computer 124 through the network 122. The remote controller 120 includes an A / D converter 106, a first-in-first-out (FIFO) controller 108, an input-output module 11, a central processing unit (CPU) 112, and a memory. Body controller 114, memory 116 and network card 1 1 8. If the A / D converter 106 samples the image data, it will generate 200 pixels (Pixe "changed 'and it will be allocated in 200 blocks, and each block has 16 * 16 pixels In the case of the compression and transmission of these image data, the system must have strong compression and reading capabilities in order to complete the processing of these image data at appropriate intervals. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs After the compression of these image data, if you want to further use the network for the transmission of compressed image data ', under the limited available bandwidth of the network, even if the compressed image data has been compressed, you want to do so magic. The real-time transmission of image data will still cause a great burden on the network. Generally speaking, 'the use of compression technology can effectively explain the formation of data transmission in the network: bottlenecks'. The network bears a large amount of image data, and in a limited bandwidth, real-time image transmission and computer control will be limited by the transmission capacity of the network. The continuous and magical image data, using limited ^ network bandwidth, for real-time image transfer and remote computer control, is a common goal of network and server administrators. Zhang Zhiji applies the Chinese gj A standard (CNS) A4 specification (2ι〇 > < 297 Gongchu) printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs 582174 V. Description of the invention () Purpose and summary of the invention: In the background, you have compressed—right, large, and huge image data, but even with limited network bandwidth and transmission capacity, the image is transmitted in real-time and connected to the remote computer.偻 The limitation of the combined force. How to effectively compress the continuous ^ ^ 1 豕 to reduce the burden on the network is a common expectation of users. One of the purposes of the present invention is to use ^ ^ 4 m species A method of continuously compressing image data through effective values through the network. Another object of the present invention is to use the Sui Ning publication to compare the current image block with the previous image block, calculate the number of different day elements, and automatically Calculation Algorithms for compressing image data to be used to compress and transmit image data. 的 Another object of the present invention is to use appropriate algorithms for compressing image data to perform different compression algorithms according to different image conditions. The need to effectively reduce the bandwidth and transmission rate of the network makes it easier and less time-delayed to transmit real-time images and control computer equipment over the network. According to the above-mentioned purpose, the present invention is an image data encoding method, It is used for continuous image transmission, which can reduce the size of the encoded image data and reduce the demand for network bandwidth. It is especially suitable for keyboard-video-mouse switch (KVM Switch) to manage the image of the remote computer through the network Required on transmission. This method includes reading continuous images, and using a fuzzy comparison method to compare the images before and after the continuous images ... Buzzing ::: · ---- 可 ......... Please (Please read the notes on the back before filling out this page) ___
本紙張尺度適用中國國家標準(CNS)A4規格(210X 297公釐) 582174 A7 五、發明説明( 連續相同之該些位元組的能力,例如“ ababab” 。 (請先閲讀背面之注意事項再填寫本頁) 本發明之改善影像壓縮及傳送的方法,跟據不同的需 求,決定所需的壓縮方法,充分利用每一種壓縮方法的優 點,有效降低影像傳輸速率及網路頻寬的需求。 圖式簡單說明: 本發明的較佳實施例將於往後之說明文字中辅以下列 圖形做更詳細的闡述,其中: 第一圖為習知的遠端控制系統; 第一圖為本發明之影像壓縮的主要分配程序之流程示 意圖; 第三圖為將標準的RGB555的畫素格式轉換分離為紅 (R)、綠(G)及藍(B)的格式; 第四圖為模糊比較程序之流程示意圖; 第五圖為本發明之較佳實施例RLEDMP演算法的編碼 程序; 經濟部智慧財產局員工消費合作社印製 第六圖為第五圖中步驟614所述之直接編碼程序的編 碼格式; 第七圖為第五圖中步驟614之直接編碼程序的流程示 意圖; 第八圖為第五圖中步驟604之列圖案指標編碼的流程 示意圖; 本紙張尺度適用中國國家標準(CNS)A4規格(210X297公釐) 五、發明説明( 第九圖為RLEDM編碼㈣ 第十A至十C圖為第五圖 :圖, 的流程示意圖。 〇之RLEDM編碼 圖號對照說明: 102 電腦1〜4 104 106 A/D轉換器 108 110 輸出輸入模組 112 114 記憶體控制器 116 118 網路卡 120 122 網路 124 202〜1434 步驟 電腦切換器 先進先出控制器 中央處理單元 記憶體 控制器 客戶端電腦 發明詳細說明·· 經濟部智慧財產局員工消費合作社印製 由上述之發明背景中可釦,碴雄 連續的影像資料,雖經壓 縮,但在有限網路頻寬下,欲推弁、*緣 欲進仃連續的影像即時傳輸與 遠端電腦控制,將受限於網路的值仏处a ^ J俗们得輸忐力。如何更有效壓 縮連續影像資料以減輕網路的負栌,亦你θ + ^ ^ 们員擔可使即時影像的傳輸 與遠端電腦控制的時間延誤大幅降低。 以下將以圖示及詳細說明,清楚描述本發明之精神, 如熟悉此技術之人員在瞭解本發明之較佳實施例後,當可 由本發明所教示之技術’加以改變及修飾,其並不脫離本 發明之精神與範圍。本發明係提供—種改進的變動長度編 7 ;本紙張尺度適用中國國家標準(CNS)A4規格(210X297公楚) •丨·:ι......•裝·........訂.......# π請先閲讀背面之注意事項再填寫本頁) 582174 經濟部智慧財產局員工消費合作社印製 A7 B7 五、發明説明() 碼(Run Length Encoding)及三角調變編碼法(Delta Modulation Encoding),以進行電腦影像的編碼,並透過網 路進行壓縮後之連續影像之傳輸。請參閱第二圖至第十C 圖,結合以下之說明,以詳述本發明之精神與範圍。 在此我們稱此編碼程序為變動長度編碼與三角調變位 置編碼(Run Length Encoding and Delta Modulation Position ; RLEDMP)。其中變動長度編碼以RLE代表,係 為一種無失真的演算法(Lossless Algorithm),其將重複的 資料,以原始資料與重複次數來表示,故可有效的減少資 料的數量。而三角調變以DM代表,係一種屬於差值訊號 編碼(Differential Pulse Code Modulation ; DPCM)的技術。 使用此種方法只有少數位元用來編碼不相同的地方,先決 定一數值後,在利用增量及減量的方式表達其他數值。p 代表位置編碼(Position Coding),係用來紀錄列圖案的資 料。 根據JPEG標準的要求,原始的影像資料首先被分割為 多個8*8的矩陣,每一矩陣代表64個畫素。由於JPEG壓 縮,首先要將RGB轉換成亮度與色度的格式。一般而言, 轉換成為Y代表亮度,Cb代表藍色色度,及Cr代表紅色 色度。因此我們定義1 6* 1 6的矩陣為一最小的需求單元。 當僅有部分的資料不相同時,使用JPEG演算法是不值得 的。在此種情況下,我們將採用上述之RLEDMP的演算法 來取代原來的JPEG演算法。 第二圖中為本發明之影像壓縮的主要分配程序之流程 本紙張尺度適用中國國家標準(CNS)A4規格(210X297公楚) ......0 裝.........訂.........# ^請先閲讀背面之注意事項再填寫本頁) 582174 經濟部智慧財產局員工消費合作社印製 A7 B7 五、發明說明() 示心圖。利用分配功能自動轉換資料進入特定的流程。首 先將衫像切割,以形成16* 16晝素的區塊,步驟202。 接著準備數據以進行模糊比較(Fuzzy Comparison)的程 序v驟2 0 4。然後比較現有的區塊與先前的區塊,以獲 得王口P 差異數量(T〇tal Differenee Nmber ; TDN),步驟 206再根據比較的結果,分配程序將影像資料進行不同的 壓縮,例如在本實施例中,使用預定的差異數為3 2而言, 當差異數不小於1但小於32時,進入步驟21〇,使用 RLEDMP的項算法進行壓縮。當差異數大於時,則進入 v驟212,使用jpeg的演算法進行壓縮。步驟214,儲存 編碼後的資料。雖然在此實施例中,選用32為一個使用 jpeg或RLEDMP的選擇門檻,但本發明並不限定此數值, 般而吕’在16*16的矩陣中,小於128的數值均可使用。 本發明利用一預定的差異數值進行每一影像區塊使用的壓 縮廣算法的判斷準則,使每一區塊根據與前一影像之比 車乂,選擇合適的壓縮演算法,進行壓縮,故能產生最為合 適此區塊的的影像壓縮種演算法,可有效的降低影像壓縮 後之資料大小,因而節省影像傳輸所需要的時間。 第三圖為將標準的RGB5 5 5的晝素格式轉換分離為紅 (R)、綠(G)及藍(B)的格式。習知的畫素格式使用RGB555 的格式,係為圖中之畫素格式4〇〇,為方便進行比較,本 發明將其分離為紅色畫素格式401、綠色畫素袼式4〇2及 藍色晝素格式403,而每一顏色的畫素格式使用8個位元 進行資料進行資料的儲存,如圖中所示,每個顏色的資料 ί本紙張尺度適用中國國家標準(CNS)A4規格(210X297公楚)----- .......Φ裝.........訂.........參 '請先閲讀背面之注意事項再填寫本頁} 582174 A7 五、發明説明() 均小於0x20 ’故定義0x20為結束字元(End of buffer character ; EOBC)。This paper size applies to the Chinese National Standard (CNS) A4 specification (210X 297 mm) 582174 A7 V. Description of the invention (the ability of the same bytes in a row, such as "ababab". (Please read the precautions on the back first) (Fill in this page) The method for improving image compression and transmission according to the present invention determines the required compression method according to different needs, making full use of the advantages of each compression method, and effectively reducing the image transmission rate and network bandwidth requirements. Brief description of the drawings: The preferred embodiment of the present invention will be described in more detail in the following explanatory text with the following figures, where: The first picture is a conventional remote control system; the first picture is the present invention The schematic diagram of the main distribution procedure of image compression; the third figure is the conversion of the standard RGB555 pixel format into red (R), green (G) and blue (B) format; the fourth figure is the fuzzy comparison program The schematic diagram of the process; The fifth diagram is the coding procedure of the RLEDMP algorithm of the preferred embodiment of the present invention; The coding format of the direct coding program in Figure 7; Figure 7 is a schematic flow chart of the direct coding program in step 614 in Figure 5; Figure 8 is a schematic flow chart of the pattern index coding in step 604 in Figure 5; This paper scale is applicable to China National Standard (CNS) A4 specification (210X297 mm) 5. Explanation of the invention (The ninth picture is the RLEDM code. The tenth A to ten C are the fifth picture: the diagram. The schematic diagram of the RLEDM code drawing number. : 102 computer 1 ~ 4 104 106 A / D converter 108 110 input / output module 112 114 memory controller 116 118 network card 120 122 network 124 202 ~ 1434 steps computer switch FIFO controller central processing unit Detailed description of the invention of the memory controller client computer. The employee co-operative society of the Intellectual Property Bureau of the Ministry of Economic Affairs printed the deductible and outstanding continuous image data in the background of the above invention. Although compressed, it is under limited network bandwidth. The continuous transmission of real-time images and remote computer control for pushing and fate, will be limited by the value of the network. ^ J folks have to lose power. How to have more Effectively compress continuous image data to reduce the burden on the network, and your staff can greatly reduce the time delay between the transmission of real-time images and remote computer control. The following will be clearly described with icons and detailed descriptions The spirit of the present invention can be changed and modified by those skilled in the art after learning about the preferred embodiments of the present invention without departing from the spirit and scope of the present invention. The present invention is Provide—an improved variable length series 7; this paper size is applicable to China National Standard (CNS) A4 specification (210X297 Gongchu) • 丨 :: ι ...... • installation ... ....... # π Please read the notes on the back before filling in this page) 582174 Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs A7 B7 V. Run Length Encoding and Triangle Modulation Coding method (Delta Modulation Encoding) is used to encode computer images and compress the transmission of continuous images through the network. Please refer to the second chart to the tenth chart C, in combination with the following description, to detail the spirit and scope of the present invention. Here we call this encoding program Run Length Encoding and Delta Modulation Position (RLEDMP). Among them, the variable length coding is represented by RLE, which is a lossless algorithm. It represents the repeated data by the original data and the number of repetitions, so it can effectively reduce the amount of data. The triangle modulation is represented by DM, which is a technology of differential pulse code modulation (Differential Pulse Code Modulation; DPCM). With this method, only a few bits are used to encode the difference. After a value is determined, other values are expressed in the form of increments and decrements. p stands for Position Coding, which is used to record the data of the column pattern. According to the requirements of the JPEG standard, the original image data is first divided into multiple 8 * 8 matrices, each matrix representing 64 pixels. Due to JPEG compression, RGB is first converted into a luminance and chrominance format. Generally speaking, conversion to Y represents brightness, Cb represents blue chromaticity, and Cr represents red chromaticity. Therefore we define a matrix of 16 * 16 as a minimum demand unit. When only a part of the data is different, it is not worth using the JPEG algorithm. In this case, we will use the above-mentioned RLEDMP algorithm to replace the original JPEG algorithm. The second figure shows the flow of the main distribution procedure of the image compression of the present invention. The paper size is applicable to the Chinese National Standard (CNS) A4 specification (210X297). ...... 0 installation ......... Order ......... # ^ Please read the notes on the back before filling out this page) 582174 Printed by A7 B7, Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs V. Description of the invention () Heart map. Use the distribution function to automatically convert data into a specific process. First cut the shirt image to form a 16 * 16 diurnal block, step 202. The data is then prepared for a fuzzy comparison procedure vstep 2 0 4. Then compare the existing block with the previous block to obtain the number of discrepancy (T〇tal Differenee Nmber; TDN), step 206, and then according to the comparison result, the distribution program compresses the image data differently. In the embodiment, when the predetermined number of differences is 32, when the number of differences is not less than 1 but less than 32, the process proceeds to step 21 and compression is performed using a term algorithm of RLEDMP. When the number of differences is greater, it enters v212 and compresses using jpeg algorithm. Step 214: Store the encoded data. Although in this embodiment, 32 is selected as a selection threshold using jpeg or RLEDMP, the present invention is not limited to this value. Generally, in the matrix of 16 * 16, any value less than 128 can be used. The invention uses a predetermined difference value to determine the compression algorithm used by each image block, so that each block can select an appropriate compression algorithm to perform compression according to the ratio of the previous image to the previous image. Generates the most suitable image compression algorithm for this block, which can effectively reduce the data size after image compression, thus saving the time required for image transmission. The third figure shows the conversion of the standard RGB5 55 daylight format into red (R), green (G), and blue (B) formats. The conventional pixel format uses the RGB555 format, which is the pixel format of 400 in the figure. For convenience of comparison, the present invention separates it into red pixel format 401, green pixel format 40 and blue. The color element format is 403, and the pixel format of each color uses 8 bits for data storage. As shown in the figure, the data of each color is applicable to the Chinese National Standard (CNS) A4 specification. (210X297 公 楚) ----- ....... Φ ............ Order ......... Please read the notes on the back before filling in this Page} 582174 A7 V. Description of the invention () are all less than 0x20 ', so 0x20 is defined as the End of buffer character (EOBC).
第四圖為模糊比較程序之流程示意圖。由於經由a/d 轉換器1 06產生資料取樣的不穩定的情形,在此先定義一 預定的數值以進行比較時使用。此預定的數值以模糊比較 範圍(Fuzzy Comparison Range ; FCR)來表示。當由 A/D 轉 換器106所傳送來的資料,如〇χ1〇, 〇χ1丨〇r 〇χ12均可被視 為相同的資料,以克服A/D轉換器丨〇6影像擷取時之不穩 定的現象’也就是說當差異值小於等於2時,可將其視為 相同。下列的公式一係用來進行模糊控制計算時使用: Result = Abs(CR-FR) <= FCR .and. Abs(CG-FG) <= FCR .and. Abs(CB-FB) <= FCR 上述之CR、CG與CB分別代表目前影像之晝素中分離 出來的目前R、G與B的色彩資料,而FR、FG與FB分別 代表先前影像之畫素中分離出來的先前r、G與b的色彩 資料。 當上述之Result為零的時候,代表先前影像之晝素與 目前影像之晝素不是相同的畫素,其他的情況代表此兩個 晝素係為相同的畫素。 首先’步驟502,係為模糊比較之起始設定值,設定 FCR之預定值’並將列計數器(Line counter ; LC)、點計數 器(Dot Counter ; DC)、列圖案(Line Pattern ; LP)及全部相 異數量(Total Difference Number; TDN)均先歸零。步驟 5 04 ’獲取欲比較影像1與影像2的影像資料。步驟506, 裝: (請先閲讀背面之注意事項再填寫本頁) -訂· # 經濟部智慧財產局員工消費合作社印製 10 582174 五、發明説明( 一點一點的比較影像中每一點的影像資料是否相同,當發 現此點的影像資料為相同時,直接進入步驟5丨4,而當此 點的影像資料並不相同時,則進入本發明之模糊比較的程 序。步驟508,將此點、的影像資料,依第三圖中之分離方 法分離為RGB顏色的資料,影像1的此點的影像資料被分 為R1、B1及G1,而影像2的相同位置的影像資料被分為 R2、G2及b 2。步驟5 1 0,利用公式一模糊比較此點之影像 資料是否在可被接受的範圍中,即此點之RGB值是否可被 視為相同,而被接受為相同的資料。若判定為相同則直接 進入步驟5 1 4,若被視為不相同則進入步驟5丨2。步驟5丨2, 將此點位置的LP設定為1,並將TDN的數值加1 ,代表全 部相異數量又增加一點。接著進入步驟5丨4,將LP的位置, 向左移動一個位元,並將DC加1。步驟5丨6,係判斷此列 的影像資料是否已比較完畢,在此以每列1 6個影像點為 例’故DC判斷是否大於等於1 6。倘若尚未比較完成,即 D C小於1 6,回到步驟5 0 4繼續比較下一個位置點,若已 比較元成’即D C大於等於1 6,則進入步驟5 1 8。步驟5 1 8, 係將上述比較所記錄的LP值存入LP緩衝器(LP Buffer ; LPBUF)中,並將LP與DC重新設定為零及將LC加丨,也 就是進入下一列中,繼續進行影像模糊比較的工作,直到 步驟520,完成此影像的模糊比較工作。 使用本發明之影像模糊比較的程序,當比較完成後, 同時產生了 一組列圖案的資料’以下將以表一、表二及表 三來加以詳述本發明之比較方法與比較結果。表一代表先 π 本紙張尺度適用中國國家標準(CNS)A4規格(210X297公楚) ...........·裝: /請先閲讀背面之注意事項再填寫本頁} 訂· # 經 濟 部 智 慧 財 產 局 員 工 消 費 合 社 印 製 582174 經濟部智慧財產局員工消費合作社印製 A7 B7 五、發明説明() 前的影像資料,表二代表目前的影像資料及表三代表經比 較後所產生之列圖案的資料。 本發明之模糊比較程序,係利用比較表一及表二中之 每一列及每一行的晝素點,以產生表三中的列圖案的資料。 0000 0000 〇〇〇〇 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 〇〇〇〇 0000 0001 0001 0001 0000 0000 0001 0001 0001 0000 0000 0000 0000 0000 0001 0001 0000 0020 0020 0020 0000 0000 0020 0020 0020 0001 0000 0000 0000 0000 0020 0020 0000 0021 0021 0021 0000 0000 0021 0021 0021 0020 0000 0000 0000 0000 0021 0021 0000 0000 0000 0000 0000 0000 0000 0000 0000 0021 0001 0001 0000 0000 0000 〇〇〇〇 0001 0400 0400 0400 0001 0001 0001 0400 0400 0000 0020 0020 0001 0000 0400 0400 0020 0020 0001 0020 0020 0020 0020 0020 0020 0400 0021 0021 0020 0000 0020 0020 0021 0001 0020 0001 0021 0021 0021 0001 0001 0020 0000 0000 0021 0000 0001 0001 0000 0000 0021 0000 0000 0000 0000 0000 0000 0001 0400 0400 0000 0000 0020 〇〇〇〇 0400 0000 0000 0020 0400 0400 0400 0000 0000 0000 0020 0020 0400 0000 0021 〇〇〇〇 0020 0000 0400 0001 0020 0020 0020 0000 0000 0000 0001 0001 0020 0000 0000 〇〇〇〇 0001 0000 0020 0000 0001 0001 0001 0000 0000 0000 0000 0000 0001 0000 0400 〇〇〇〇 0000 0000 0001 0000 0000 0000 0000 0000 0000 0000 0000 0020 0000 0000 0020 〇〇〇〇 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0001 0000 0000 0001 〇〇〇〇 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 〇〇〇〇 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 表一:先前的影像的資料。 0001 0000 〇〇〇〇 0000 0000 0000 7000 7000 0000 0000 0000 0001 0000 0000 0000 0000 0020 0000 0001 0000 0000 0000 0000 7000 0000 0001 0000 0020 0000 0000 0000 0000 0021 0000 0020 0000 0000 0000 0001 7000 0001 0020 0001 0021 0001 0001 0001 0001 0000 0001 0021 0000 0001 0000 0020 7000 0020 0021 0020 0000 0020 0020 0020 0020 0400 0020 〇〇〇〇 0000 0020 0001 0021 7000 0021 0000 0021 0400 0021 0021 0021 0021 0020 0021 0400 0000 0021 0020 0000 7000 0000 0400 0000 0020 0000 0000 0000 0000 0001 0000 0020 0001 0000 0021 0400 7000 0001 0020 0400 0001 0400 0400 0400 0400 0000 0400 0001 0020 0400 0000 0020 7000 0020 0001 0001 0020 0020 0020 0020 0020 0000 0020 0000 0021 0001 0001 0001 7001 0021 0001 0020 0021 0001 0001 0001 0001 0000 0001 〇〇〇〇 0000 0020 0020 0000 7000 0000 0020 0021 0000 0020 0000 0000 0020 0000 0000 〇〇〇〇 0400 0021 0021 0400 7000 0400 0021 0000 0400 0021 0000 0000 0021 0000 0000 〇〇〇〇 0020 0000 0000 0020 7000 0020 0000 0400 0020 0000 0000 0000 0000 0000 0000 〇〇〇〇 0001 0400 0400 0001 7000 0001 0400 0020 0001 0400 0000 0000 0400 0000 0000 〇〇〇〇 0000 0020 0020 0000 7000 0000 0020 0001 0000 0020 0000 0000 0020 0000 0000 0000 0000 0001 0001 0000 7000 0000 0001 0000 0000 0001 0000 0000 0001 0000 0000 〇〇〇〇 0000 0000 0000 7001 7000 0001 0000 0000 0000 0000 0000 0000 0000 表二:目前的影像的資料。 本紙張尺度適用中國國家標準(CNS)A4規格(210X297公釐) ..........#裝.........訂......... (、請先閲讀背面之注意事項再填寫本頁) 582174 A7 B7 五 '發明説明() 0 0 經濟部智慧財產局員工消費合作社印製 表三:經比較後所產生之列圖案的資料。 經比較後寫入LPBUF中的資料: 0x0300, 0x0100, 0x0100, 0x0100, 0x0100, 0x0100, 0x0100, 0x0100, 0x0100, 0x0100, 0x0100, 0x0100, 0x0100, 0x0100, 0x0100, 0x0300, 即為表三中展開的列圖案的資料。 由於全部相異數量為1 8個點,所以根據第二圖中所說 明的判斷方式及預定的TDN門檻,上述之影像比較結果將 採用RLEDMP的編碼方式進行此影像區塊的壓縮方式。 第五圖係說明本發明之較佳實施例RLEDMP演算法的 編碼程序。步驟602,首先判斷TDN的數量,若TDN的數 量小於4則直接進入步驟614,反之若TDN的數量大於等 於4則進入下一個步驟604。步驟6〇4至步驟6〇6,在此稱 之為P方法(P Method),其中步驟604進行列圖案指標(Line Pattern Index)的編碼,而步驟606則進行LpBUF的編碼。 在下一個步驟608中’將LP標示有差異影像晝素點的:素 資料分離成為R、GAB的顏色資料。接下來的步驟 13The fourth figure is a flow chart of the fuzzy comparison procedure. Since the data sampling is unstable by the a / d converter 106, a predetermined value is defined here for comparison. The predetermined value is expressed by a fuzzy comparison range (FCR). When the data transmitted by the A / D converter 106, such as 〇χ1〇, 〇χ1 丨 〇r 〇χ12 can be regarded as the same data, in order to overcome the A / D converter 丨 〇6 image capture Unstable phenomenon 'means that when the difference is less than or equal to 2, it can be regarded as the same. The following formulas are used for fuzzy control calculations: Result = Abs (CR-FR) < = FCR .and. Abs (CG-FG) < = FCR .and. Abs (CB-FB) < = FCR The above CR, CG, and CB represent the current color data of R, G, and B, which are separated from the daytime elements of the current image, and FR, FG, and FB represent the previous r, G and b color data. When the above Result is zero, it means that the diurnal element of the previous image is not the same pixel as the diurnal element of the current image, and other cases represent that the two diurnal elements are the same pixel. First, step 502 is the initial set value of the fuzzy comparison, and the predetermined value of the FCR is set. The line counter (Line counter; LC), dot counter (Dot Counter; DC), line pattern (LP), and All Total Difference Numbers (TDNs) are reset to zero first. Step 5 04 ′ acquire image data of image 1 and image 2 to be compared. Step 506, install: (Please read the precautions on the back before filling out this page) -Order · # Printed by the Employee Consumption Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs 10 582174 V. Description of the invention (point by point comparison of each point in the image Whether the image data are the same. When the image data at this point are found to be the same, go directly to step 5 丨 4, and when the image data at this point are not the same, then enter the fuzzy comparison procedure of the present invention. Step 508 The point and image data are separated into RGB color data according to the separation method in the third figure. The image data at this point in image 1 is divided into R1, B1, and G1, and the image data at the same position in image 2 is divided R2, G2, and b 2. Step 5 1 0. Use formula 1 to fuzzy compare whether the image data at this point is in the acceptable range, that is, whether the RGB values at this point can be regarded as the same and accepted as the same. Data. If it is determined to be the same, go directly to step 5 1 4; if it is not the same, go to step 5 丨 2. Step 5 丨 2, set the LP at this point to 1, and increase the value of TDN by 1, Represents the increase in the total number of distinct Point. Then proceed to step 5 丨 4, move the position of LP, one bit to the left, and increase DC by 1. Step 5 丨 6, it is to determine whether the image data of this row has been compared, and here each row is 1 6 image points are taken as an example, so the DC judges whether it is greater than or equal to 16. If the comparison is not completed, that is, if the DC is less than 16, return to step 5 0 4 and continue to compare the next position point. If the comparison has been made, then DC is greater than or equal to 16 then go to step 5 1 8. Step 5 1 8 is to store the LP value recorded in the above comparison into the LP buffer (LP Buffer; LPBUF), reset LP and DC to zero and add LC丨, that is, enter the next column, and continue the work of image blur comparison, until step 520, complete the image blur comparison work. Using the image blur comparison program of the present invention, when the comparison is completed, a set of columns is generated at the same time The information of the pattern 'will be described in detail in Table 1, Table 2 and Table 3. The comparison methods and comparison results of the present invention are shown in Table 1. Table 1 represents the first π. This paper size applies the Chinese National Standard (CNS) A4 specification (210X297). ........... · Install: / Please read first Read the notes on the back and fill in this page again} Order · # Printed by the Consumers ’Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs 582174 Printed by the Consumers’ Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs A7 B7 Represents the current image data and Table 3 represents the column pattern data generated after comparison. The fuzzy comparison program of the present invention uses a comparison of each column and each row in Table 1 and Table 2 to generate a table The data of the three middle column patterns. 0000 0000 〇〇〇〇0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 〇〇〇〇0000 0001 0001 0001 0000 0000 0001 0001 0001 0000 0000 0000 0000 0001 0001 0000 0020 0020 0020 0000 0000 0020 0020 0020 0001 0000 0000 0000 0000 0020 0020 0000 0021 0021 0021 0000 0000 0021 0021 0021 0020 0000 0000 0000 0000 0021 0021 0000 0000 0000 0000 0000 0000 0000 0000 0000 0021 0001 0001 0000 0000 0000 0000 0001 0400 0400 0400 0001 0001 0001 0400 0400 0000 0020 0020 0001 0000 0400 0400 0020 00 20 0001 0020 0020 0020 0020 0020 0020 0020 0400 0021 0021 0020 0000 0020 0020 0021 0001 0020 0001 0021 0021 0021 0021 0001 0001 0020 0000 0000 0021 0000 0001 0001 0000 0000 0000 0021 0000 0000 0000 0000 0000 0000 0001 0400 0400 0000 0000 0000 0020 〇00〇〇〇〇〇〇〇〇 0400 0000 0000 0020 0400 0400 0400 0000 0000 0000 0020 0020 0400 0000 0021 〇〇〇0020 0000 0400 0001 0020 0020 0020 0000 0000 0000 0001 0001 0020 0000 0000 〇〇〇〇0001 0000 0020 0000 0001 0001 0001 0000 0000 0000 0000 0000 0000 0001 0000 0400 〇00〇000000 0000 0001 0000 0000 0000 0000 0000 0000 0000 0000 0020 0000 0000 0020 〇〇〇〇0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0001 0000 0000 0001 0000 000000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 〇〇〇〇0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 Table 1: Data of previous images. 0001 0000 〇00〇000000 0000 0000 7000 7000 0000 0000 0000 0001 0000 0000 0000 0000 0020 0000 0001 0000 0000 0000 0000 0000 7000 0000 0001 0000 0020 0000 0000 0000 0000 0021 0000 0020 0000 0000 0000 0001 7000 0001 0020 0001 0021 0001 0001 0001 0001 0000 0001 0021 0000 0001 0000 0020 7000 0020 0021 0020 0000 0020 0020 0020 0020 0400 0020 〇〇〇〇0000 0020 0001 0021 7000 0021 0000 0021 0400 0021 0021 0021 0021 0020 0021 0400 0000 0021 0020 0000 7000 0000 0400 0000 0020 0000 0000 0000 0000 0001 0000 0020 0001 0000 0021 0400 7000 0001 0020 0400 0001 0400 0400 0400 0400 0000 0400 0001 0020 0400 0000 0020 7000 0020 0001 0020 0020 0020 0020 0020 0020 0020 0000 0020 0000 0021 0001 0001 0001 0001 7001 0021 0001 0020 0021 0001 0001 0000 0001 〇〇〇〇〇0000 0020 0020 0000 7000 0000 0020 0021 0000 0020 0000 0000 0020 0000 0000 〇〇〇〇〇00400 0021 0021 0400 7000 0400 0021 0000 0400 0021 0000 0000 0021 0000 0000 00000020 0000 0000 0020 7000 0020 0000 0400 0020 0000 0000 0000 0000 0000 0000 〇〇〇〇0001 0400 0400 0001 7000 0001 0400 0020 0001 0400 0000 0000 0400 0000 0000 〇〇〇〇0000 0020 0020 0000 7000 0000 0020 0001 0000 0020 0000 0000 0020 0000 0000 0000 0000 0001 0001 0000 7000 0000 0001 0000 0000 0001 0000 0000 0001 0000 0000 〇00〇0000 0000 0000 7001 7000 0001 0000 0000 0000 0000 0000 0000 0000 Table 2: The current image data. This paper size applies to China National Standard (CNS) A4 specification (210X297 mm) .......... # 装 ......... Order ... (、, Please read the precautions on the back before filling out this page) 582174 A7 B7 Five 'invention description () 0 0 Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs Table 3: The information of the pattern produced after comparison. Data written into LPBUF after comparison: 0x0300, 0x0100, 0x0100, 0x0100, 0x0100, 0x0100, 0x0100, 0x0100, 0x0100, 0x0100, 0x0100, 0x0100, 0x0100, 0x0100, 0x0100, 0x0300. Pattern information. Since the total number of differences is 18 points, according to the judgment method described in the second figure and the predetermined TDN threshold, the above-mentioned image comparison results will use the RLEDMP coding method to compress this image block. The fifth figure illustrates the encoding procedure of the RLEDMP algorithm of the preferred embodiment of the present invention. In step 602, the number of TDNs is determined first. If the number of TDNs is less than 4, the process proceeds directly to step 614; otherwise, if the number of TDNs is greater than or equal to 4, the process proceeds to the next step 604. Steps 604 to 606 are referred to herein as a P method, where step 604 encodes a Line Pattern Index, and step 606 encodes LpBUF. In the next step 608, the LP is marked with the difference image day pixels: the voxel data is separated into R and GAB color data. Next steps 13
;-! ........#裝.........訂......... (請先閲讀背面之注意事項再填寫本頁) 582174 經濟部智慧財產局員工消費合作社印製 A7 B7 五、發明説明() 則進行RLE DM的編碼。再比較編碼後的尺寸是否大於 TDN*3。若編碼後的尺寸大於TDN*3,則進行步驟614之 直接編碼的程序,但若編碼後的尺寸小於等於T d n * 3 ,則 使用RLEDM編碼後的資料。 為更清楚說明本發明之方法,請參考第六圖,第六圖 係為第五圖中步驟6 1 4所述之直接編碼程序的編碼格式。 如圖中所示,前面的字元用來儲存列7〇2的資料,緊接的 字元用來記錄行704的資料其餘於的位元則紀錄晝素7〇6 的資料。 第七圖係為根據本發明之較佳實施例第五圖中步驟 6 1 4之直接編碼程序的流程示意圖。步驟8〇2係讀取起始 資料’將列圖案緩衝器之位址(LPB Adress ; LPBA)與影像 緩衝器位址(Image Buffer Adress; IBA)讀入。步驟804,設 定列計數器之起始值為零。步驟806,依LPBA讀取LP的 資料’ LPBA加2,同時設定行計數器(c〇lumil counter ; cc) 為零。步驟808 ,判斷LP是否為零,倘若為零直接進入步 驟8 1 8,倘若不為零,則進入步驟8丨〇,判斷大部分的 的位το資料是否零,若為零進入步驟816,若不為零進入 步驟812。步驟812,將目前的行計數器與列計數器編碼寫 入輸出字串(Output Stream)中。步驟814,將所需的畫素資 料寫入輸出字串中。步驟816,將LP向左移一個位元及 CC加1。流程再回到步驟808,直到LP = 〇時,進入步驟 818,將LC加1。步驟820 ,判斷LC是否大於等於16,若 小於1 6回到步驟806,若大於等於i 6,則結束此程序。 ...................、一^.........0'L (請先閲讀背面之注意事項再填寫本頁);-! ........ # 装 ......... Order ......... (Please read the notes on the back before filling this page) 582174 Intellectual Property of the Ministry of Economic Affairs A7 B7 printed by the Bureau ’s Consumer Cooperative. V. Invention Description () The RLE DM is coded. Then compare whether the size after encoding is greater than TDN * 3. If the size after encoding is greater than TDN * 3, the direct encoding procedure of step 614 is performed, but if the size after encoding is less than or equal to T d n * 3, the data after RLEDM encoding is used. For a clearer explanation of the method of the present invention, please refer to FIG. 6, which is an encoding format of the direct encoding program described in step 6 1 4 of the fifth figure. As shown in the figure, the preceding characters are used to store the data of column 702, and the following characters are used to record the data of row 704. The remaining bits are used to record the data of day 706. The seventh diagram is a schematic flowchart of the direct encoding procedure of step 6 1 4 in the fifth diagram according to the preferred embodiment of the present invention. Step 802 is to read the initial data ′ and read in the address of the column pattern buffer (LPB Adress; LPBA) and the image buffer address (Image Buffer Adress; IBA). In step 804, the initial value of the column counter is set to zero. In step 806, the data of the LP is read according to the LPBA, and LPBA is increased by two, and at the same time, a row counter (cclumin counter; cc) is set to zero. Step 808, determine whether LP is zero, if it is zero, go directly to step 8 1 8; if it is not zero, go to step 8 丨 0, determine whether most of the bit το data is zero, and if it is zero, go to step 816, if Go to step 812 if it is not zero. In step 812, the current row counter and column counter codes are written into the output stream (Output Stream). Step 814: Write the required pixel data into the output string. Step 816: LP is shifted to the left by one bit and CC is incremented by one. The flow returns to step 808 again, and when LP = 0, it proceeds to step 818 and increases LC by 1. In step 820, it is judged whether the LC is greater than or equal to 16, and if it is less than 16, it returns to step 806. If it is greater than or equal to i 6, the procedure ends. ......... 、 一 ^ ......... 0'L (Please read the precautions on the back before filling this page)
582174 五、 經濟部智慧財產局員工消費合作社印製 發明説明() RLEDMP演算法,將非線性的資料轉換成為線性的資 枓’而由P方法負擔起這個責任。P方法首先建立16位元 的列圖案指標(Line Pattern Index)。每一個位元代表一個列 圖案,當列圖案不為零的時候,相對應的位元將被設定為 1,以代表此列中包含資料。且p方法並將列圖案暫存器進 行壓縮。 第八圖為根據本發明之較佳實施例第五圖中步驟6〇4 之列圖案指標編碼的流程示意圖。首先,步驟9〇2,讀取 列圖案緩衝器位址。步驟9〇4,將列圖案指標(Une Index ; LPI)及一計數器之起始值設定為零。步驟9〇6,讀 取LPB中的LP的資料。步驟908判斷LP的資料是否為零。 當其不為零時,進入步驟9 1 0,設定列圖案指標之最小的 有效位元為1。當LP的資料為零或步驟91〇結束後,進入 步驟9 1 2,列圖案指標向左移一個字元,計數器及列圖案 緩衝器位址均增加。步驟9丨6,如果計數器小於i 6,則回 到步驟906。當計數器累計至16以後,進入步驟9丨8,將 列圖案指標與全部相異數量儲存。 以下將詳細說明本發明之較佳實施例第五圖中步驟 606之列圖案暫存器編碼方法。由於,列圖案指標已經包 含列的資料,在此P方法僅需處理這些LP不等於零的情 況。P方法將每一列圖案分成四個半字節(Nibble)。如果半582174 V. Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economics Invention description () The RLEDMP algorithm converts non-linear data into linear data, and P method bears this responsibility. The P method first establishes a 16-bit Line Pattern Index. Each bit represents a row pattern. When the column pattern is not zero, the corresponding bit will be set to 1 to indicate that the row contains data. And the p method compresses the column pattern register. The eighth figure is a schematic flow chart of pattern index coding in step 604 in the fifth figure according to a preferred embodiment of the present invention. First, in step 902, the column pattern buffer address is read. Step 904: Set the initial value of the column pattern index (Une Index; LPI) and a counter to zero. Step 906: Read the data of LP in LPB. Step 908 determines whether the data of the LP is zero. When it is not zero, it proceeds to step 9 1 0 and sets the minimum effective bit of the column pattern index to 1. When the data of the LP is zero or step 910 is completed, the process proceeds to step 9 1 2 and the column pattern indicator is shifted to the left by one character, and the counter and the column pattern buffer addresses are increased. Step 9 丨 6, if the counter is less than i 6, go back to step 906. When the counter has accumulated to 16, enter step 9 丨 8 to store the column pattern index and all the different numbers. The following will describe in detail the coding method of the pattern register of step 606 in the fifth figure of the preferred embodiment of the present invention. Since the column pattern index already contains column data, the P method only needs to deal with these cases where LP is not equal to zero. The P method divides each column pattern into four nibbles. If half
1 X X X X 如果半字節為零時,但是剩下的半字節並不為零時 輸出位元0與位元1至輸出字串 15 47本紙張尺度適用中國國家標準(CNS)A4規格(210X297公楚) ..........I (請先閲讀背面之注意事項再場寫本頁} -1 XXXX If the nibble is zero, but the remaining nibble is not zero, output bit 0 and bit 1 to output string 15 47 This paper size applies Chinese National Standard (CNS) A4 specification (210X297 Gongchu) .......... I (Please read the notes on the back before writing this page}-
Aw 582174Aw 582174
五、發明説明() 如果剩下的半字節為零時,輸出兩個位元〇至輸出字 〇 0 串 p方法也將目前區塊的位置訊息加以編碼,故可進行這 些分散數據的處理。RLEDMP演算法,透過列圖案資料集 結這些資料,並測試每一個列圖案緩衝器的位元。如果發 現非零的位元,將根據目前位置取得影像暫存器中的晝素 資料。在影像暫存器中的畫素資料,立刻由RGB555被分 離為R、G與B的資料。全部的R資料會被儲存在rgb取 樣暫存器(RGBSampleBuffer)中,全部的g資料則會被儲存 在RGB取樣暫存器(RGBSampleBuffer)中並根據TDN位 移,同樣的,全部的B資料則會被儲存在rgB取樣暫存器 (RGBSampleBuffer)中並根據 TDN * 2 位移。 例如,將以下的影像資料進行處理: P1 P2 P3 P4 P5 P6 P7 P8 其中,P1〜P8係使用 RGB555的格武,分別句.厶V. Description of the invention () If the remaining nibble is zero, output two bits 0 to output word 0 The string p method also encodes the position information of the current block, so it can process these scattered data . The RLEDMP algorithm gathers these data through the row pattern data and tests the bits of each row pattern buffer. If a non-zero bit is found, the daylight data in the image register will be obtained based on the current position. The pixel data in the image register is immediately separated into R, G, and B data by RGB555. All R data will be stored in the rgb sampling buffer (RGBSampleBuffer), all g data will be stored in the RGB sampling buffer (RGBSampleBuffer) and shifted according to TDN. Similarly, all B data will be It is stored in the rgB sampling buffer (RGBSampleBuffer) and shifted according to TDN * 2. For example, the following image data will be processed: P1 P2 P3 P4 P5 P6 P7 P8 Among them, P1 ~ P8 use RGB555 gewu, respectively. 厶
RpiGpiBpi 〜RpsGpsBps 且其值均小於 0x20。 上述之區塊共計有8個不相同的晝素,將其分離並轉 換後,可得下列之線性數據:RpiGpiBpi ~ RpsGpsBps and their values are all less than 0x20. The above blocks have a total of 8 different daylight elements. After separating and converting them, the following linear data can be obtained:
RpiRp2Rp3Rp4Rp5Rp6Rp7Rp8GpiGp2Gp3Gp4Gp5Gp6Gp7Gp8BpiBp2Bp3Bp4Bp5Bp6Bp7Bps 以下列的線性數據為例: 0x08,0x09, 0x09, 0x08, 0x08, 0x08, 0x09, 0x08 16 本紙張尺度適用中國國家標準(CNS)A4規格(210X297公釐) ..........•襞.........訂.........#:、 {請先閲讀背面之注意事項再填寫本頁} 經濟部智慧財產局員工消費合作社印製 582174 A7 B7 五、發明説明( 當原始資料為:0x06,0x06,0x06,0x06時, 裝· _ (請先閲讀背面之注意事項再填寫本頁) 經壓縮後之資料為:6(5位元),1(1位元),3(2位元)。 由於重複次數並不會小於零,故在數量攔位中,以〇 代表一次,依序增加,而旗標值為丨表示壓縮袼式,資料 則以原始資料填入。 但當重複次數大於四次且小於1 29次時,RLE〇 1編碼 的方法利用下列的方法進行編碼。 資料(Base Value) 旗標(Flag) 數量(Unit Amount、 X |X X x X 0 0 N N N N n |n n 當原始資料為:0x06,0x06 ’ 0x06,0x06,0x06,0x06,0x06 時, 經壓縮後之資料為:6(5位元),00(2位元>,6(7位元)。 雖然在此利用數量為7個位元為例,當然其可使用在 任何數量大於2個位元的情況。 接下來再進行RLE02編碼方法的說明。 資料(Bas丨 e Value) 旗標fflag) 資料(Base Value) 數量(Unit Amount) A |a A A A 0 1 1 0 0 B B B B |B n |n |n In |n |n |n 當原始資料為:0x11,0x06,0x11,0x06,0x11,〇χ〇6 時, 經壓縮後之資料為:11(5位元),01(2位元),1〇〇(3位 元),6(5位元),2(7位元)。 經濟部智慧財產局員工消費合作社印製 雖然使用數量為7個位元為例,其可使用在任何數量 的位元情況。 若資料的變化範圍在3至-3之間的時候,DM編碼方 法將會使用下列的編碼方法。 資料差異(Value Difference) 壓縮後的資料(Compressed Code) 0 000 1 001 2 010 18 本紙張尺度適用中國國家標準(CNS)A4規格(210X297公楚) 582174 A7 B7 五、發明説明() 3 Oil -1 101 -2 110 -3 111 EOF ~ 100 資料(Base Value) 旗標(Flag) 壓縮後的資料 結束旗標(End Flag) A A A A A 0 l 0 1 0 0 當原始資料為:0x06,0x07,0x05,0x08,0x05 時, <請先閲讀背面之注意事項再場寫本頁) 經壓縮後之資料為:6(5位元),010(3位元),〇〇1(3位 元),110(3 位元),〇ιι(3 位元),111(3 位元),1〇〇(3 位元)。 參閱第十A圖、十b圖及十C圖,係為根據本發明之 較佳實施例第五圖中步驟6 1 0之RLEDM編碼的流程示意 圖。第十A圖主要描述起始狀態的設定,rle〇 1編碼及結 束流程。 經濟部智慧財產局員工消費合作社印製 步驟1202,起始條件設定,讀取RGBSampleBuffer之 開始位址(RSBA),計算而得RGBSampleBuffer之結束位址 (RGBEA),為RSBA再加上TDN * 3,並將結束字元(End of Buffer Character ; EOBC)力口在 RGBSampleBuffer 之後。設 定編碼模式(EncodeMode)的預設值為Normal,其代表使用 第九圖中之EM—Normal的編碼模式’並設定'^先前三角調 變差異值(Previous Delta-modulation Difference Value ; PrevDM)為0。最後,讀取RGBSampleBuffer之一位元組以 設定先前字元(Previous Character ; PrevCH)的起始值,並 將RSBA增加1。 步驟1 204,判斷RSBA是否大於等於RSBEA。當RSBA 大於等於RSBEA時,接步驟1206,再判斷EncodeMode是 19 本紙張尺度適用中國國家標準(CNS)A4規格(210X 297公釐) 582174 A7 B7 經濟部智慧財產局員工消費合作社印製 五、發明説明( 否等於EM一PRE — DM,如果相等,則進入步驟1208,送出 01至輸出字串。步驟1212,輸出prevD Μ的三個位元至輸 出字串,與步驟1214 ,送出1〇〇至輸出字串。然後,步 驟1 2 1 6將編碼暫存器中的資料清除。但如果步驟1 2 〇 6中, EncodeMode與EM_PRE —DM並不相同時,則進入步驟 1210 ’以判斷EncodeMode是否與EM_DM相同,如果相同 進入步驟1 2 1 2 ’但是如果其並不相同,則進入步驟i 2 1 6。 其中步驟1204中,如果rsBA小於RSBEA時,則進 入步驟1218,將重複次數(Repeat c〇unter ; RC)設定為1。 此流程將決定一先前字元(prevCh)之重複的次數。步驟 1220,現有字元(Current Character; Ch)由 RGBSampleBuffer 中讀取’並將RSBA加1。步驟1 222,判斷Ch是否與PrevCh 相同’如果相同進入步驟1224,將RC加1,用來計算連 續的相同的字元數量,並再回到步驟122〇。如果Ch與 PrevCh並不相同,則進入步驟1226,判斷RC是否等於1。 如果等於1,則進入第十B圖中的步驟1302,而如果不等 於1,則此流程將依據下列的說明方式,決定使用哪一個 編碼流程較為合適。 此流程如何決定使用何種編碼方式進行這些重複的資 料的編碼。首先,步驟1230,判斷Enc〇deM〇de是否等於 EM-NORMAL。如果 EncodeMode 等於 EM_NORMAL,進入 步驟1254,判斷RC是否小於等於4。如果RC小於等於4 則進入步驟1256,將RC減1。步驟1 258,將PrevCh的值 向左移三個位元並將其加上RC的值。步驟1260,將8位 ...................、可.........# (I請先閲讀背面之注意事項再場寫本頁) 20 582174RpiRp2Rp3Rp4Rp5Rp6Rp7Rp8GpiGp2Gp3Gp4Gp5Gp6Gp7Gp8BpiBp2Bp3Bp4Bp5Bp6Bp7Bps Take the following linear data as examples: 0x08, 0x09, 0x09, 0x08, 0x08, 0x08, 0x08, 0x08 .. • 襞 ......... Order ......... # :, {Please read the notes on the back before filling out this page} Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs 582174 A7 B7 V. Description of the invention (when the original data is: 0x06, 0x06, 0x06, 0x06, please install _ (Please read the precautions on the back before filling in this page) The compressed data is: 6 (5 bits) , 1 (1 bit), 3 (2 bits). Because the number of repetitions will not be less than zero, in the number block, 0 is represented once, and it is increased in order, and the flag value 丨 means compression. , The data is filled with the original data. But when the number of repetitions is more than four times and less than 129 times, the method of RLE〇1 encoding uses the following methods to encode. Base Value Flag Amount (Unit Amount) , X | XX x X 0 0 NNNN n | nn When the original data is: 0x06, 0x06 '0x06, 0x06, 0x06, 0x06, 0x06, the compressed data is: 6 (5 bits), 00 (2 bits >, 6 (7 bits). Although the number used here is 7 bits As an example, of course, it can be used in any case where the number is greater than 2 bits. Next, the RLE02 encoding method will be described. Data (Bas 丨 e Value) Flag fflag) Data (Base Value) Amount (Unit Amount) A | a AAA 0 1 1 0 0 BBBB | B n | n | n In | n | n | n When the original data is: 0x11, 0x06, 0x11, 0x06, 0x11, 0x〇6, the compressed data For: 11 (5 bits), 01 (2 bits), 100 (3 bits), 6 (5 bits), 2 (7 bits). Although printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs, The number of bits is 7 as an example, it can be used in any number of bits. If the range of the data is between 3 and -3, the DM encoding method will use the following encoding method. Data difference ( Value Difference) Compressed Code 0 000 1 001 2 010 18 This paper size applies to China National Standard (CNS) A4 (210X297) Chu 582174 A7 B 7 V. Description of the invention (3) Oil -1 101 -2 110 -3 111 EOF ~ 100 Data (Base Value) Flag Compressed Data End Flag AAAAA 0 l 0 1 0 0 When The original data is: 0x06, 0x07, 0x05, 0x08, 0x05, < Please read the notes on the back before writing this page) The compressed data is: 6 (5 bits), 010 (3 bits) 001 (3 bits), 110 (3 bits), 00 (3 bits), 111 (3 bits), 100 (3 bits). Refer to the tenth diagram A, ten b, and ten c, which are schematic diagrams of the RLEDM coding of step 6 10 in the fifth diagram according to the preferred embodiment of the present invention. The tenth figure A mainly describes the setting of the initial state, rleO 1 encoding and the end process. The Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs prints step 1202, sets the initial conditions, reads the starting address (RSBA) of RGBSampleBuffer, and calculates the ending address (RGBEA) of RGBSampleBuffer, which is RSBA plus TDN * 3, The end character (End of Buffer Character; EOBC) is behind the RGBSampleBuffer. Set the default encoding mode (EncodeMode) to Normal, which means to use the EM-Normal encoding mode in the ninth figure, and set '^ Previous Delta-modulation Difference Value (PrevDM) to 0 . Finally, read a byte of RGBSampleBuffer to set the starting value of the previous character (PrevCH), and increase the RSBA by one. Step 1 204: Determine whether RSBA is greater than or equal to RSBEA. When RSBA is greater than or equal to RSBEA, then proceed to step 1206, and then judge that the EncodeMode is 19. The paper size is applicable to the Chinese National Standard (CNS) A4 specification (210X 297 mm) 582174 A7 B7 Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs Explanation (No is equal to EM_PRE — DM. If they are equal, go to step 1208, and send 01 to the output string. Step 1212, output the three bits of prevD Μ to the output string, and step 1214, send 100 to The output string. Then, the data in the encoding register is cleared in step 1 2 1 6. However, if the EncodeMode and EM_PRE —DM are not the same in step 1 2 06, then proceed to step 1210 'to determine whether the EncodeMode is the same as EM_DM is the same, if they are the same, go to step 1 2 1 2 'but if they are not the same, then go to step i 2 1 6. Wherein in step 1204, if rsBA is less than RSBEA, then go to step 1218 and repeat the number of times (Repeat c. unter; RC) is set to 1. This process will determine the number of repetitions of a previous character (prevCh). Step 1220, the existing character (Current Character; Ch) is read from RGBSampleBuffer 'and RSBA adds 1. Step 1 222, determine whether Ch is the same as PrevCh 'If the same goes to step 1224, add RC to 1 to calculate the number of consecutive identical characters, and then return to step 122. If Ch and PrevCh are combined If not, go to step 1226 to determine whether RC is equal to 1. If it is equal to 1, go to step 1302 in the tenth B diagram, and if it is not equal to 1, then this process will decide which encoding to use according to the following description. The process is more appropriate. How does this process decide which encoding method to use to encode these repeated data. First, step 1230, determine whether EncodeMode is equal to EM-NORMAL. If EncodeMode is equal to EM_NORMAL, proceed to step 1254, determine whether RC Less than or equal to 4. If RC is less than or equal to 4, then go to step 1256 and reduce RC by 1. Step 1 258, shift the value of PrevCh three bits to the left and add the value of RC. Step 1260, 8 bits. ........ 、 可 ......... # (I Please read the notes on the back before writing this page) 20 582174
..........•裝: f請先閲讀背面之注意事項再填寫本Ϊ -訂· 兀的PreVCh送出至輸出字串。步驟1270,將PrevCh設定 等於Ch,以追蹤下一個連續字元。流程再回到步驟丨2〇4 中。如果步驟1254中,RC大於4,則進入步驟丨262。步 驟1262,將RC減1。然後步驟1264,將PrevCh向左移二 個位兀。步驟1266,將七個位元的PrevCh送出至輸出字 串中。步驟1268,將七個位元的Rc送出至輸出字串中。 然後,步驟1270,將prevCh設定等於ch,流程並回到步 驟1204中。 而其中之步驟1230中,若EncodeMode不等於 EM_NORMAL時’進入步驟1232,判斷RC是否大於七。 如果RC大於七,則進入步驟1252,送出1〇〇至輸出字串 中’並重新設定EncodeMode為EM一NORMAL。而當RC不 大於七時,則進入步驟1234。步驟1234,比較PrevCh減 去Ch的絕對值,是否大於三角調變差異範圍(Deha.......... • Installation: f Please read the precautions on the reverse side before filling out this pre-order. PreVCh is sent to the output string. In step 1270, PrevCh is set equal to Ch to track the next consecutive character. The process returns to step 丨 204. If the RC is greater than 4 in step 1254, proceed to step 262. In step 1262, RC is decreased by one. Then in step 1264, PrevCh is shifted two positions to the left. In step 1266, the seven-bit PrevCh is sent to the output string. In step 1268, the seven-bit Rc is sent to the output string. Then, in step 1270, prevCh is set equal to ch, and the flow returns to step 1204. In step 1230, if EncodeMode is not equal to EM_NORMAL, the process proceeds to step 1232 to determine whether the RC is greater than seven. If RC is greater than seven, go to step 1252, send 100 to the output string 'and reset EncodeMode to EM_NORMAL. When RC is not greater than seven, step 1234 is entered. Step 1234, comparing whether the absolute value of PrevCh minus Ch is greater than the range of the delta modulation difference (Deha
Modulation Difference Range ; DMRANGE)。若絕對值 (PrevCh-Ch)大於DMRANGE時,且步驟1236中之RC大於 等於3時,流程回到步驟1 252。而當步驟1236中之RC小 於3時,進入步驟 1238,判定EncodeMode是否等於 EM — PRE —DM。若 EncodeMode 等於 EM — PRE — DM,進入步 驟1272將10送出至輸出字串,並進入步驟1274,將此流 程送出PrevDM的三個位元、000及1〇〇至輸出字串中,並 設定 PrevCh 等於 Ch。而當現有的編碼模式不是 EM_PRE_DM時,直接進入步驟1274。當步驟1234中之絕 對值(PrevCh-Ch)不大於DMRANGE時,此流程進入步驟 本紙張尺度適用中國國家標準(CNS)A4規格(210X297公釐) i 經濟部智慧財產局員工消費合作社印製 582174 A7 B7 五、發明説明() 1240,判斷現有的編碼模式是否為EM — PRE — DM。當現有 的編碼模式不是為EM一PRE一DM時,進入步驟1244,將送 出PrevDM的三個位元至輸出字串中,且RC減1。而當現 有的編碼模式是EM_PRE__DM時,進入步驟1242,送出10 至輸出字串中,並設定EncodeMode為EM —DM,然後再進 入步驟1244。在步驟1244之後,流程進入步驟1246,判 斷RC是否等於零。當RC不是等於零時,進入步驟1248, 此流程將重複送出000至輸出字串中,並將RC減1,直到 RC等於零時,進入步驟1250,PrevDM設定為Ch減去 PrevCh,並將PrevCh再設定等於Ch,接著流程回到步驟 1204 中。 第十B圖主要係說明RLE02編碼。首先,步驟1 302, 儲存輸入字串現有的位置RSBA,接著進入步驟1304,由 RSB A讀取兩個字元,將第一個字元寫入為下一字元(Next Character ; NextCh),並將 RSBA 加 1。步驟 1 306,將第二 個字元送出至比較字元(Comparison Character ; CCh),並 將RSBA再加1。步驟1 308,判斷PrevCh是否等於NextCh, 倘若PrevCh等於NextCh,及步驟1310,判斷Ch是否等於 CCh。若步驟1 308或步驟1310中任一不成立的情況,進 入步驟1312’將RSBA減二。倘若兩者均成立的情況下進 入步驟1314,係將RC設定為二。步驟1316、1318、1322、 Π24及1326,比較接下來RSBA的兩個字元與PrevCh及 Ch,以計算連續重複的字元。當計算完成後,如果有任何 差異發生時,流程進入步驟1 3 3 2。步驟1 3 3 2,比較RC是 ----------ijil· I (請先閲讀背面之注意事項再填寫本頁} -訂· 經濟部智慧財產局員工消費合作社印製8··! 22 582174 A7 ___B7 五、發明説明() 否大於2,如果RC大於2,接下來進入步驟1338,比較目 前的編碼模弍是否為EM一NORMAL,如果是,此流程將把 相關的資料根據RLE02的規則,送出至輸出字串中。當步 驟1332中,RC不大於2時,進入步驟1334,比較絕對值 (PrevCh-Ch)是否大於 DMRANGE。當絕對值(prevch-Ch)大 於D MR AN GE時’進入步驟1336,儲存輸入字串的位置, 並進入第十C圖中。如果步驟1338中,目前的編碼模式不 是EM 一 NORMAL時,送出1〇〇至輸出字串中,步驟134〇。 根據RLE02編碼規則,首先進入步驟1 342,比較RC 是否等於零,如果RC等於零時,進入步驟1360。步驟 13 60,由輸人字串中取得資料,並儲存至PrevCh,再將rSBA 加1與設定目前的編碼模式為EM—NORMAL。當步驟1342 中,RC不等於零時,且步驟1344,RC大於128時,設定 計數器等於127。當步驟1344中,RC不大於128時,計 數器等於RC減卜接著進入步驟1350、1352、1354及1356, 送出PrevCh的五個位元、1〇與1〇〇、ch的五個位元及計 數器的七個位元,至輸出字串。步驟1358將RC設定為RC 減去計數器減去1,並回到步驟1342,直到RC等於零。 第十C圖主要係說明DM的編碼。首先,步驟1402, 比較絕對值(PrevCh-Ch)是否大於DMRANGE。當絕對值 (PrevCh-Ch)不大於DMRANGE時,進入步驟1404,比較目 前的編碼模式是否為EM一NORMAL。當目前的編碼模式是 EM—NORM AL時’進入步驟1406,將EncodeMode設定為 EM — PRE — D Μ。步驟1 40 8,送出PrevCh的五個位元至輸出 1 裝: (請先閲讀背面之注意事項再填寫本頁) -訂- 經濟部智慧財產局員工消費合作社印製8" 23Modulation Difference Range; DMRANGE). If the absolute value (PrevCh-Ch) is greater than DMRANGE and the RC in step 1236 is greater than or equal to 3, the flow returns to step 1 252. When the RC in step 1236 is less than 3, it proceeds to step 1238 to determine whether EncodeMode is equal to EM — PRE — DM. If EncodeMode is equal to EM — PRE — DM, go to step 1272 and send 10 to the output string, and then go to step 1274 to send the three bits of PrevDM, 000 and 100 to the output string, and set PrevCh Equal to Ch. When the existing encoding mode is not EM_PRE_DM, it proceeds directly to step 1274. When the absolute value (PrevCh-Ch) in step 1234 is not greater than DMRANGE, the process enters this step. The paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm). I. A7 B7 V. Description of the invention () 1240, to determine whether the existing coding mode is EM — PRE — DM. When the existing encoding mode is not EM-PRE-DM, the process proceeds to step 1244, and three bits of PrevDM are sent to the output string, and RC is decremented by one. When the existing encoding mode is EM_PRE__DM, it proceeds to step 1242, sends 10 to the output string, and sets EncodeMode to EM-DM, and then proceeds to step 1244. After step 1244, the flow proceeds to step 1246, and it is determined whether RC is equal to zero. When RC is not equal to zero, proceed to step 1248. This process will repeatedly send 000 to the output string and reduce RC by 1. Until RC is equal to zero, proceed to step 1250. PrevDM is set to Ch minus PrevCh, and PrevCh is set again. Equal to Ch, then the flow returns to step 1204. The tenth B diagram mainly illustrates RLE02 encoding. First, in step 1 302, the current position RSBA of the input string is stored, and then the process proceeds to step 1304. RSB A reads two characters and writes the first character as the next character (Next Character; NextCh). And increase RSBA by 1. Step 1 306: Send the second character to the comparison character (Comparison Character; CCh), and add 1 to RSBA. In step 1308, it is determined whether PrevCh is equal to NextCh. If PrevCh is equal to NextCh, and in step 1310, it is determined whether Ch is equal to CCh. If any of step 1308 or step 1310 does not hold, proceed to step 1312 'to reduce the RSBA by two. If both are true, go to step 1314, set RC to two. Steps 1316, 1318, 1322, Π24, and 1326, compare the next two characters of RSBA with PrevCh and Ch to calculate consecutively repeated characters. When the calculation is completed, if any difference occurs, the process proceeds to step 1 3 3 2. Step 1 3 3 2, compare RC is ---------- ijil · I (Please read the notes on the back before filling out this page} -Order · Printed by the Employee Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs 8 · ·! 22 582174 A7 ___B7 V. Description of the invention () No is greater than 2, if RC is greater than 2, then proceed to step 1338 to compare whether the current encoding mode is EM_NORMAL. If so, this process will be based on relevant information The rule of RLE02 is sent to the output string. When RC is not greater than 2 in step 1332, proceed to step 1334 and compare whether the absolute value (PrevCh-Ch) is greater than DMRANGE. When the absolute value (prevch-Ch) is greater than D MR AN When GE, go to step 1336, store the position of the input string, and enter the tenth figure C. If the current encoding mode is not EM_NORMAL in step 1338, send 100 to the output string, step 134. According to the RLE02 coding rules, first enter step 1 342 and compare whether RC is equal to zero. If RC is equal to zero, proceed to step 1360. Step 13 60: Obtain the data from the input string and store it in PrevCh, then increase rSBA by 1 and Set the current encoding mode to EM— NORMAL. When RC is not equal to zero in step 1342 and RC is greater than 128 in step 1344, set the counter to 127. When RC is not greater than 128 in step 1344, the counter is equal to RC decrement and then proceed to steps 1350, 1352, 1354 And 1356, send the five bits of PrevCh, 10 and 100, the five bits of ch and the seven bits of the counter to the output string. Step 1358 sets RC to RC minus the counter minus 1 , And return to step 1342 until RC is equal to zero. The tenth figure C mainly describes the encoding of DM. First, step 1402, compare whether the absolute value (PrevCh-Ch) is greater than DMRANGE. When the absolute value (PrevCh-Ch) is not greater than DMRANGE If the current encoding mode is EM_NORM AL, when the current encoding mode is EM_NORM AL, go to step 1406 and set the EncodeMode to EM_PRE_DM. Step 1 40 8 and send PrevCh's five bits to output 1 Pack: (Please read the precautions on the back before filling out this page) -Order-Printed by the Intellectual Property Bureau of the Ministry of Economic Affairs and Consumer Cooperatives 8 " 23
582174 A7 五、發明説明() 字串,步驟1418,將PrevDM設定為Ch減去Prevch,及 步驟1434,將PrevCh設定為Ch。然後此流程回到第十a 圖中之步驟1204。582174 A7 V. Description of the invention () string, step 1418, set PrevDM to Ch minus Prevch, and step 1434, set PrevCh to Ch. Then the process returns to step 1204 in the tenth a diagram.
如果步驟1404中,當目前的編碼模式不是 EM—NORMAL·時,進入步驟1410,比較目前的編碼模式是 否為EM 一 PRE —DM。如果是,進入步驟1412,將EncodeMode 設疋為DM’步驟1414與步驟1416,送出1〇與prevDM 的三個位元至輸出字串。如果步驟1410中,目前的編碼模 式不是EM — PRE —DM時,流程直接進入步驟ι416,接著進 入步驟1 4 1 8。 如果步骤1 402中’絕對值(PrevCh-Ch)大於DM RANGE 時,進入步驟1418 ’比較目前的編碼模式是否為 EM —PRE — DM。如果是,進入步驟1420、步驟1422與步驟 1424,送出10、prevDM的三個位元與i 00至輸出字串。 進入步驟1432,將EncodeMode設定為EM一NORMAL,接 著流程進入步驟1434。如果步驟1418中,目前的編碼模 式不是EM一PRE — DM時,流程進入步驟1426,比較目前的 編碼模式是否為EM一DM。如果是EM_DM,進入步驟1422。 如果不是EM —DM,進入步驟1428,將PrevCh的值位移三 個位元並加上100,然後進入步驟1 430,將PrevCh的八個 位元送出至輸出字串中。流程接著進入步驟1 4 3 2。 如第一圖中所述之遠端控制系統,當使用本發明之改 善影像壓縮及傳送的方法時,具有許多的優點。倘若欲傳 送一電腦影像具有800*600的影像晝素,而每一個畫素使 裝: (請先閲讀背面之注意事項再填寫本頁) -訂· # 經濟部智慧財產局員工消費合作社印製Q-5; 24If in step 1404, when the current coding mode is not EM_NORMAL, go to step 1410, and compare whether the current coding mode is EM_PRE_DM. If yes, go to step 1412, set EncodeMode to DM ’, step 1414 and step 1416, and send three bits of 10 and prevDM to the output string. If in step 1410, the current encoding mode is not EM — PRE — DM, the flow directly proceeds to step 416, and then proceeds to step 1 4 1 8. If the 'absolute value (PrevCh-Ch)' in step 1 402 is greater than DM RANGE, go to step 1418 'to compare whether the current encoding mode is EM — PRE — DM. If yes, go to step 1420, step 1422, and step 1424, and send the three bits of 10 and prevDM and i 00 to the output string. Go to step 1432, set EncodeMode to EM_NORMAL, and then go to step 1434. If the current encoding mode is not EM-PRE-DM in step 1418, the flow proceeds to step 1426 to compare whether the current encoding mode is EM-DM. If it is EM_DM, go to step 1422. If it is not EM-DM, proceed to step 1428, shift the value of PrevCh by three bits and add 100, and then proceed to step 1 430 to send the eight bits of PrevCh to the output string. The flow then proceeds to steps 1 4 3 2. The remote control system shown in the first figure, when using the method for improving image compression and transmission of the present invention, has many advantages. If you want to send a computer image with 800 * 600 image day pixels, and each pixel is installed: (Please read the precautions on the back before filling out this page) -Order · # Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs Q-5; 24
582174 A7 B7 五、發明説明( 用1 6個位元。假設使用JPEG演算法具有1 5 : i的壓縮比, 經壓縮後之影像資料,約為64,000字元大小。如果有2000 個晝素被分配在200個16*16的區塊中,若僅使用JPEG 演算法,約需要6726個位元組。如果使用RLEDMP時, 即使僅利用直接編碼法進行壓縮,經壓縮後的資料大小約 僅為3000位元組。若RLEDMP的壓縮率為2 : 1時,則僅 需1 500位元組。故使用本發明之改善影像壓縮及傳送的方 法’可大幅降低影像信號傳輸速率及網路頻寬的需求,使 得在網路上有效傳輸壓縮影像,變得更為有效率。 本發明提供一種結合變動長度編碼與三角調變編碼的 演算法,更提供一種結合兩種編碼的可行方式。本發明之 改善影像壓縮及傳送的方法,可跟據不同的需求,決定所 需的壓縮方法’充分利用母一種壓縮方法的優點,有效降 低影像傳輸速率及網路頻寬的需求。 如熟悉此技術之人員所暸解的,以上所述僅為本發明 之較佳實施例而已,並非用以限定本發明之申請專利範 圍;凡其它未脫離本發明所揭示之精神下所完成之等效改 變或修飾,均應包含在下述之申請專利範圍内。 ......•裝.........訂.........#’ (請先閲讀背面之注意事項再填寫本頁) 經濟部智慧財產局員工消費合作社印製582174 A7 B7 V. Description of the invention (16 bits are used. Assuming that the JPEG algorithm has a compression ratio of 15: i, the compressed image data is about 64,000 characters in size. If there are 2000 daylight elements, Allocated in 200 16 * 16 blocks, if only the JPEG algorithm is used, about 6726 bytes are required. If RLEDMP is used, even if only the direct encoding method is used for compression, the compressed data size is only about 3000 bytes. If the compression ratio of RLEDMP is 2: 1, only 1,500 bytes are required. Therefore, the method for improving image compression and transmission of the present invention can greatly reduce the image signal transmission rate and network bandwidth. The requirements of the present invention make it more efficient to efficiently transmit compressed images on the Internet. The present invention provides an algorithm combining variable length coding and triangular modulation coding, and a feasible way to combine two coding methods. Improve the method of image compression and transmission. According to different needs, determine the required compression method. 'Make full use of the advantages of a mother compression method, effectively reduce the image transmission rate and network bandwidth. As understood by those familiar with this technology, the above is only a preferred embodiment of the present invention, and is not intended to limit the scope of patent application for the present invention; all others can be completed without departing from the spirit disclosed by the present invention Equivalent changes or modifications should be included in the scope of patent application as described below ....... • 装 ......... Order ......... # '(Please first (Read the notes on the back and fill out this page) Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs
Claims (1)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW91118121A TW582174B (en) | 2002-08-12 | 2002-08-12 | Method for image compression |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW91118121A TW582174B (en) | 2002-08-12 | 2002-08-12 | Method for image compression |
Publications (1)
Publication Number | Publication Date |
---|---|
TW582174B true TW582174B (en) | 2004-04-01 |
Family
ID=32960531
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
TW91118121A TW582174B (en) | 2002-08-12 | 2002-08-12 | Method for image compression |
Country Status (1)
Country | Link |
---|---|
TW (1) | TW582174B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101998154B (en) * | 2009-08-28 | 2013-02-27 | 宏正自动科技股份有限公司 | Multi-computer switcher for transmitting image signals to a plurality of far-end control computers and method thereof |
-
2002
- 2002-08-12 TW TW91118121A patent/TW582174B/en not_active IP Right Cessation
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101998154B (en) * | 2009-08-28 | 2013-02-27 | 宏正自动科技股份有限公司 | Multi-computer switcher for transmitting image signals to a plurality of far-end control computers and method thereof |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108141505B (en) | Compression and decompression method for high bit depth medical gray level image | |
WO2018184458A1 (en) | Picture file processing method and device, and storage medium | |
CN101107838B (en) | Image processing device, image processing method, and imaging device | |
CN102821279A (en) | Picture compression method keeping with alpha channel | |
CN102289829A (en) | Method and device for storing and restoring image and system for processing image | |
WO2020135357A1 (en) | Data compression method and apparatus, and data encoding/decoding method and apparatus | |
CN103841424B (en) | The system and method for compressed data in random access memory | |
US20130251261A1 (en) | Method And Apparatus For Image Data Compression | |
CN103124350B (en) | Based on method for compressing image and the device of RLE | |
WO2021031877A1 (en) | Methods and apparatus for image coding and decoding, and chip | |
TW201228247A (en) | Image encoding device, and image encoding method | |
WO2023246047A1 (en) | Jpeg image compression method and system, device, and storage medium | |
CN1984228A (en) | Image forming system and method | |
CN108881915B (en) | Device and method for playing video based on DSC (differential scanning sequence) coding technology | |
CN106817584A (en) | A kind of MJPEG compressions implementation method and FPGA based on FPGA | |
KR102626217B1 (en) | Frame buffer compressor and Image processing device comprising the same | |
KR101100829B1 (en) | System and method for lossless compression and decompression of image data | |
CN104575427A (en) | High-compression-ratio lossless image compression method suitable for embedded system decoding | |
TW582174B (en) | Method for image compression | |
US11153586B2 (en) | Image processing device and frame buffer compressor | |
JP2004032698A (en) | Differential image compressing method and apparatus | |
CN109413445B (en) | Video transmission method and device | |
JP2003189109A (en) | Image processor and image processing method, and computer program | |
US7142721B2 (en) | Method of image compression | |
US20050018910A1 (en) | Method and apparatus for reducing the bandwidth required to transmit image data |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
MK4A | Expiration of patent term of an invention patent |