200307457 Ο) 玖、發明說明 【發明所屬之技術領域】 本發明乃關於’檢出以影像信號所顯示的晝像角度之 畫像角度檢出裝置及具備此之掃描線內插裝置，以及畫像 角度檢出方法。 【先前技術】 爲了將父錯ί币描（I n t e r 1 a c e S c a η n i n g )的影像信號轉 換爲漸進式描（P r o g r e s s i v e S c a η n i n g )的影像信號，此 外’爲了將漸進式掃描的影像信號轉換爲放大或是縮小的 景> 像信號’而採用進行掃描線的內插處理之內插電路。於 如此的內插電路中’基於由內插處理所必須製作的像素（ 以下稱爲內插像素）的周圍像素之値，來算出內插像素之 値。例如，在具有斜方向的邊緣的畫像或是具有細斜線的 畫像當中‘從周圍像素的亮度分布來檢出畫像的角度，並 採用位於相關性較高的方向上的像素，來算出內插像素之 値。 於日本特開平9 - 3 7 2 1 4號公報中，揭示了可於具有斜 方向的邊緣的畫像當中，進行掃描線內插之漸進掃描線內 插裝置。 於此漸進掃描線內插裝置中，在將交錯掃描的影像信 號轉換爲漸進式掃描的影像信號之際，從位於以內插像素 爲中心之具有點對稱關係的原像素的組合當中，選擇用於 算出像素値的差分絕對値之候選的原像素組合。然後各自 -5- (2) (2)200307457 算出所選擇的組合的像素値之差分絕對値，並基於各個組 合的個別原像素的邊緣資訊，來校正這些差分絕對値，在 檢出校正後的差分絕對値爲最小的原像素組合後，基於檢 出的原像素組合來製作內插像素。 如此，可對具有斜邊緣的畫像進行掃描線內插。 然而，於以往的漸進掃描線內插裝置中，無法辨識出 畫像的邊緣爲直線形狀或是圓弧形狀等曲線形狀。因此， 在具有圓弧形狀等曲線形狀的邊緣之畫像中進行內插處理 的情況下，則無法得到平滑圓潤的畫像。 【發明內容】 本發明的目的在於，提供可以正確的檢出以影像信號 來顯示的畫像角度及形狀之畫像角度檢出裝置。 本發明的其他目的在於，提供可以進行適合於以影像 信號來顯示的畫像角度及形狀的內插之掃描線內插裝置。 本發明的其他目的在於，提供可以正確的檢出以影像 信號來顯示的畫像角度及形狀之畫像角度檢出方法。 依循本發明的型態之畫像角度檢出裝置，乃基於所輸 入的影像信號，來檢出關於掃描線之間的各條內插掃描線 中之應予內插的像素之畫像角度，其特徵爲具備：於包含 多數的掃描線及應予內插的像素之所定檢出區域內，對所 輸入的影像信號進行2値化處理而產生2値化模式之2値 化模式產生器；以及產生具有不同方向的2値畫像，以做 爲多數的參照模式之參照模式產生器；以及將藉由2値化 (3) (3)200307457 模式產生器所產生的2値化模式，與藉由參照模式產生器 所產生的多數的各個參照模式比較，並基於比較結果，來 檢出關於應予內插的像素的畫像角度之比較器；以及基於 比較器所檢出之關於應予內插的像素的畫像角度、及於上 方與下方的內插掃描線中所檢出的畫像角度之組合，來檢 出畫像的形狀之形狀檢出器。 於本發明的畫像角度檢出裝置中，藉由2値化模式產 生器，對所輸入的影像信號於所定的檢出區域內進行2値 化處理，來產生2値化模式。此外，藉由參照模式產生器 ，來產生具有多數方向的2値化畫像，以做爲多數的參照 模式。然後，藉由比較器，比較2値化模式與多數的各個 參照模式，並基於比較結果，來檢出關於應予內插的像素 的畫像角度。再者，基於比較器所檢出之關於應予內插的 像素的畫像角度、及於上方與下方的內插掃描線中所檢出 的畫像角度之組合，藉由形狀檢出器來檢出畫像的形狀。 於此情況下，因爲進行二次元的模式比較，因此與採 用2像素之間的差分値的情況相較，可抑制錯誤的檢出， 並可正確的檢出具有斜方向的邊緣之畫像角度。此外，藉 由採用二次元的參照模式，不僅可檢出，連接檢出角度位 於以內插像素爲中心之點對稱的位置上之像素之間的直線 角度，還可以檢出位於這些角度之間的角度。因此，可於 更精密的間隔檢出角度。 此外，若是想要藉由比較2値化模式與多數的各個參 照模式來檢出畫像的曲線形狀的話’則必須具有包含至少 200307457200307457 〇). Description of the invention [Technical field to which the invention belongs] The present invention relates to an image angle detection device that detects a daylight image angle displayed by an image signal, a scanning line interpolation device having the same, and an image angle detection device.出 方法。 Out method. [Prior art] In order to convert the image signal of the parent coin (Inter 1 ace Sca ning) into the image signal of progressive tracing (Progressive Sca ning), in addition, 'in order to convert the image of progressive scanning The signal is converted into an enlarged or reduced scene > image signal 'and an interpolation circuit is used to perform interpolation processing of the scanning lines. In such an interpolation circuit, 値 calculates the 像素 of the interpolated pixel based on the 周围 of the surrounding pixels of the pixel (hereinafter referred to as the interpolated pixel) that must be produced by the interpolation process. For example, in an image with oblique edges or an image with thin oblique lines, 'the angle of the image is detected from the brightness distribution of surrounding pixels, and the pixels located in the direction with high correlation are used to calculate the interpolated pixels.値. Japanese Patent Application Laid-Open No. 9-3 7 2 1 4 discloses a progressive scan line interpolation device capable of performing scan line interpolation in an image having oblique edges. In this progressive scanning line interpolation device, when an interlaced scanning image signal is converted into a progressive scanning image signal, it is selected from a combination of original pixels having a point-symmetric relationship with the interpolated pixels as the center. Calculate the original pixel combination of the candidate of the absolute difference of the pixel 値. Then -5- (2) (2) 200307457 respectively calculate the difference absolute 値 of the pixels 所 of the selected combination, and correct these difference absolute 基于 based on the edge information of the individual original pixels of each combination. After detecting the corrected 値After the original pixel combination with the smallest absolute difference, the interpolation pixel is produced based on the detected original pixel combination. In this way, scan-line interpolation can be performed on an image with a beveled edge. However, in the conventional progressive scanning line interpolation device, it is impossible to recognize that the edge of the image is a linear shape or a curved shape such as an arc shape. Therefore, when an interpolation process is performed on an image having curved edges such as an arc shape, a smooth and round image cannot be obtained. SUMMARY OF THE INVENTION An object of the present invention is to provide an image angle detection device that can accurately detect an image angle and a shape displayed by a video signal. Another object of the present invention is to provide a scanning line interpolation device capable of performing interpolation of image angles and shapes suitable for display by video signals. Another object of the present invention is to provide an image angle detection method capable of accurately detecting an image angle and a shape displayed by a video signal. The image angle detection device according to the form of the present invention detects the image angles of pixels to be interpolated in each of the interpolated scanning lines between the scanning lines based on the input image signal, and its characteristics In order to provide: a 2D mode generator that generates a 2D mode by performing 2D processing on an input image signal in a predetermined detection area including a plurality of scan lines and pixels to be interpolated; and 2 値 portraits with different directions as the reference pattern generator for most reference patterns; and 2 値 patterns generated by the 2 値 (3) (3) 200307457 pattern generator, and by reference A comparison of most of the reference patterns generated by the pattern generator, and a comparator for detecting the image angle of the pixel to be interpolated based on the comparison result; and a pixel of the pixel to be interpolated based on the comparison result A shape detector that detects the shape of an image by a combination of the angle of the image and the angle of the image detected in the upper and lower interpolation scan lines. In the image angle detection device of the present invention, a 2D mode is generated by performing a 2D process on the input image signal in a predetermined detection area by a 2D mode generator. In addition, a reference pattern generator is used to generate a two-dimensional image with a majority direction as the majority reference pattern. Then, the comparator compares the 2D mode with a plurality of reference modes, and based on the comparison result, detects the image angle of the pixel to be interpolated. Furthermore, based on a combination of the image angles of the pixels to be interpolated detected by the comparator and the image angles detected by the interpolation scan lines above and below, the shape detector is used to detect them. Portrait shape. In this case, since the comparison of the two-dimensional mode is performed, compared with the case where a difference chirp between 2 pixels is used, erroneous detection can be suppressed, and the angle of the image with oblique edges can be accurately detected. In addition, by adopting the two-dimensional reference mode, not only can be detected, but also the straight line angle between pixels whose detection angle is located at a point symmetrical with the interpolated pixel as the center, can also be detected between these angles. angle. Therefore, the angle can be detected at more precise intervals. In addition, if you want to detect the curvilinear shape of an image by comparing the 2D mode with most of the reference modes, you must have at least 200307457
3條掃描線之巨大數目的參照模式，亦使電路規模變大， 因此並不實用。於本發明的畫像角度檢出裝置中，首先於 小區域中’各自檢出關於各條內插掃描線的各個應予內插 的像素之局部性的畫像角度，並基於所檢出的角度與其上 下方向的組合來檢出畫像形狀。於此情況下，可經由改良 藉由比較2値化模式與多數的各個參照模式來檢出畫像角 度的構成的一部分，來檢出畫像形狀。 因此’經由畫像形狀的檢出，而不會於畫像角度的檢 出處理中產生延遲’且不會使線路規模變大，來正確的檢 出畫像的角度及畫像的形狀。 於本發明的畫像角度檢出裝置中，於比較器所檢出之 關於應予內插的像素的畫像角度，位於在上方的內插掃描 線中所檢出的畫像角度以及在下方的內插掃描線中所檢出 的畫像角度之間，並且在上方的內插掃描線中所檢出的畫 像角度的絕對値大於在下方的內插掃描線中所檢出的畫像 角度的絕對値的情況下’於關於應予內插的像素所檢出的 畫像角度、以及在上方的內插掃描線中所檢出的畫像角度 、以及在下方的內插掃描線中所檢出的畫像角度均爲正値 之際，則形狀檢出器輸出顯示畫像形狀爲朝右下方凸出的 形狀之形狀檢出信號，而於關於應予內插的像素所檢出的 畫像角度、以及在上方的內插掃描線中所檢出的畫像角度 、以及在下方的內插掃描線中所檢出的畫像角度均爲負値 之際，則形狀檢出器輸出顯示畫像形狀爲朝左下方凸出的 形狀之形狀檢出信號；而於比較器所檢出關於上述應予內 -8- (5) (5)200307457 插的像素的畫像角度，位於在上方的內插掃描線中所檢出 的畫像角度以及在下方的內插掃描線中所檢出的畫像角度 之間，並且在上方的內插掃描線中所檢出的畫像角度的絕 對値小於在下方的內插掃描線中所檢出的畫像角度的絕對 値的情況下’於關於應予內插的像素所檢出的畫像角度、 以及在上方的內插掃描線中所檢出的畫像角度、以及在下 方的內插掃描線中所檢出的畫像角度均爲正値之際，則形 狀檢出器輸出顯示畫像形狀爲朝左上方凸出的形狀之形狀 檢出信號，而於關於應予內插的像素所檢出的畫像角度、 以及在上方的內插掃描線中所檢出的畫像角度、以及在下 方的內插掃描線中所檢出的畫像角度均爲負値之際，則形 狀檢出器輸出顯示畫像形狀爲朝右上方凸出的形狀之形狀 檢出信號。 於此情況下，因應關於所檢出之應予內插的像素的畫 像角度、以及於上方及下方的內插掃描線中所檢出的畫像 角度的組合，可檢出畫像形狀爲朝像哪邊凸出的形狀。 於本發明的畫像角度檢出裝置中，在檢出畫像形狀爲 圓弧狀的情況下，形狀檢出器亦可輸出顯示圓弧的內側方 向之形狀檢出信號。 於此情況下，可基於顯示圓弧內側的形狀檢出信號， 來判別具有圓弧形狀的畫像中之圓弧的方向。 於本發明的畫像角度檢出裝置中，2値化模式產生器 亦可包含：基於檢出區域內的影像信號的亮度，算出用於 2値化的閾値之閾値算出裝置；以及採用以閾値算出裝置 -9 - (6) (6)200307457 所算出的閾値，藉由對所輸入的影像信號進行2値化處理 ’來產生2値化模式之2値化裝置。 於此情況下，因爲可基於檢出區域內的影像信號的亮 度’算出用於2値化的閾値，因此不須從外部設定閾値， 可產生不受限於影像信號的亮度水準之2値化模式。 於本發明的畫像角度檢出裝置中，又具備於檢出區域 內的影像信號中，判定各掃描線的水平方向的亮度分佈爲 單調遞增或是單調遞減之判定器，而在判定器判定亮度分 佈不是單調遞增亦不是單調遞減的情況下，比較器亦可不 進行2値化模式與多數的各個參照模式之比較。 在檢出區域內的影像信號中，各掃描線的水平方向的 亮度分佈不是單調遞增亦不是單調遞減的情況下，並不進 行2値化模式與多數的各個參照模式之比較，且不檢出畫 像角度。藉此，以抑制因雜訊所造成的錯誤檢出。 於本發明的畫像角度檢出裝置中，又具備檢出檢出區 域內的影像信號的對比之對比檢出器，而在對比檢出器所 檢出的對比小於所定値的情況下，比較器亦可不進行2値 化模式與多數的各個參照模式之比較。 在影像信號的對比較低的情況下，採用斜方向的像素 之內插處理的效果較小。因此，檢出區域內的影像信號的 對比小於所定値的情況下，並不進行2値化模式與多數的 各個參照模式之比較，且不檢出畫像角度。藉此，使用伴 隨著雜訊的斜方向的像素之內插處理，可以僅在效果較大 的情況下才採用。 -10- (7) (7)200307457 於本發明的畫像角度檢出裝置中，由參照模式產生器 所產生的多數的各個參照模式’亦可包含配置於應予內插 的像素的上方的掃描線之第1像素列以及配置於應予內插 的像素的下方的掃描線之第2像素列，而第i像素列具有 從第1像素値往第2像素値變化的1個變化點，第2像素 列具有從第1像素値往第2像素値變化的i個變化點，並 且第1像素列中之從第1像素値往第2像素値變化的變化 方向，與第2像素列中之從第1像素値往第2像素値變化 的變化方向相同。 於此參照模式中，配置於上方的掃描線的像素列以及 配置於下方的掃描線的像素列，均具有亮度變化，且具有 相同方向的亮度梯度。如此的參照模式相當於斜邊緣的畫 像。因此，當2値化模式與參照模式一致的情況下，可以 確實的特定化斜邊緣的角度。 依循本發明的其他型態之掃描線內插裝置，其特徵爲 具備：基於所輸入的影像信號，來檢出關於應予內插的像 素的畫像角度及畫像形狀之畫像角度檢出裝置；以及基於 由畫像角度檢出裝置所檢出的角度及形狀，選擇用於內插 處理的像素，並藉由採用所選擇的像素來算出應予內插的 像素之値，來生成內插掃描線之內插電路；其中，畫像角 度檢出裝置包含··於包含多數的掃描線及各內插掃描線中 之應予內插的像素之所定檢出區域內，對所輸入的影像信 號進行2値化處理而產生2値化模式之2値化模式產生器 ;以及產生具有不同方向的2値畫像，以做爲多數的參照 -11 - (8) (8)200307457 模式之參照模式產生器；以及將藉由2値化模式產生器所 產生的2値化模式，與藉由參照模式產生器所產生的多數 的各個參照模式比較，並基於比較結果，來檢出關於應予 內插的像素的畫像角度之比較器；以及基於比較器所檢出 之關於應予內插的像素的畫像角度、及於上方與下方的內 插掃描線中所檢出的畫像角度之組合，來檢出畫像的形狀 之形狀檢出器。 於本發明的掃描線內插裝置中，藉由畫像角度檢出裝 置’基於所輸入的影像信號，來正確的檢出關於應予內插 的像素的畫像角度及畫像形狀，並基於由畫像角度檢出裝 置所檢出的角度及形狀，選擇用於內插處理的像素，並採 用由內插電路所選擇的像素來算出應予內插的像素之値， 來生成內插掃描線。 於此情況下，因爲因應畫像形狀來選擇用於內插的像 素，因此不僅可對直線形狀進行內插，亦可對曲線形狀的 邊緣進行平滑的內插。 因此，可進行適用於影像信號所顯示之畫像的角度及 畫像的形狀之平滑的內插。 於本發明的掃描線內插裝置中，在檢出畫像形狀爲圓 弧狀的情況下，形狀檢出器輸出顯示圓弧的內側方向之形 狀檢出信號；而內插電路基於形狀檢出器所輸出的形狀檢 出信號，從圓弧的內側選擇用於內插處理的像素，並藉由 採用所選擇的像素來算出應予內插的像素之値，來生成內 插掃描線。 -12- 200307457 Ο) 於此情況下，基於顯示圓弧的內側方向的形狀檢出信 號，從圓弧的內側選擇用於內插處理的像素，並採用所選 擇的像素來算出應予內插的像素之値。藉此，可生成沿著 圓弧形狀之平滑的內插。 於本發明的掃描線內插裝置中，在檢出畫像形狀爲圓 弧狀的情況下，形狀檢出器輸出顯示圓弧的內側方向之形 狀檢出信號；而內插電路基於形狀檢出器所輸出的形狀檢 出信號，選擇對關於應予內插的像素所檢出的畫像角度的 方向中之上方與下方的掃描線位置，僅僅往圓弧的內側方 向移動0.5像素量的位置，並藉由採用所選擇的位置的像 素値來算出應予內插的像素之値，來生成內插掃描線。 於此情況下，基於顯示圓弧的內側方向的形狀檢出信 號，從圓弧內側的特定像素當中選擇用於內插處理的像素 ，並採用所選擇的像素來算出應予內插的像素之値。藉此 ，可生成沿著圓弧形狀之平滑的內插。 再者，因爲可因應圓弧形狀來選擇用於內插的像素， 因此不僅可對直線形狀進行內插，亦可對曲線形狀的邊緣 進行平滑的內插。 依循本發明的其他型態之畫像角度檢出方法，乃基於 所輸入的影像信號，來檢出關於掃描線之間的各條內插掃 描線中之應予內插的像素之畫像角度，其特徵爲具備：於 包含多數的掃描線及應予內插的像素之所定檢出區域內， 對所輸入的影像信號進行2値化處理而產生2値化模式之 步驟；以及產生具有不同方向的2値畫像，以做爲多數的 -13- (10) (10)200307457 參照模式之步驟；以及將藉由2値化模式產生器所產生的 2値化模式，與藉由參照模式產生器所產生的多數的各個 參照模式比較，並基於比較結果，來檢出關於應予內插的 像素的畫像角度之步驟；以及基於比較器所檢出之關於應 予內插的像素的畫像角度、及於上方與下方的內插掃描線 中所檢出的畫像角度之組合，來檢出畫像的形狀之步驟。 於本發明的畫像角度檢出方法中，對所輸入的影像信 號於所定的檢出區域內進行2値化處理，來產生產生2値 化模式。此外，產生具有多數方向的2値化畫像，以做爲 多數的參照模式。然後，比較2値化模式與多數的各個參 照模式，並基於比較結果，來檢出關於應予內插的像素的 畫像角度。再者，基於關於應予內插的像素的畫像角度、 及於上方與下方的內插掃描線中所檢出的畫像角度之組合 ，來檢出畫像的形狀。 於此情況下，因爲進行二次元的模式比較，因此與採 用2像素之間的差分値的情況相較，可抑制錯誤的檢出， 並可正確的檢出具有斜方向的邊緣之畫像角度。此外，藉 由採用二次元的參照模式，不僅可檢出，連接檢出角度位 於以內插像素爲中心之點對稱的位置上之像素之間的直線 角度，還可以檢出位於這些角度之間的角度。因此，可於 更精密的間隔檢出角度。 此外，首先於小區域中，各自檢出關於各條內插掃描 線的各個應予內插的像素之局部性的畫像角度，並基於所 檢出的角度與其上下方向的組合來檢出畫像形狀。於此情 -14 - (11) (11)200307457 況下，可經由改良藉由比較2値化模式與多數的各個參照 模式來檢出畫像角度的構成的一部分，來檢出畫像形狀。 因此，經由畫像形狀的檢出，而不會於畫像角度的檢 出處理中產生延遲，且不會使線路規模變大，來正確的檢 出畫像的角度及畫像的形狀。 【實施方式】 第1圖係顯示本發明的實施型態中之晝像角度檢出裝 置的構成之方塊圖。 第1圖的畫像角度檢出裝置10包含線記憶體1 a、以 及線記憶體1 b、以及線記憶體1 c、以及2値化部2、以 及角度檢出部3、以及圓弧形狀檢出部4、以及檢出視窗 內影像信號處理部5、以及參照模式產生部6、以及A/D 轉換器7。 A/D轉換器7對類比的影像信號VA進行類比數位轉 換，並輸出數位影像信號VD 1。從A/D轉換器7輸出之 影像信號VD 1，被輸入於線記憶體1 a及檢出視窗內影像 信號處理部5。線記憶體1 a對A/D轉換器7所輸出的影 像信號VD 1，進行1條線（1條掃描線）的量的延遲而輸 出。從線記憶體1 a所輸出的影像信號V D 2，被賦予至2 値化部2及檢出視窗內影像信號處理部5。 於本例當中，影像信號VD1、VD2具有25 6色調的亮 度。亦即’影像信號 VD1、VD2的亮度的最小値爲，，〇”， 最大値爲”255 ”。 (12) (12)200307457 2値化部2以之後所述之從檢出視窗內影像信號處理 部5所賦予的平均亮度値LU爲閾値，對A/D轉換器7所 輸出的影像信號VD i以及線記憶體i a所輸出的影像信號 VD2進行2値化處理，並輸出由”1”及，，〇，，所組成的2値化 模式BI。2値化模式BI具有檢出視窗的大小。 在此’檢出視窗例如爲，包含影像信號VD1的7個 像素及影像信號V D 2的7個像素之7 X 2像素的矩形區域 、或是包含影像信號VD1的15個像素及影像信號VD 2的 1 5個像素之1 5 X 2像素的矩形區域等。於以下的說明中， 以9 X 2像素爲檢出視窗的大小。於此情況下，2値化模式 BI的大小爲9x 2像素。檢出視窗的大小並不限定於此， 於本發明的範圍內可任意設定。 檢出視窗內影像信號處理部5於所輸出的影像信號 VD 1以及由線記憶體1 a所輸出的影像信號V D 2上，設定 檢出視窗，並算出檢出視窗內的影像信號V D 1、V D 2的亮 度平均値，並賦予平均亮度値LU於2値化部2，以做爲 用於2値化的閾値。 於本實施型態中，乃採用檢出視窗內所有像素的亮度 平均値，來做爲用於2値化的閾値，但是並不限定於此， 亦可以採用檢出視窗內像素値的最大値與最小値的平均値 ’來做爲用於2値化的閾値，或是可採用以亮度大小來排 列之際的中央値，來做爲用於2値化的閾値，或是可採用 位於以亮度大小來排列之際的中央値附近的多數像素的平 均値等，來做爲用於2値化的閾値。 -16- (13) (13)200307457 此外’檢出視窗內影像信號處理部5判定檢出視窗內 的影像信號VD1、VD2的水平方向的亮度分佈爲單調遞增 或是單調遞減，若不是單調遞增亦不是單調遞減的情況下 ’則亦可以賦予最小値爲” 0 ”或是最大値爲，，2 5 5，，於2値 化部2以做爲閾値。藉此，2値化部2輸出所有由，，1，，或 是”〇”所組成的2値化模式BI。於此情況下，則依序算出 與影像信號VD1、VD2相鄰的2個像素之間的差分値，若 差分値的正負符號相同的話，則可判定爲單調遞增或是單 調遞減。 此外，檢出視窗內影像信號處理部5檢出視窗內影像 信號處理部5算出檢出視窗內的影像信號VD1、VD2的亮 度的最大値與最小値的差來做爲對比，於所算出的對比小 於所定値的情況下，則賦予最小値爲”0”或是最大値爲 ”25 5 ”於2値化部2以做爲閾値。藉此，2値化部2輸出 所有由”1”或是”0”所組成的2値化模式BI。 參照模式產生部6產生由”1”及”0”所組成的多數的參 照模式RA，並輸出至角度檢出部3。各參照模式RA的大 小與檢出視窗的大小相同。 角度檢出部3將由2値化部2所賦予的2値化模式 BI，與參照模式產生部6所賦予的多數的各個參照模式 RA進行比較，並輸出一致的參照模式RA的角度以做爲 角度資訊S 1。關於此角度之後將會詳述。以下，將2値 化模式BI與各個參照模式R A的比較動作稱爲模式比對 (Pattern Matching)。 (14) 200307457 如上所述，在檢出視窗內的影像信號 VD 1及影像信 號 VD2的亮度分佈均不是單調遞增亦不是單調遞減的情 況下，可從2値化部2輸出所有由”1”或是”〇”所組成的2 値化模式BI。於此情況下，並不從角度檢出部3輸出角 度資訊S 1。 此外，檢出視窗內的影像信號V D 1、V D 2的對比小於 所定値的情況下，則從2値化部2輸出所有由” 1 ”或是”0” 所組成的2値化模式BI，因此，並不從角度檢出部3輸 出角度資訊S 1。 在影像信號VD1、VD2的對比較小的情況下，採用斜 方向的像素之內插處理的效果較小。於採用斜方向的像素 之內插處理當中，若未檢出正確的角度的話，則可能產生 雜訊，因此，在效果較小的情況下，並不進行採用斜方向 的像素之內插處理且不輸出角度資訊S 1。 線記億體1 b對角度檢出部3所輸出的角度資訊S 1， 進行1條線（1條掃描線）的量的延遲，並輸出延遲1條 線的量的角度資訊S 2於圓弧形狀檢出部4及線記憶體1 c 。線記憶體lc對線記憶體lb所輸出的角度資訊S2，進 行1條線（1條掃描線）的量的延遲，並輸出延遲1條線 的量的角度資訊S3於圓弧形狀檢出部4。 在此，將包含內插像素的掃描線稱爲內插掃描線。圓 弧形狀檢出部4不僅從上方1條線的內插掃描線的角度資 訊S 3、及下方1條線的內插掃描線的角度資訊S 1、及爲 對象的內插掃描線的角度資訊S 2的組合當中，輸出畫像 -18 - (15) (15)200307457 邊緣角度資訊τ 1於爲對象的內插掃描線上’還辨識圓弧 形狀，並輸出圓弧形狀資訊τ 2於爲對象的內插掃描線上 。關於角度檢出及圓弧形狀的辨識，將於之後詳述。 第2圖係顯示由第1圖的2値化部2所輸出之2値化 模式ΒΙ的一例之模式圖。 於第2圖中，IN顯示內插像素，IL顯示內插掃描線 。此外，AL顯示內插掃描線IL的上方掃描線，BL顯示 內插掃描線I L的下方掃描線。 於第2圖的例子中，以”0”顯示較低的部分（較暗的 部分），以”1”顯示較高的部分（較亮的部分）。於2値 化模式BI中，畫像邊緣的角度爲45 °。在此，水平方向 的角度爲0°，右上的斜方向的角度爲正。 第3圖、第4圖、第5圖、第6圖係顯示由第1圖的 參照模式產生部6所產生之參照模式的例子之模式圖。反 白的像素爲，用於算出以粗線所顯示的內插像素的値之上 方及下方掃描線的像素。 弟 3 圖的（a) 、 ( b ) 、 ( c ) 、 ( d ) 、 （e) 、 （f )各自顯不 45°、34。、27。、22。、18。、16。之參照 模式。於第3圖的例子中，左上方顯示較暗的部分，右下 方顯示較亮的部分。第4圖的（a ) 、（ b ) 、 （ c ) 、 （ d )、（e) 、 （f)各自顯示 45。、34。、27。、22。、 18。、1 ό。之參照模式。於第4圖的例子中，左上方顯示 較亮的部分，右下方顯示較暗的部分。 第 5 圖的（a) 、 （b) 、 （c) 、 （d) 、 （e) 、 （f •19- (16) (16)200307457 )各自顯示-45°、-34。、-27。、-22。、-18。、- 16。之 參照模式。於第5圖的例子中，右上方顯示較暗的部分， 左下方顯示較亮的部分。第6圖的（a ) 、（ b ) 、（ c ) 、（d) 、 （e) 、 （f)各自顯示- 45°、-34°、-27。、 -22 °、-18 °、- 16 °之參照模式。於第6圖的例子中，右 上方顯示較亮的部分，左下方顯示較暗的部分。 於角度檢出部3中，第3圖〜第6圖所示之參照模式 與2値化部2所輸出之2値化模式BI比較，然後角度檢 出部3輸出一致的參照模式所具有的角度資訊S工。 此外，如第3圖〜第6圖所示般，於依據二次元的亮 度分佈之參照模式中，不僅可以設定連接以內插像素爲中 心之點對稱的位置之像素間的直線的角度，還可以設定位 於兩者之間的角度。例如，可以設定位於45 °與 2 7。之 間以及45°與18°之間的角度之34°及22°。 例如，第2圖的2値化模式BI可與第4圖（a )的6 個參照模式當中的1個參照模式一致。於此情況下，第1 圖的角度檢出部3輸出第4圖（a )的參照模式所顯示之 45°，做爲角度資訊S1。 而由第1圖的參照模式產生部6所產生的參照模式 RA，並不限定於第3圖〜第6圖所示之例子，可採用任何 大小的參照模式。 第7圖係顯示用於說明第1圖的圓弧形狀檢出部4的 處理之圖式。第7圖係顯示，因應內插像素的檢出角度、 及下方的內插掃描線中之檢出角度、及上方的內插掃描線 •20- (17) (17)200307457 中之檢出角度的組合，所辨識之圓弧形狀的例子。 具體而言，圓弧形狀可分類爲第7圖的5種狀況A、 及B、及C、及D、及E之組合。 於第7圖的狀況A的組合例當中，顯示圓弧形狀的 邊緣（圓弧邊緣）的凸狀方向爲”右下”’圓弧形狀內側所 朝的方向爲”左”。於此情況下，內插像素的檢出角度的絕 對値，爲下方的內插掃描線中之檢出角度的絕對値與上方 的內插掃描線中之檢出角度的絕對値的中間値，並且上方 的內插掃描線中之檢出角度的絕對値大於下方的內插掃描 線中之檢出角度的絕對値，所有的檢出角度爲正値。 圓弧形狀檢出部4不僅輸出內插像素的檢出角度，來 做爲角度資訊T1，還於辨識圓弧形狀的情況下，輸出圓 弧形狀內側所朝的方向，來做爲圓弧形狀資訊T2。 此外，於第7圖的最右邊，顯示圓弧形狀的模式圖及 顯示圓弧內側的例子。多數的細長箭頭的連接，顯示所辨 識出的圓弧，細長的各個箭頭的方向則顯示沿著圓弧所檢 出的角度，而粗箭頭顯示圓弧的內側。 於第7圖的狀況B、C、D的各個組合當中，與第7 圖的狀況A相同，均顯示圓弧邊緣的凸狀方向及圓弧形 狀內側所朝的方向。 於第7圖的狀況B的組合例當中，顯示圓弧邊緣的凸 狀方向爲”右上”，圓弧形狀內側·所朝的方向爲”左”。於此 情況下，內插像素的檢出角度的絕對値，爲下方的內插掃 描線中之檢出角度的絕對値與上方的內插掃描線中之檢出 -21 - (18) (18)200307457 角度的絕對値的中間値，並且上方的內插掃描線中之檢出 角度的絕對値小於下方的內插掃描線中之檢出角度的絕對 値，所有的檢出角度爲負値。 於第7圖的狀況C的組合例當中，顯示圓弧邊緣的凸 狀方向爲”左上”，圓弧形狀內側所朝的方向爲”右”。於此 情況下，內插像素的檢出角度的絕對値，爲下方的內插掃 描線中之檢出角度的絕對値與上方的內插掃描線中之檢出 角度的絕對値的中間値，並且上方的內插掃描線中之檢出 角度的絕對値小於下方的內插掃描線中之檢出角度的絕對 値，所有的檢出角度爲正値。 於第7圖的狀況D的組合例當中，顯示圓弧邊緣的 凸狀方向爲”左下”，圓弧形狀內側所朝的方向爲”右”。於 此情況下，內插像素的檢出角度的絕對値，爲下方的內插 掃描線中之檢出角度的絕對値與上方的內插掃描線中之檢 出角度的絕對値的中間値，並且上方的內插掃描線中之檢 出角度的絕對値小於下方的內插掃描線中之檢出角度的絕 對値，所有的檢出角度爲負値。 於第7圖的狀況E的組合例當中，顯示無法辨識圓弧 邊緣之檢出角度的組合。亦即，所有不屬於於第7圖的狀 況A、B、C、D的組合例之組合，均屬於第7圖的狀況E. 〇 在此，所謂的中間値，是指處於2個數値X、Y之間 的値，若X < Y的話，則所有大於X小於Y的値均爲中間 値。 -22- (19) (19)200307457 於上方的內插掃描線中，上方的內插掃描線中之檢出 角度的參照位置乃對應於，位於對內插像素以該內插像素 的檢出角度所決定的方向上的1點，而於下方的內插掃描 線中，下方的內插掃描線中之檢出角度的參照位置乃對應 於，位於對內插像素以該內插像素的檢出角度來決定的方 向上的1點。 於上方的內插掃描線中，上方的內插掃描線中之檢出 角度的參照位置亦可以採用，包含位於對內插像素以該內 插像素的檢出角度來決定的方向上的1點之水平方向上的 一定寬度的區域（多數像素量的區域）。此外，於下方的 內插掃描線中，下方的內插掃描線中之檢出角度的參照位 置亦可以採用，包含位於對內插像素以該內插像素的檢出 角度來決定的方向上的1點之水平方向上的一定寬度的區 域（多數像素量的區域）。 關於第7圖中之內插像素的檢出角度、及下方的內插 掃描線中之檢出角度、及上方的內插掃描線中之檢出角度 的組合，僅僅舉出上述各個例子，但並不限定於此，亦可 採用圖中未顯示之組合。 於本實施型態的畫像角度檢出裝置1 〇中，將檢出視 窗內的影像信號VD1、VD2的亮度分佈轉換爲2値化模式 BI，並進行2値化模式BI與預先設定的多數的參照模式 RA進行模式比對，藉此，可以以極小的電路規模來檢出 畫像的斜邊緣角度。 於此情況下，因爲採用檢出視窗內的平均亮度値做爲 -23- (20) (20)200307457 2値化的閾値，因此不須從外部設定2値化的閾値’可產 生不受限於影像信號的亮度水準之必定包含”0”及”1”兩者 的2値化模式BI。 此外，因爲進行依據二次元的亮度分佈之模式比對’ 因此與採用2像素之間的差分値的情況相較’可抑制錯誤 的檢出，並可正確的檢出具有斜方向的邊緣之畫像角度。 再者，藉由採用依據二次元的亮度分佈的參照模式 RA，不僅可檢出，連接檢出角度位於以內插像素爲中心 之點對稱的位置上之像素之間的直線角度，還可以檢出位 於這些角度之間的角度。因此，可採用較少容量的線記憶 體1 a，以更精密的間隔來檢出角度。 此外，因爲可由內插像素的檢出角度、及下方的內插 掃描線中之檢出角度、及上方的內插掃描線中之檢出角度 的組合，來辨識圓弧形狀，因此不需準備必須有3條以上 的掃描線之用於辨識圓弧形狀的參照模式。因此，不需增 加畫像角度檢出裝置1 〇的電路規模或是計算規模，來檢 出畫像角度並辨識出圓弧形狀。 第8圖係顯示具備第i圖的畫像角度檢出裝置之掃描 線內插裝置的構成之方塊圖。 於第8圖中’掃描線內插裝置j 〇 〇由畫像角度檢出裝 置以及內插電路20所構成。影像信號VA輸入於畫像 角度檢出裝置10以及內插電路20。 畫像角度檢出裝置10由第1圖的畫像角度檢出裝置 1〇所構成。畫像角度檢出裝置基於影像信號VA，檢 -24 - (21) (21)200307457 出畫像的斜邊緣角度及圓弧形狀，並輸出畫像邊緣角度資 訊T1與圓弧形狀資訊T2。內插電路20基於角度資訊T1 與圓弧形狀資訊T2，從上下方的掃描線來選擇內插像素 的斜方向的像素，然後採用選擇的像素的亮度値算出內插 像素的亮度値，而輸出內插影像信號V Ο U T。 於第8圖的掃描線內插裝置1 0 0中，不僅可藉由畫像 角度檢出裝置1 〇正確的檢出直線形狀，亦可正確的檢出 具有包含圓弧形狀的斜方向邊緣之畫像角度。因此，即使 在不僅包含直線形狀，亦包含具有圓弧形狀的斜方向邊緣 之畫像當中，可選擇斜方向的適當像素，來進行平滑的內 插處理。 第9圖係顯示第8圖的掃描線內插裝置1 00中之內插 電路20的構成之方塊圖。 第9圖的內插電路20包含A/D (類比·數位）轉換 器21、及線記憶體22、及內插像素選擇電路23、及平均 値運算電路24。 A/D轉換器21對類比的影像信號VA進行類比數位 轉換，並輸出數位影像信號VD1。從A/D轉換器21輸出 之影像信號VD 1，被輸入於線記憶體22及內插像素選擇 電路23。線記憶體22對A/D轉換器21所輸出的影像信 號VD 1，進行1條線（1條掃描線）的量的延遲而輸出。 從線記憶體22所輸出的影像信號VD2，被賦予至內插像 素選擇電路23。 內插像素選擇電路23採用所賦予的影像信號VD 1、 (22) (22)200307457 V D 2、及畫像角度檢出裝置1 〇的角度資訊τ 1與圓弧形狀 資訊Τ2，從上方掃描線當中，選擇內插參照像素ρ 1並輸 出於平均値運算電路24，從下方掃描線當中，選擇內插 參照像素Ρ 2並輸出於平均値運算電路2 4。 平均値運算電路24從內插參照像素ρ 1及內插參照像 素Ρ2，算出內插像素的亮度値並輸出。 依據內插像素選擇電路23之內插參照像素的選擇， 乃於內插之際使畫像中之邊緣變得平滑的方式來進行。依 據內插像素選擇電路23選擇性的進行，在畫像具有直線 形狀的邊緣的情況的動作，以及畫像具有圓弧形狀的邊緣 的情況的動作。亦即，內插像素選擇電路23在畫像具有 直線形狀的邊緣的情況下，基於角度資訊Τ1，從上下方 的掃描線當中’選擇第3圖〜第6圖所示之反白部分的像 素來做爲內插參照像素。而內插參照像素的中心位置則爲 內插參照位置。此外，內插像素選擇電路23在畫像具有 圓弧形狀的邊緣的情況下’於上下方的掃描線中，對內插 像素來特定化角度資訊Τ 1所示方向的位置，並基於圓弧 形狀資訊Τ2，對於該特定化的位置，選擇於圓弧形狀的 內側當中往水平方向移動之位置，來做爲內插參照位置。 在此採用第7圖，以一個例子來說明。在檢出狀況Α 的圓弧形狀的情況下’因爲圓弧形狀的內側方向在粗箭頭 所示般之左側，因此內插像素選擇電路23在上下方的掃 描線當中’對內插像素來特定化角度資訊T 1所示方向的 位置，並對於該特定化的位置，從移往左側的位置當中， -26· (23) (23)200307457 選擇上下方各自的內插參照位置。 此外，於畫像具有圓弧形狀之際，在選擇將內插參照 位置’對畫像具有直線形狀邊緣的情況之際所選擇的內插 爹照位置’移動0.5像素來做爲內插參照位置的情況下， 乃經由實驗，確認了此方式乃可藉由內插來形成最爲平滑 的圓弧形狀的方式。 平均値運算電路24運算內插參照位置所屬像素的平 均亮度値，並決定內插像素的亮度値。雖然圖中未顯示， 平均値運算電路24亦可包含相關運算電路，其可進行因 應內插參照位置所屬像素之間的相關性的大小，亦即因應 內插參照位置所屬像素的差分大小之內插運算。於此情況 下’即使因某種原因而錯誤檢出角度及圓弧形狀，亦可降 低因內插所造成的雜訊。 第1 〇圖係顯示由第1圖的畫像角度檢出裝置1 〇所檢 出的畫像角度的一例之模式圖。於第1 〇圖中，顯示具有 圓弧形狀邊緣的畫像的例子。第11圖係顯示採用由第1 圖的晝像角度檢出裝置1 0所檢出的畫像的角度資訊、以 及圓弧形狀資訊之像素的內插的例子之圖式。相對的，第 12圖係顯示僅採用由第1圖的畫像角度檢出裝置1〇所檢 出的畫像的角度資訊之像素的內插的例子之圖式。 於第10圖〜第12圖中，IL1、IL2、及IL3顯示內插 掃描線，AL、BL、CL、及DL顯示內插掃描線。於第10 圖〜第12圖中，AL、BL、CL、及DL上的各個値顯示各 個像素的亮度値。於第1 〇圖中，內插掃描線IL1、IL2、 (24) 200307457 中 個 爲 線 〇 照 象 中 爲 的 內 而 素 處 的 爲 R2 的 上 及IL3上的各個値顯示角度資訊，於第11圖及第丨2圖 ，內插掃描線IL1、IL2、及IL3上的各個値，顯示各 內插像素的亮度値。 在此，以IN爲對象內插像素。於此情況下，：[L2 對象內插掃描線’ BL爲上方掃描線，CL爲下方掃描 ’ IL1爲上方的內插掃描線，IL3爲下方的內插掃描線 此外，於第1 1圖及第12圖中，以X記號來顯示內插參 位置P1、P 2、P 3、P 4。於第1 〇圖的例子當中，關於對 內插像素IN的畫像之角度爲27。。 於弟12圖的例子當中，於上下方掃描線b L、C L ，選擇對對象內插像素IN的2 7。的方向之位置，來做 內插參照位置P3、P4。因爲內插參照位置p3、P4所屬 像素Q 3、Q 4的平均亮度値各爲” χ 〇 〇，，，因此藉由採用 插參照位置P 3、P 4所屬的像素Q 3、Q 4的內插運算， 使對象內插像素IN的亮度値爲” χ 〇〇，，。關於其他內插像 的亮度値亦以相同方法來算出，因此可得到第12圖的 理結果。 於第1 1圖的例子當中，以上方的內插掃描線IL1 參照像素爲R 1 ’以下方的內插掃描線IL2的參照像素 R2。因爲參照像素R1的角度資訊爲45 °，參照像素 的角度資訊爲18 °，因此第1 1圖的例子相當於第7圖 狀況A，其圓弧形狀內側所朝的方向爲，，左”。因此，於 下方掃描線B L、C L中’特定化對對象內插像素I n的 2 7的方向之位置p 1、p 2，並選擇對特定化的位置p 1 -28- (25) (25)200307457 p 2往水平方向移動〇 . 5像素的位置，來做爲內插參照位置 PI、P2。內插參照位置P1 ' P2所屬的像素的亮度値，各 爲相鄰的2個像素Q 1 ;l、q i 2的亮度値，，1 〇 〇，，及，，〇，，的平均 亮度値” 5 0 ” °關於其他內插像素的亮度値亦以相同方法來 算出，因此可得到第1 1圖的處理結果。 於第11圖及第12圖中，內插掃描線IL1、IL2、及 IL3上的各個値，顯示各個內插像素的亮度値，黑點爲於 各條掃描線的亮度的中央値，虛線顯示以直線連接各個亮 度中央値的點者。亦即，虛線表示畫像的邊緣。於本例的 情況下，中央値爲亮度値”50”者。 若比較第1 1圖的處理結果及第1 2圖的處理結果的話 ’則在僅僅採用第1 2圖所示之畫像的角度資訊來算出內 插像素的亮度値的情況下，進行內插後的畫像邊緣形狀爲 折線形狀，無法以內插得到平滑的邊緣。亦即，因爲第 12圖的處理是從局部所辨識的角度方向來進行內插，因 此從大範圍來觀看其連續性的話，則並不見得可達到平滑 的內插。 相對於此，在基於第11圖所示之畫像的角度資訊及 圓弧形狀資訊來算出內插像素的亮度値的情況下，則可以 表現出近似於圓弧形狀的平滑邊緣。 於本例中，乃以辨識爲圓弧形狀的情況下的內插參照 位置的移動量爲0.5爲例子來說明，但並不限定於此，亦 可設定任意的移動量。 於本實施型態中，2値化部2及檢出視窗內影像信號 -29- (26) (26)200307457 處理部5相當於2値化模式產生器，參照模式產生部6相 當於參照模式產生器，角度檢出部3相當於比較器。此外 ，圓弧形狀檢出部4相當於形狀檢出器。再者，檢出視窗 內影像信號處理部5相當於閾値算出裝置及判別器，2値 化部2相當於2値化裝置。 【圖式簡單說明】 第1圖係顯示本發明的實施型態中之畫像角度檢出裝 置的構成之方塊圖。 第2圖係顯示由同裝置的2値化部所輸出之2値化模 式的一例。 桌3圖係顯不由第1圖的參照模式產生部所產生之參 照模式的例子之模式圖。 第4圖係顯不由第1圖的參照模式產生部所產生之參 照模式的例子之模式圖。 第5圖係顯示由第1圖的參照模式產生部所產生之參 照模式的例子之模式圖。 第6圖係顯示由第1圖的參照模式產生部所產生之參 照模式的例子之模式圖。 第7圖係顯示用於說明第1圖的圓弧形狀檢出部的處 理之模式圖。 第8圖係顯示具備第1圖的畫像角度檢出裝置之掃描 線內插裝置的構成之方塊圖。 第9圖係顯示第8圖的掃描線內插裝置中之內插電路 -30- (27) (27)200307457 的構成之方塊圖。 第10圖係顯示由第1圖的畫像角度檢出裝置所檢出 的畫像角度的一例之模式圖。 第11圖係顯示採用由第1圖的畫像角度檢出裝置所 檢出的畫像的角度資訊、以及圓弧形狀資訊之像素的內插 的例子之模式圖。 第1 2圖係顯示僅採用由第1圖的畫像角度檢出裝置 所檢出的畫像的角度資訊之像素的內插的例子之模式圖。 〔主要元件對照表〕 la ， lb ， lc ， 22線記憶體 2 2値化部 3角度檢出部 4圓弧形狀檢出部 5檢出視窗內影像信號處理部 6參照模式產生部 7，21 A/D轉換器 1〇畫像角度檢出裝置 20內插電路 23內插像素選擇電路 24平均値運算電路 1〇〇掃描線內插裝置 AL，BL，CL，DL內插掃描線的上方掃描線 ΒΙ 2値化模式 -31 - (28) (28)200307457 BL內插掃描線的下方掃描線 IL，IL1，IL2，IL3 內插掃描線 I N內插像素 LU平均亮度値 p 1，P 2位置 P1，P 2，P 3，P 4 內插參照位置 Q3，Q4，Qll，Q12 像素The large number of reference patterns of 3 scan lines also increases the circuit scale and is therefore not practical. In the image angle detection device of the present invention, first, a local image angle of each pixel to be interpolated for each interpolation scan line is detected in a small area, and based on the detected angle and The combination of the up and down directions detects the image shape. In this case, the shape of the image can be detected by improving a part of the configuration that detects the angle of the image by comparing the 2D mode with a plurality of reference modes. Therefore, 'the angle of the image and the shape of the image can be accurately detected by' detecting the shape of the image without delay in the detection process of the image angle 'and without increasing the scale of the line. In the image angle detection device of the present invention, the image angles of the pixels to be interpolated detected by the comparator, the image angles detected in the interpolation scan lines located above, and the interpolations below When the absolute angle of the image angle between the image angles detected in the scan line and the upper interpolation scan line is greater than the absolute angle of the image angle detected in the lower interpolation scan line The image angles detected by the pixels below which should be interpolated, the image angles detected by the upper interpolation scan line, and the image angles detected by the lower interpolation scan line are When it is right, the shape detector outputs a shape detection signal that shows the shape of the image as a shape protruding toward the lower right, and the angle of the image detected by the pixel to be interpolated and the interpolation above When the angle of the image detected in the scan line and the angle of the image detected in the interpolated scan line below are both negative, the shape detector output shows that the shape of the image is a shape protruding toward the lower left. Shape inspection Signal; and the angle of the image detected by the comparator about the above-mentioned interpolation should be -8- (5) (5) 200307457, the angle of the image detected by the interpolated scan line above and below The absolute angle between the image angles detected in the interpolated scan line and the image angle detected in the upper interpolated scan line is smaller than the absolute value of the image angle detected in the lower interpolated scan line. In the case of ', the angle of the image detected by the pixel to be interpolated, the angle of the image detected by the upper interpolation scan line, and the image detected by the lower interpolation scan line When the angles are all positive, the shape detector outputs a shape detection signal indicating that the shape of the image is a shape protruding toward the upper left, and the angle of the image detected by the pixel to be interpolated and above When the image angle detected in the interpolated scan line and the image angle detected in the lower interpolated scan line are both negative, the shape detector output shows that the image shape is convex toward the upper right Shape checkout letter number. In this case, depending on the combination of the image angles of the detected pixels to be interpolated and the image angles detected in the interpolation scan lines above and below, it can be detected where the image shape is. Convex shape. In the image angle detection device of the present invention, when the shape of the detected image is an arc shape, the shape detector may also output a shape detection signal showing the inner direction of the arc. In this case, the direction of the arc in the image having the shape of an arc can be determined based on the shape detection signal showing the inside of the arc. In the image angle detection device of the present invention, the 2D mode generator may further include: a threshold value calculation device that calculates a threshold value for 2D based on the brightness of the image signal in the detection area; and adopts a threshold value calculation Device-9-(6) (6) 200307457 The threshold value calculated is used to generate a 2D device in a 2D mode by performing 2D processing on the input image signal. In this case, since the threshold value for 2 digitization can be calculated based on the brightness of the image signal in the detection area, there is no need to set the threshold value from the outside, and it is possible to generate 2 digitization without limitation on the brightness level of the image signal mode. The image angle detection device of the present invention is further provided with a determiner for determining the brightness distribution of each scanning line in the horizontal direction in the image signal in the detection area as a monotonically increasing or monotonically decreasing, and determining the brightness in the determiner. In the case where the distribution is not monotonically increasing or monotonically decreasing, the comparator may not compare the 2 値 mode with most of the reference modes. In the case where the luminance distribution in the horizontal direction of each scanning line is not monotonically increasing or monotonically decreasing in the image signal in the detection area, the comparison between the 2 値 mode and most of the reference modes is not performed, and no detection is performed. Portrait angle. This is to suppress false detection caused by noise. The image angle detection device of the present invention further includes a contrast detector that detects the contrast of the image signals in the detection area. When the contrast detected by the contrast detector is smaller than the predetermined threshold, the comparator It is not necessary to compare the 2D mode with most of the reference modes. When the contrast of the video signal is low, the effect of interpolation using pixels in the oblique direction is small. Therefore, when the contrast of the image signals in the detection area is smaller than the predetermined frame, the comparison between the 2 frame mode and most of the reference modes is not performed, and the image angle is not detected. This makes it possible to use the interpolation of pixels with oblique directions accompanied by noise, which can be used only when the effect is large. -10- (7) (7) 200307457 In the image angle detection device of the present invention, the majority of each reference pattern generated by the reference pattern generator may include a scan arranged above a pixel to be interpolated. The first pixel column of the line and the second pixel column of the scan line arranged below the pixel to be interpolated, and the i-th pixel column has a change point that changes from the first pixel 値 to the second pixel ，. The 2 pixel column has i change points that change from the first pixel 値 to the second pixel ，, and the direction of the change from the first pixel 値 to the second pixel 中 in the first pixel column is the same as that in the second pixel column. The direction of change from the first pixel 値 to the second pixel 相同 is the same. In this reference mode, the pixel columns arranged on the upper scanning line and the pixel columns arranged on the lower scanning line both have a brightness change and a brightness gradient in the same direction. Such a reference pattern corresponds to a picture with a beveled edge. Therefore, when the 2D mode is consistent with the reference mode, the angle of the oblique edge can be specified with certainty. A scanning line interpolation device according to another aspect of the present invention is characterized by: an image angle detection device that detects an image angle and an image shape of a pixel to be interpolated based on an input image signal; and Based on the angle and shape detected by the image angle detection device, a pixel for interpolation processing is selected, and the selected pixel is used to calculate the number of pixels to be interpolated to generate an interpolation scan line. Interpolation circuit; among them, the image angle detection device includes: 2 値 the input image signal in a predetermined detection area including a plurality of scanning lines and pixels to be interpolated in each of the interpolation scanning lines; Transformation process to generate a 2D pattern generator of a 2D pattern; and a 2D portrait with different orientations as a reference for most references-11-(8) (8) 200307457 pattern reference pattern generator; and The 2D pattern generated by the 2D pattern generator is compared with the majority of each reference pattern generated by the reference pattern generator, and based on the comparison result, it is detected that interpolation should be performed. A pixel image angle comparator; and a detection based on a combination of the image angle of the pixel to be interpolated detected by the comparator and the image angle detected in the upper and lower interpolation scan lines Shape detector of the shape of the image. In the scanning line interpolation device of the present invention, the image angle detection device 'accurately detects the image angle and image shape of the pixels to be interpolated based on the input image signal, and based on the image angle The angle and shape detected by the detection device select pixels for interpolation processing, and use the pixels selected by the interpolation circuit to calculate the number of pixels to be interpolated to generate an interpolation scan line. In this case, since the pixels to be used for interpolation are selected in accordance with the shape of the image, not only linear shapes but also curved edges can be smoothly interpolated. Therefore, it is possible to perform smooth interpolation suitable for the angle of the image and the shape of the image displayed on the video signal. In the scanning line interpolation device of the present invention, when the shape of the detected image is an arc shape, the shape detector outputs a shape detection signal showing the inner direction of the arc; and the interpolation circuit is based on the shape detector The output shape detection signal selects pixels for interpolation processing from the inside of the arc, and uses the selected pixels to calculate the number of pixels to be interpolated to generate an interpolation scan line. -12- 200307457 〇) In this case, based on the shape detection signal showing the inside direction of the arc, the pixels for interpolation processing are selected from the inside of the arc, and the selected pixels are used to calculate the interpolation to be performed. Of pixels. Thereby, smooth interpolation can be generated along the arc shape. In the scanning line interpolation device of the present invention, when the shape of the detected image is an arc shape, the shape detector outputs a shape detection signal showing the inner direction of the arc; and the interpolation circuit is based on the shape detector The output shape detection signal selects positions above and below the scanning line in the direction of the image angle detected by the pixel to be interpolated, and moves only a position of 0.5 pixels toward the inside of the arc, and The interpolated scan line is generated by calculating the number of pixels to be interpolated by using the pixel φ of the selected position. In this case, based on the shape detection signal showing the inside direction of the arc, a pixel for interpolation processing is selected from specific pixels inside the arc, and the selected pixel is used to calculate the number of pixels to be interpolated. value. With this, smooth interpolation can be generated along the arc shape. Furthermore, since the pixels to be used for the interpolation can be selected in accordance with the arc shape, it is possible to perform smooth interpolation not only on the linear shape but also on the edges of the curved shape. According to other types of image angle detection methods of the present invention, based on the input image signal, the image angles of pixels to be interpolated in each interpolated scan line between scan lines are detected. The method is characterized in that: in a predetermined detection area including a plurality of scanning lines and pixels to be interpolated, a step of generating a 2D mode by performing 2D processing on an input image signal; and generating a 2D mode with different directions; 2 値 portrait, as a majority of the steps of -13- (10) (10) 200307457 reference pattern; and the 2 値 pattern generated by the 2 値 pattern generator and the reference pattern generator A step of comparing the majority of each of the generated reference patterns, and detecting the image angle of the pixel to be interpolated based on the comparison result; and the image angle of the pixel to be interpolated based on the comparison, and The step of detecting the shape of the image by combining the image angles detected in the upper and lower interpolation scan lines. In the image angle detection method of the present invention, the input image signal is subjected to a 2D processing in a predetermined detection area to generate a 2D mode. In addition, a two-dimensional image with a majority direction is generated as a majority reference mode. Then, the two-dimensionalization mode is compared with a plurality of reference modes, and based on the comparison result, the image angles of the pixels to be interpolated are detected. Furthermore, the shape of the image is detected based on a combination of the image angles of the pixels to be interpolated and the image angles detected in the interpolation scan lines above and below. In this case, since the comparison of the two-dimensional mode is performed, compared with the case where a difference chirp between 2 pixels is used, erroneous detection can be suppressed, and the angle of the image with oblique edges can be accurately detected. In addition, by adopting the two-dimensional reference mode, not only can be detected, but also the straight line angle between pixels whose detection angle is located at a point symmetrical with the interpolated pixel as the center, can also be detected between these angles. angle. Therefore, the angle can be detected at more precise intervals. In addition, first, in a small area, a local image angle of each pixel to be interpolated for each interpolation scan line is detected, and the image shape is detected based on the combination of the detected angle and its up-down direction. . In this case, -14-(11) (11) 200307457, the shape of the image can be detected by improving a part of the configuration that detects the angle of the image by comparing the 2D mode with most of the reference modes. Therefore, by detecting the shape of the image, the angle of the image and the shape of the image can be accurately detected without delay in the detection process of the image angle and without increasing the scale of the line. [Embodiment] Fig. 1 is a block diagram showing a configuration of a day image angle detecting device in an embodiment of the present invention. The image angle detection device 10 shown in FIG. 1 includes a line memory 1 a, a line memory 1 b, and a line memory 1 c, and a 2 unit 2, an angle detection unit 3, and an arc shape detection unit. The output unit 4, the video signal processing unit 5 in the detection window, the reference pattern generating unit 6, and the A / D converter 7. The A / D converter 7 performs an analog-to-digital conversion on the analog video signal VA, and outputs a digital video signal VD1. The video signal VD 1 output from the A / D converter 7 is input to the line memory 1 a and the video signal processing unit 5 in the detection window. The line memory 1 a delays the image signal VD 1 output from the A / D converter 7 by one line (one scanning line) and outputs the same. The video signal V D 2 output from the line memory 1 a is given to the video processing unit 2 and the video signal processing unit 5 in the detection window. In this example, the video signals VD1, VD2 have a brightness of 256 tones. That is, the minimum value of the brightness of the image signals VD1 and VD2 is, 0, and the maximum value is "255". (12) (12) 200307457 2 The imager 2 will detect the image signal in the detection window as described later. The average luminance “LU” provided by the processing unit 5 is a threshold value, and the image signal VD i output from the A / D converter 7 and the image signal VD2 output from the line memory ia are subjected to a two-dimensional processing, and the output is “1”. And, 〇 ,, constitutes the 2D mode BI. The 2D mode BI has the size of the detection window. Here, the 'detection window' includes, for example, 7 pixels of the video signal VD1 and the video signal VD2. A rectangular area of 7 pixels by 7 pixels, or a rectangular area of 15 pixels by 15 pixels of the video signal VD1 and 15 pixels by 15 X 2 pixels of the video signal VD 2. In the following description, 9 X 2 pixels is the size of the detection window. In this case, the size of the BI mode is 9x 2 pixels. The size of the detection window is not limited to this, and can be arbitrarily set within the scope of the present invention. The video signal processing unit 5 in the out-of-window The detection window is set on the image signal VD 2 output from the body 1 a, and the brightness average 値 of the image signals VD 1 and VD 2 in the detection window is calculated, and the average brightness 値 LU is given to the 2 unit 2 to It is used as the threshold 値 for 2 値. In this embodiment, the brightness average 値 of all pixels in the detection window is used as the threshold 用于 for 2 値, but it is not limited to this, and it may be The average 値 of the maximum and minimum 値 of the pixel 内 in the detection window is used as the threshold 用于 for 2 値, or the central 际 when the brightness is arranged as the 2 値The threshold threshold used for the two thresholds, or the average threshold of most pixels located near the center threshold when the brightness is arranged, can be used as the threshold threshold for the two thresholds. -16- (13) (13) 200307457 In addition 'The image signal processing unit 5 in the detection window determines that the luminance distribution in the horizontal direction of the image signals VD1 and VD2 in the detection window is monotonically increasing or monotonically decreasing, and if it is not monotonically increasing or monotonically decreasing', it may be Give the smallest 値 as "0" or the most For example, 2 5 5 is used as the threshold value in the 2 conversion unit 2. In this way, the 2 conversion unit 2 outputs all the 2 conversion modes BI composed of,, 1, or "0". In this case, the difference 値 between two pixels adjacent to the video signals VD1 and VD2 is sequentially calculated, and if the sign of the difference 相同 is the same, it can be determined as monotonically increasing or monotonically decreasing. The image signal processing unit 5 in the detection window detects the image signal processing unit 5 in the detection window and calculates the difference between the maximum 値 and minimum 的 of the brightness of the image signals VD1 and VD2 in the detection window for comparison, and the calculated comparison is smaller than a predetermined value. In the case of 値, the minimum 値 is set to “0” or the maximum 値 is “25 5” to the 2 conversion unit 2 as the threshold 値. As a result, the 2 conversion unit 2 outputs all 2 conversion patterns BI composed of "1" or "0". The reference pattern generation unit 6 generates a plurality of reference patterns RA composed of "1" and "0", and outputs them to the angle detection unit 3. The size of each reference pattern RA is the same as the size of the detection window. The angle detection unit 3 compares the 2D pattern BI provided by the 2D pattern unit 2 with the majority of each of the reference patterns RA provided by the reference pattern generation unit 6 and outputs the angle of the consistent reference pattern RA as the Angle information S 1. This angle will be detailed later. Hereinafter, a comparison operation between the 2 値 pattern BI and each reference pattern R A is referred to as a pattern matching. (14) 200307457 As described above, when the brightness distribution of the video signal VD 1 and the video signal VD2 in the detection window is neither monotonically increasing nor monotonically decreasing, all of the "1" can be output from the 2 conversion unit 2. Or "〇" in a 2-division mode BI. In this case, the angle information S 1 is not output from the angle detection section 3. In addition, if the contrast of the image signals VD 1 and VD 2 in the detection window is smaller than the predetermined value, all the 2 conversion modes BI consisting of “1” or “0” are output from the 2 conversion unit 2. Therefore, the angle information S 1 is not output from the angle detection unit 3. In the case where the contrast of the video signals VD1 and VD2 is small, the effect of the interpolation processing using the pixels in the oblique direction is small. In the pixel interpolation process using the oblique direction, if the correct angle is not detected, noise may be generated. Therefore, when the effect is small, the pixel interpolation process using the oblique direction is not performed and No angle information S 1 is output. The line recorder 1 b delays the angle information S 1 output by the angle detection unit 3 by 1 line (1 scan line), and outputs the angle information S 2 delayed by 1 line in a circle. The arc shape detection unit 4 and the line memory 1 c. The line memory lc delays the angle information S2 output by the line memory lb by one line (one scan line), and outputs the angle information S3 delayed by one line by the arc shape detection unit. 4. Here, a scanning line including an interpolation pixel is referred to as an interpolation scanning line. The arc shape detection unit 4 not only includes the angle information S 3 of the interpolation scan line of the upper line, and the angle information S 1 of the interpolation scan line of the lower line, but also the angle of the interpolation scan line of the object. In the combination of information S 2, output the image -18-(15) (15) 200307457 edge angle information τ 1 on the object's interpolated scan line, and also recognize the arc shape, and output the arc shape information τ 2 as the object. Interpolated scan lines. The angle detection and the identification of the arc shape will be described in detail later. Fig. 2 is a pattern diagram showing an example of a 2-division pattern BI output by the 2-division unit 2 of Fig. 1. In Figure 2, IN shows interpolated pixels and IL shows interpolated scan lines. In addition, AL shows the upper scan line of the interpolation scan line IL, and BL shows the lower scan line of the interpolation scan line IL. In the example in Figure 2, the lower part (the darker part) is displayed with "0", and the higher part (the lighter part) is displayed with "1". In 2D mode BI, the angle of the image edge is 45 °. Here, the angle in the horizontal direction is 0 °, and the angle in the upper right oblique direction is positive. Figs. 3, 4, 5, and 6 are schematic diagrams showing examples of reference patterns generated by the reference pattern generation unit 6 of Fig. 1. The pixels that are highlighted are pixels used to calculate the scan lines above and below the interpolated pixels displayed in thick lines. (A), (b), (c), (d), (e), and (f) in the third figure show 45 ° and 34, respectively. , 27. ,twenty two. , 18. , 16. Reference mode. In the example in Figure 3, the darker part is displayed in the upper left and the lighter part is displayed in the lower right. (A), (b), (c), (d), (e), (f) in Figure 4 each show 45. , 34. , 27. ,twenty two. , 18. , 1 ό. Reference model. In the example in Figure 4, the brighter part is shown in the upper left and the darker part is shown in the lower right. (A), (b), (c), (d), (e), (f • 19- (16) (16) 200307457) in Figure 5 show -45 ° and -34, respectively. , -27. ,-twenty two. , -18. ,-16. The reference model. In the example of Fig. 5, the darker part is displayed in the upper right and the lighter part is displayed in the lower left. (A), (b), (c), (d), (e), (f) in Figure 6 show -45 °, -34 °, -27, respectively. , -22 °, -18 °,-16 ° reference mode. In the example in Figure 6, the brighter part is displayed in the upper right and the darker part is displayed in the lower left. In the angle detection unit 3, the reference patterns shown in Figs. 3 to 6 are compared with the two conversion patterns BI output by the two conversion units 2, and then the angle detection unit 3 outputs the same reference pattern Angle information. In addition, as shown in FIGS. 3 to 6, in the reference pattern based on the quadratic element luminance distribution, not only the angle of a straight line connecting pixels connected to a point-symmetrical position centered on the interpolated pixel can be set, but also Set the angle between the two. For example, it can be set at 45 ° and 2 7. 34 ° and 22 ° between 45 ° and 18 °. For example, the two-dimensionalization pattern BI of FIG. 2 may be consistent with one of the six reference patterns of FIG. 4 (a). In this case, the angle detection unit 3 in FIG. 1 outputs 45 ° shown in the reference pattern in FIG. 4 (a) as the angle information S1. The reference pattern RA generated by the reference pattern generation unit 6 in Fig. 1 is not limited to the examples shown in Figs. 3 to 6, and any reference pattern may be used. Fig. 7 is a diagram for explaining processing of the arc-shaped detection unit 4 of Fig. 1. Figure 7 shows the detection angle corresponding to the detection angle of the interpolated pixels, the detection angle in the interpolation scan line below, and the detection angle in the interpolation scan line above. 20- (17) (17) 200307457 Combination, examples of recognized arc shapes. Specifically, the arc shape can be classified into a combination of five conditions A, and B, and C, and D, and E in FIG. 7. In the combination example of the situation A in FIG. 7, the convex direction of the edge (arc edge) showing the arc shape is “lower right” and the direction toward the inside of the arc shape is “left”. In this case, the absolute angle 値 of the detection angle of the interpolated pixel is the middle 値 of the absolute angle 値 of the detection angle in the lower interpolation scan line and the absolute 値 of the detection angle in the upper interpolation scan line. In addition, the absolute detection angle of the upper interpolation scan line is larger than the absolute detection angle of the lower interpolation scan line, and all the detection angles are positive. The arc shape detection unit 4 not only outputs the detection angle of the interpolated pixels as the angle information T1, but also recognizes the arc shape, and outputs the direction toward the inside of the arc shape as the arc shape. Information T2. In addition, on the far right of Fig. 7, an arc-shaped pattern diagram and an example of the inside of the arc are displayed. The connection of most slender arrows shows the recognized arc, the direction of each slender arrow shows the angle detected along the arc, and the thick arrow shows the inside of the arc. Among the combinations of conditions B, C, and D in FIG. 7, the same as the condition A in FIG. 7 shows the convex direction of the arc edge and the direction toward the inside of the arc shape. In the combination example of the situation B in FIG. 7, the convex direction of the edge of the arc is "upper right", and the inside direction of the arc shape is "left". In this case, the absolute angle 値 of the detection angle of the interpolated pixels is the absolute angle 値 of the detection angle in the lower interpolation scan line and the detection -21 in the upper interpolation scan line. ) 200307457 The absolute value of the middle angle of the angle, and the absolute value of the detected angle in the upper interpolation scan line is smaller than the absolute value of the detected angle in the lower interpolation scan line. All the detected angles are negative. In the combination example of the situation C in Fig. 7, the convex direction of the arc edge is "upper left", and the inner direction of the arc shape is "right". In this case, the absolute angle 値 of the detection angle of the interpolated pixel is the middle 値 of the absolute angle 値 of the detection angle in the lower interpolation scan line and the absolute 値 of the detection angle in the upper interpolation scan line. And the absolute angle 値 of the detection angle in the upper interpolation scan line is smaller than the absolute value 値 of the detection angle in the lower interpolation scan line, and all the detection angles are positive. In the combination example of the situation D in FIG. 7, the convex direction of the arc edge is "lower left", and the inner direction of the arc shape is "right". In this case, the absolute angle 値 of the detection angle of the interpolated pixel is the middle 値 of the absolute angle 値 of the detection angle in the lower interpolation scan line and the absolute 値 of the detection angle in the upper interpolation scan line. And the absolute angle 値 of the detection angle in the upper interpolation scan line is smaller than the absolute value 値 of the detection angle in the lower interpolation scan line, and all the detection angles are negative 値. In the combination example of the situation E in Fig. 7, a combination of detection angles at which the edges of the arc cannot be recognized is displayed. That is, all the combinations of the combination examples of the conditions A, B, C, and D that do not belong to FIG. 7 belong to the condition E of FIG. 7. 〇 Here, the so-called middle 値 refers to being in the two numbers 値値 between X and Y, if X < Y, all 値 greater than X and less than Y are intermediate 値. -22- (19) (19) 200307457 In the upper interpolation scan line, the reference position of the detection angle in the upper interpolation scan line corresponds to the detection of the interpolation pixel with the interpolation pixel 1 point in the direction determined by the angle, and in the lower interpolation scan line, the reference position of the detection angle in the lower interpolation scan line corresponds to the detection of the interpolation pixel with the interpolation pixel 1 point in the direction determined by the angle. In the upper interpolation scan line, the reference position of the detection angle in the upper interpolation scan line can also be adopted, including 1 point in a direction determined by the detection angle of the interpolation pixel for the interpolation pixel. An area of a certain width in the horizontal direction (area of a large number of pixels). In addition, in the lower interpolation scan line, the reference position of the detection angle in the lower interpolation scan line can also be adopted, including the position in the direction determined for the interpolation pixel by the detection angle of the interpolation pixel. An area of a certain width in the horizontal direction at one point (area of a large number of pixels). Regarding the combination of the detection angle of the interpolation pixel in FIG. 7 and the detection angle in the interpolation scan line below and the detection angle in the interpolation scan line above, only the above examples are given, but It is not limited to this, and combinations not shown in the figure may be adopted. In the image angle detection device 10 of this embodiment mode, the brightness distribution of the image signals VD1 and VD2 in the detection window is converted into a 2D mode BI, and the 2D mode BI and a preset majority The pattern comparison is performed with reference to the pattern RA, whereby the oblique edge angle of the image can be detected with a very small circuit scale. In this case, because the average brightness 値 in the detection window is used as the -23- (20) (20) 200307457 2-threshold threshold, there is no need to set the 2-threshold threshold from the outside. The brightness level of the image signal must include the two-division mode BI of both "0" and "1". In addition, because the pattern comparison based on the brightness distribution of the two-dimensional element is performed, 'compared with the case where the difference between two pixels is used,' false detection can be suppressed, and the image with oblique edges can be accurately detected. angle. Furthermore, by adopting the reference pattern RA based on the brightness distribution of the quadratic element, not only can it be detected, but also a straight line angle between pixels whose detection angle is located at a point symmetrical about the interpolated pixel can be detected. The angle between these angles. Therefore, a smaller amount of line memory 1 a can be used to detect the angle at more precise intervals. In addition, since the combination of the detection angle of the interpolated pixels, the detection angle in the lower interpolation scan line, and the detection angle in the upper interpolation scan line can be used to identify the arc shape, no preparation is required. There must be more than 3 scan lines for the reference pattern used to identify the arc shape. Therefore, it is not necessary to increase the circuit scale or calculation scale of the image angle detection device 10 to detect the image angle and recognize the arc shape. Fig. 8 is a block diagram showing the configuration of a scanning line interpolation device provided with the image angle detecting device of Fig. I. In FIG. 8, the 'scan line interpolation device j 〇 〇 is composed of an image angle detection device and an interpolation circuit 20. The video signal VA is input to the image angle detection device 10 and the interpolation circuit 20. The image angle detection device 10 is composed of the image angle detection device 10 shown in Fig. 1. Based on the image signal VA, the image angle detection device detects the angle of the beveled edge and the arc shape of the image, and outputs the image edge angle information T1 and the arc shape information T2. Based on the angle information T1 and the arc shape information T2, the interpolation circuit 20 selects pixels in the oblique direction of the interpolation pixel from the upper and lower scanning lines, and then uses the brightness 値 of the selected pixel to calculate the brightness 値 of the interpolation pixel and outputs The video signal V Ο UT is interpolated. In the scanning line interpolation device 100 of FIG. 8, not only the linear shape can be accurately detected by the image angle detection device 10, but also an image having an oblique edge including an arc shape can be accurately detected. angle. Therefore, even in images that include not only linear shapes but also oblique edges with circular arc shapes, appropriate pixels in the oblique direction can be selected for smooth interpolation processing. Fig. 9 is a block diagram showing the configuration of the interpolation circuit 20 in the scanning line interpolation device 100 of Fig. 8. The interpolation circuit 20 in FIG. 9 includes an A / D (Analog / Digital) converter 21, a line memory 22, an interpolation pixel selection circuit 23, and an average unitary operation circuit 24. The A / D converter 21 performs analog-to-digital conversion on the analog video signal VA and outputs a digital video signal VD1. The video signal VD 1 output from the A / D converter 21 is input to the line memory 22 and the interpolation pixel selection circuit 23. The line memory 22 delays the video signal VD 1 outputted from the A / D converter 21 by one line (one scanning line) and outputs it. The video signal VD2 output from the line memory 22 is supplied to the interpolation pixel selection circuit 23. The interpolation pixel selection circuit 23 uses the given image signals VD 1, (22) (22) 200307457 VD 2, and the angle information τ 1 and the arc shape information T 2 of the image angle detection device 10, from among the upper scanning lines. , The interpolation reference pixel ρ 1 is selected and outputted to the average unitary operation circuit 24, and the interpolation reference pixel P 2 is selected from the lower scan lines and outputted to the average unitary operation circuit 24. The average unitary arithmetic circuit 24 calculates the brightness 値 of the interpolated pixel from the interpolated reference pixel ρ 1 and the interpolated reference pixel P2 and outputs it. The selection of the interpolation reference pixels by the interpolation pixel selection circuit 23 is performed in such a manner that the edges in the image are smoothed during the interpolation. According to the interpolation pixel selection circuit 23, the operation in the case where the image has a straight edge and the operation in the case where the image has an arc-shaped edge are selectively performed. That is, when the image has a straight edge in the portrait, the interpolation pixel selection circuit 23 selects the pixels in the highlighted part shown in FIGS. 3 to 6 from the upper and lower scanning lines based on the angle information T1. As an interpolation reference pixel. The center position of the interpolation reference pixel is the interpolation reference position. In addition, the interpolation pixel selection circuit 23 specifies the position of the direction indicated by the angle information T 1 for the interpolation pixel in the upper and lower scanning lines when the image has an arc-shaped edge, and is based on the arc shape. Information T2, for the specified position, a position moved horizontally among the inner sides of the arc shape is selected as the interpolation reference position. Here, FIG. 7 is used as an example. When the arc shape of the condition A is detected, 'the inner pixel direction of the arc shape is to the left as indicated by the thick arrow, the interpolation pixel selection circuit 23 is among the upper and lower scanning lines' to specify the interpolation pixel The position of the direction indicated by the angle information T 1 is selected from the positions shifted to the left from the specified position. -26 · (23) (23) 200307457 selects the respective interpolation reference positions above and below. In addition, when the image has a circular arc shape, when the interpolation reference position 'is selected when the image has a straight-line edge, the selected interpolation position' is shifted by 0.5 pixels as the interpolation reference position. Next, it was confirmed through experiments that this method can form the smoothest arc shape by interpolation. The average 値 calculation circuit 24 calculates the average brightness 値 of the pixel to which the interpolation reference position belongs, and determines the brightness 値 of the interpolation pixel. Although not shown in the figure, the average unitary operation circuit 24 may also include a correlation operation circuit, which can perform the correlation between the pixels belonging to the interpolation reference position, that is, within the difference size of the pixels belonging to the interpolation reference position. Interpolation. In this case, 'even if the angle and arc shape are incorrectly detected for some reason, the noise caused by interpolation can be reduced. Fig. 10 is a schematic diagram showing an example of an image angle detected by the image angle detection device 10 of Fig. 1. Fig. 10 shows an example of a portrait having an arc-shaped edge. Fig. 11 is a diagram showing an example of interpolation of pixels using angle information of an image detected by the day image angle detection device 10 of Fig. 1 and arc shape information. In contrast, FIG. 12 is a diagram showing an example of interpolation of pixels using only the angle information of the image detected by the image angle detection device 10 of FIG. 1. In Figs. 10 to 12, IL1, IL2, and IL3 show interpolated scan lines, and AL, BL, CL, and DL show interpolated scan lines. In Figures 10 to 12, each frame on AL, BL, CL, and DL shows the brightness frame of each pixel. In Fig. 10, the interpolated scan lines IL1, IL2, (24) 200307457 are line 0. The inner position of the image is R2 and each frame on IL3 displays angle information. In FIG. 11 and FIG. 2, each 値 on the scan lines IL1, IL2, and IL3 is interpolated to display the brightness 値 of each interpolated pixel. Here, pixels are interpolated with IN as a target. In this case: [L2 object interpolation scan line 'BL is the upper scan line, CL is scan below' IL1 is the upper interpolation scan line, IL3 is the lower interpolation scan line In addition, as shown in Figure 11 and In Fig. 12, the interpolation parameter positions P1, P2, P3, and P4 are displayed by X marks. In the example in FIG. 10, the angle of the portrait of the interpolated pixel IN is 27. . In the example in FIG. 12, in the upper and lower scanning lines b L and C L, 2 7 of the pixel IN is selected for the object interpolation. Position in the direction of, to do the interpolation reference position P3, P4. Since the average luminances of the pixels Q 3 and Q 4 to which the reference positions p3 and P4 belong are “× 〇〇”, therefore, the pixel Q 3 and Q 4 to which the reference positions P 3 and P 4 belong are interpolated. The interpolation operation makes the brightness 値 of the interpolation pixel IN of the object to be “χ 〇〇,”. The brightness 値 of other interpolated images is also calculated in the same way, so the result of Fig. 12 can be obtained. In the example in FIG. 11, the reference pixel of the interpolation scan line IL1 above is the reference pixel R2 of the interpolation scan line IL2 below R 1 ′. Because the angle information of the reference pixel R1 is 45 ° and the angle information of the reference pixel is 18 °, the example in FIG. 11 is equivalent to the condition A in FIG. 7, and the direction of the inside of the arc shape is, left. " Therefore, in the lower scan lines BL and CL, the positions p 1 and p 2 in the direction of 2 7 of the interpolation pixel I n are specified for the target, and the positions p 1 -28- (25) (25 ) 200307457 p 2 moves the position of 0.5 pixels in the horizontal direction as the interpolation reference positions PI, P2. The brightness 値 of the pixel to which the interpolation reference position P1 ′ P2 belongs, each being two adjacent pixels Q 1 ; the brightness l of l, qi 2 ,, 〇〇, and, 〇 ,, the average brightness 値 "50" ° The brightness 其他 of other interpolated pixels is also calculated in the same way, so the first The processing results of Figure 1. In Figures 11 and 12, each frame on the scan lines IL1, IL2, and IL3 is interpolated to display the brightness of each interpolated pixel, and the black points are the brightness on each scan line. The center 値 of the image, and the dotted line shows the dots connecting the center 値 of each brightness in a straight line. In the case of this example, the central 値 is the brightness 値 "50". If you compare the processing results of Figure 11 and the processing results of Figure 12 ', then only the portrait shown in Figure 12 is used In the case where the brightness 値 of the interpolated pixels is calculated by the angle information, the edge shape of the image after interpolation is a polyline shape, and smooth edges cannot be obtained by interpolation. That is, because the processing of FIG. 12 is locally identified Interpolation is performed in the angle direction, so if the continuity is viewed from a wide range, smooth interpolation may not be achieved. In contrast, based on the angle information and arc shape information of the portrait shown in Figure 11 When the brightness 値 of the interpolated pixel is calculated, a smooth edge similar to an arc shape can be displayed. In this example, the movement amount of the interpolation reference position when the shape is recognized as an arc shape is 0.5. As an example, it is not limited to this, and an arbitrary amount of movement may be set. In this embodiment, the image processing unit 2 and the detection signal in the detection window -29- (26) (26) 200307457 processing Part 5 is equivalent to 2 units In the pattern generator, the reference pattern generator 6 corresponds to a reference pattern generator, and the angle detection unit 3 corresponds to a comparator. In addition, the arc shape detection unit 4 corresponds to a shape detector. Furthermore, the image in the window is detected. The signal processing unit 5 corresponds to a threshold value calculation device and a discriminator, and the 2D unit 2 corresponds to a 2D device. [Brief Description of the Drawings] FIG. The block diagram of the structure. Fig. 2 shows an example of a 2D mode output from the 2D unit of the same device. Table 3 shows an example of a reference mode not generated by the reference mode generating unit of Fig. 1 Pattern illustration. Fig. 4 is a pattern diagram showing an example of a reference pattern generated by the reference pattern generating section of Fig. 1. Fig. 5 is a pattern diagram showing an example of a reference pattern generated by the reference pattern generating section of Fig. 1. Fig. 6 is a pattern diagram showing an example of a reference pattern generated by the reference pattern generating section of Fig. 1. Fig. 7 is a schematic diagram showing the processing of the arc-shaped detection section of Fig. 1; Fig. 8 is a block diagram showing a configuration of a scanning line interpolation device including the image angle detecting device of Fig. 1; Fig. 9 is a block diagram showing the constitution of the interpolation circuit -30- (27) (27) 200307457 in the scanning line interpolation device of Fig. 8. Fig. 10 is a schematic diagram showing an example of an image angle detected by the image angle detection device of Fig. 1. Fig. 11 is a schematic diagram showing an example of interpolation of pixels using angle information of an image detected by the image angle detection device of Fig. 1 and arc shape information. Fig. 12 is a schematic diagram showing an example of interpolation of pixels using only angle information of an image detected by the image angle detection device of Fig. 1. [Comparison table of main components] la, lb, lc, 22-line memory 2 2 digitizing section 3 angle detecting section 4 arc shape detecting section 5 detecting image signal processing section 6 in window view reference pattern generating section 7, 21 A / D converter 10 Image angle detection device 20 Interpolation circuit 23 Interpolation pixel selection circuit 24 Average 値 Operation circuit 100 Scan line interpolation device AL, BL, CL, DL Interpolation scan line above scan line ΒΙ 2 Scanning mode-31-(28) (28) 200307457 BL Interpolation scan line below scan line IL, IL1, IL2, IL3 Interpolation scan line IN Interpolation pixel LU average brightness 値 p 1, P 2 position P1 , P 2, P 3, P 4 interpolate reference positions Q3, Q4, Qll, Q12 pixels
Rl，R2參照像素 0 RA參照模式 S1，S2，S3 角度資訊 T 1畫像邊緣角度資訊 T2圓弧形狀資訊 VA類比影像信號 VD1，VD2數位影像信號 VOUT內插影像信號 參 -32-R1, R2 reference pixels 0 RA reference mode S1, S2, S3 Angle information T 1 Image edge angle information T2 Arc shape information VA analog image signal VD1, VD2 digital image signal VOUT interpolated image signal Reference -32-