200946899 六、發明說明： 【發明所屬之技術領域】 本發明有關於光學式檢查膜上之缺陷之缺陷檢測方法 及裝置。 【先前技術】 近年來在擴大需要之液晶顯示器，將作爲偏極光板之 保護膜或光學補償膜之具有光學式異向性之相位差膜黏貼 在螢幕。相位差膜（以下簡稱爲「膜」）利用在透明之支持 〇 體形成定向膜之步驟，和在形成有定向膜之支持體塗布液 晶後形成液晶層之步驟製造（例如日本專利特開平9 -7 3 08 1號公報）。在各個步驟進行嚴格之品質管理，但是要 完全消除由於製造中之異物之混入、附著造成之分子定向 不均（定向缺陷），液晶層之塗布厚度不均（相位差缺陷）等之 缺陷會有困難。 當在該膜有缺陷時，在該處發生光之散射，所以在膜 之製造步驟設置光學式之缺陷檢測裝置，檢測膜之缺陷。 〇 該缺陷檢測裝置之構成包含：投光器，對膜照射光；受光 器，接受從該膜發出之光；和處理電路，用來解析來自受 光器之輸出信號。該受光器使用具有攝像裝置（線感測器或 區域感測器）之視訊攝像機，輸出在膜之寬度方向線掃描之 掃描信號（攝像信號）作爲輸出信號。 另外，以在投光器和受光器之間使偏極光方向互相正 交（正交尼科耳（crossed Nichol))之方式，配置一對偏極光 板。該一對偏極光板在膜上沒有缺陷時，光不會進入到受 光器，在膜上有缺陷時，因該缺陷而散射、彌散之光進入 -4 - 200946899 到受光器，利用此種方式使缺陷之檢測容易（日本專利特開 平6— 14 8095號公報）。另外，由於藉由對來自受光器之輸 出信號進行微分處理，強調缺陷部份之信號，和減小缺陷 部份以外之信號之所謂雜訊信號，因此亦容易進行缺陷之 檢測。 近年來由於要求更高品質之膜，所以希望有亦可以檢 測至微細之缺陷之精度佳之缺陷檢測方法。缺陷之檢測精 度一般利用以缺陷部份以外之信號（雜訊（noise)信號），除來 〇 自受光器之輸出信號中之缺陷部份之信號（缺陷信號）之 値（S/N)表示，該S/N越大，缺陷之精度越高。 因此，爲獲得更大S/N，須使缺陷信號變大，和將雜 訊信號抑制至最小限度。使缺陷信號變大之方法可以考慮 使高亮度之光對膜照射，用來使在膜上之缺陷散射、彌散 之光之強度變大。 但是，如同金屬鹵化物燈泡，在發出高亮度之光之投 光器存在有複數急激尖峰値之所謂亮線。在使具有此種複 © 數亮線之光照射在如同膜之薄透明體之情況下，會在膜上 產生具有光相互增強化部份之干涉條紋。因此，強光亦照 射在沒有缺陷之部份，所以雜訊信號變大，結果造成S/N 降低。另外，在將一對偏極光板配置成正交尼科耳，同時 相對於膜之遲相軸傾斜情況下，即使沒有缺陷，仍因從一 對偏極光板透過之光很少，所以當存在有複數亮線之光照 射在膜時，雜訊會因干涉條紋之影響而變大。因此，即使 是具有高亮度之投光器，在具有複數亮線情況下，S/N反 而會因干涉條紋之影響而降低。 200946899 另外，在將一對偏極光板配置成正交尼科耳而檢測缺 陷情況下，來自受光器之輸出信號因膜之光學特性或受光 器之敏感度特性而使膜之寬度方向之中央部之亮度變高， 從其中央部朝向兩端部使亮度逐漸變小。在對此種輸出信 號進行微分處理情況下，缺陷信號會受到其周邊之雜訊影 響。 例如，如第7(A)圖所示，在膜上存在有複數相同程度 之缺陷，在各個缺陷之散射、彌散之光之亮度相同情況下， 〇 當對輸出信號進行微分處理時，如（B)所示，雜訊信號相對 於膜寬度方向之斜率爲負之區域的缺陷信號小於斜率爲正 之區域的缺陷信號。因此，如第7(B)圖所示，在將缺陷檢 測用之臨限値設定在一定之亮度値Th情況下，雜訊信號之 斜率爲正之區域之缺陷信號檢測在Th以上，另一方面，斜 率爲負之區域之缺陷信號檢測爲不滿Th，所以未被檢測 到。由於任一缺陷均爲相同程度，因此，儘管散射、彌散 之光之亮度相同，仍會發生不能依照雜訊信號之斜率爲正 〇 或負來檢測缺陷信號之問題。 【發明内容】 本發明之目的是提供不會降低S/N，可進行精度佳之 缺陷檢測之缺陷檢査方法及裝置。 另外，本發明之目的是提供可防止缺陷部份之檢測失 誤之缺陷檢査方法及裝置。 爲達成該目的和其他目的，本發明是一種缺陷檢測裝 置，其在配置成正交尼科耳之第1偏極光板和第2偏極光 板之間配置有膜之狀態下，檢測該膜上之缺陷，具備：投 200946899 光手段，使未具有或只具有一亮線之光經由該第1偏極光 板’投光在該膜；受光手段，經由該第2偏極光板，接受 從該膜發出之光；和缺陷檢測手段，根據來自該受光器之 輸出信號，檢測該膜之缺陷。 較佳係使該膜爲相位差膜，該第1和第2偏極光板之 偏極光方向相對於該相位差膜之遲相軸朝向45°之方向。較 佳係使該投光手段爲鹵素燈泡。較佳係使該第1和第2偏 極光板爲碘素系偏極光板》 © 本發明之缺陷檢測方法具備：使未具有或只具有一亮 線之光經由該第1偏極光板，投光在該膜之步驟；經由該 .第2偏極光板，以受光手段接受從該膜發出之光之步驟； 和根據來自該受光器之輸出信號，檢測該膜之缺陷之步驟。 本發明是一種缺陷檢測裝置，用來檢測膜之缺陷，具 備：受光手段，具有排列成線狀之複數像素，在該膜之寬 度方向掃描，輸出時間系列之輸出信號；微分處理手段， 對該輸出信號施以微分處理；極値信號檢測手段，檢測在 © 利用該微分處理手段所獲得之微分處理信號中既定像素範 圍內亮度値成爲最大之極大信號，和既定像素範圍內亮度 値成爲最小之極小信號；差分値算出手段，求得預先設定 之基準信號之亮度値和該極大信號之亮度値之差分作爲第 1差分値，和該基準信號之亮度値和該極小信號之亮度値 之差分作爲第2差分値；加算手段，求得第1差分値和第 2差分値相加後之加算値；和缺陷信號指定手段，在該加 算値爲一定値以上情況下，指定該極値信號檢測手段所檢 測到之極大信號和極小信號成爲缺陷信號。 200946899 光板，配置成偏 該第1偏極光板 係使該膜爲相位 向相對於該相位 備排列成線狀之 掃描用來獲得時 施加微分處理之 範圍內亮度値成 度値成爲最小之 號之亮度値和該 ，和求得該基準 分作爲第2差分 相加後之加算値 下，指定該極大 投光手段照明該 之缺陷檢測。另 號之最大値和最 檢測失誤。 備定向膜形成裝 I 13及捲取裝置 另外，較佳係具備：第1和第2偏極 極光方向互相正交；和投光手段，對位於 和該第2偏極光板之間之該膜投光。較佳 差膜，該第1和第2偏極光板之偏極光方 差膜之遲相朝向45°之方向》 本發明之缺陷檢測方法具備：利用具 複數像素之受光手段，在該膜之寬度方向 間系列之輸出信號之步驟；對該輸出信號 © 步驟；檢測在該微分處理信號中既定像素 爲最大之極大信號，和既定像素範圍內亮 極小信號之步驟；求得預先設定之基準信 .極大信號之亮度値之差分作爲第1差分値 信號之亮度値和該極小信號之亮度値之差 値之步驟：求得第1差分値和第2差分値 之步驟；和在該加算値爲一定値以上情況 信號和極小信號成爲缺陷信號之步驟。 〇 依照本發明時，因爲使用亮線少之; 膜，所以不會降低S/ N，可以進行精度佳 外，依照本發明時，因爲利用微分處理信 小値檢測缺陷，所以可以防止缺陷部份之 【實施方式】200946899 VI. Description of the Invention: [Technical Field of the Invention] The present invention relates to a defect detecting method and apparatus for optical inspection of defects on a film. [Prior Art] In recent years, in order to expand the liquid crystal display required, a retardation film having an optical anisotropy as a protective film or an optical compensation film of a polarizing plate is attached to the screen. The retardation film (hereinafter simply referred to as "film") is produced by a step of forming an alignment film in a transparent supporting body, and a step of forming a liquid crystal layer after coating a liquid crystal on a support having an oriented film (for example, Japanese Patent Laid-Open No. 9- 7 3 08 1 bulletin). Strict quality management is carried out in each step, but the molecular orientation unevenness (orientation defect) caused by the incorporation and adhesion of foreign matter in manufacturing is completely eliminated, and the coating thickness unevenness (phase difference defect) of the liquid crystal layer may be defective. difficult. When the film is defective, light scattering occurs there, so that an optical defect detecting device is provided in the film forming step to detect defects of the film. 〇 The defect detecting device comprises: a light projector that emits light to the film; a light receiver that receives light emitted from the film; and a processing circuit that analyzes an output signal from the light receiver. The light receiver uses a video camera having an image pickup device (line sensor or area sensor) to output a scan signal (image pickup signal) scanned in the width direction of the film as an output signal. Further, a pair of polarizing plates are disposed such that the directions of the polarization directions are orthogonal to each other (crossed Nichol) between the light projector and the light receiver. When the pair of polarized light plates have no defects on the film, the light does not enter the light receiver. When there is a defect on the film, the light scattered and diffused by the defect enters the -4 - 200946899 to the light receiver. The detection of defects is easy (Japanese Patent Laid-Open No. Hei 6-148095). Further, since the output signal from the photoreceiver is subjected to differential processing, the signal of the defective portion is emphasized, and the so-called noise signal of the signal other than the defective portion is reduced, the defect is easily detected. In recent years, since a film of higher quality is required, it is desirable to have a defect detecting method capable of detecting fine defects. The detection accuracy of the defect is generally determined by the signal (noise signal) other than the defective portion, and the signal (defect signal) of the defective portion in the output signal of the photoreceptor is removed (S/N). The larger the S/N, the higher the accuracy of the defect. Therefore, in order to obtain a larger S/N, it is necessary to make the defect signal large and to suppress the noise signal to a minimum. In order to increase the defect signal, it is conceivable to irradiate the film with high-intensity light to increase the intensity of the light scattered and diffused on the film. However, like a metal halide bulb, there is a so-called bright line of a plurality of sharp spikes in a light projector that emits high-intensity light. In the case where light having such a complex bright line is irradiated on a thin transparent body like a film, interference fringes having a mutually reinforcing portion of light are generated on the film. Therefore, the glare is also irradiated in the portion where there is no defect, so the noise signal becomes large, resulting in a decrease in S/N. In addition, when a pair of polarized light plates are arranged in crossed Nicols and inclined with respect to the retardation axis of the film, even if there is no defect, there is little light transmitted from the pair of polarized plates, so when there is When light with a plurality of bright lines is applied to the film, the noise becomes larger due to the influence of the interference fringes. Therefore, even in the case of a light projector having high luminance, in the case of having a plurality of bright lines, the S/N may be lowered by the influence of interference fringes. 200946899 In addition, when a pair of polarized light plates are arranged in crossed Nicols to detect defects, the output signal from the light receiver causes the central portion of the width direction of the film due to the optical characteristics of the film or the sensitivity characteristics of the light receiver. The brightness is increased, and the brightness is gradually reduced from the central portion toward both end portions. In the case of differential processing of such an output signal, the defective signal is affected by noise around it. For example, as shown in Fig. 7(A), there are a plurality of defects of the same degree on the film, and when the brightness of the scattered and diffused light of each defect is the same, when the output signal is differentially processed, such as ( As shown in B), the defect signal in the region where the slope of the noise signal with respect to the film width direction is negative is smaller than the defect signal in the region where the slope is positive. Therefore, as shown in Fig. 7(B), when the threshold for defect detection is set to a certain brightness 値Th, the defect signal in the region where the slope of the noise signal is positive is detected above Th. The defect signal in the region where the slope is negative is detected as dissatisfied with Th, so it is not detected. Since any of the defects are of the same degree, even if the brightness of the scattered and diffused light is the same, there is a problem that the defect signal cannot be detected in accordance with the slope of the noise signal being positive or negative. SUMMARY OF THE INVENTION An object of the present invention is to provide a defect inspection method and apparatus which can perform defect detection with high precision without lowering S/N. Further, it is an object of the present invention to provide a defect inspection method and apparatus which can prevent detection of a defective portion. In order to achieve the object and other objects, the present invention provides a defect detecting device for detecting a film in a state in which a film is disposed between a first polarizing plate and a second polarizing plate disposed in a crossed Nicols. The defect includes: applying a light means of 200946899 to cause light having no or only one bright line to be emitted to the film via the first polarizing plate; and receiving light from the film via the second polarizing plate The emitted light; and the defect detecting means detect the defect of the film based on the output signal from the photodetector. Preferably, the film is a retardation film, and the polarization directions of the first and second polarizing plates are oriented in a direction of 45 with respect to the slow axis of the retardation film. Preferably, the light projecting means is a halogen bulb. Preferably, the first and second polarizing plates are made of an iodine-based polarizing plate. The defect detecting method of the present invention includes: passing light having no or only one bright line through the first polarizing plate a step of illuminating the film; a step of receiving light emitted from the film by the light receiving means via the second polarizing plate; and a step of detecting a defect of the film based on an output signal from the photoreceiver. The present invention relates to a defect detecting device for detecting a defect of a film, comprising: a light receiving means having a plurality of pixels arranged in a line shape, scanning in a width direction of the film, outputting an output signal of a time series; differential processing means, The output signal is subjected to differential processing; the extreme signal detection means detects the maximum luminance of the luminance 値 in the predetermined pixel range in the differential processing signal obtained by the differential processing means, and the luminance 値 in the predetermined pixel range becomes the minimum The minimum signal; the difference 値 calculation means obtains the difference between the brightness 値 of the preset reference signal and the brightness 値 of the maximum signal as the first difference 値, and the difference between the brightness 该 of the reference signal and the brightness 値 of the minimum signal The second difference 値; the addition means obtains the addition 値 after the addition of the first difference 値 and the second difference 値; and the defect signal specifying means specifies the detection method of the pole signal when the addition 値 is equal to or greater than 値The detected maximal signal and very small signal become defect signals. 200946899 The light plate is configured to be biased to the first polarizing plate so that the film is phase-aligned with respect to the phase, and the scanning is used to obtain the brightness in the range where the differential processing is applied. The brightness 値 is summed, and the reference point is obtained as the addition of the second difference, and the maximum light-emitting means is designated to illuminate the defect detection. The biggest error and the most detection error. Further, the alignment film forming apparatus I 13 and the winding device further preferably include: the first and second polarized polar directions are orthogonal to each other; and the light projecting means for the film between the second polarizing plate and the second polarizing plate Cast light. Preferably, the retardation film of the first and second polarizing plates has a retardation direction of 45°. The defect detecting method of the present invention includes: a light receiving means having a plurality of pixels, in a width direction of the film The step of outputting the signal between the series; the step of detecting the output signal; detecting the maximum signal of the predetermined pixel in the differential processing signal, and the step of brightening the minimum signal within a predetermined pixel range; obtaining a predetermined reference signal. The difference between the brightness of the signal 作为 as the difference between the brightness 第 of the first differential 値 signal and the brightness 値 of the minimum signal 步骤: the steps of obtaining the first difference 値 and the second difference ;; and the addition 値 is constant The above situation signal and the minimum signal become the steps of the defect signal. 〇In accordance with the present invention, since the film is used, the S/N is not lowered, and the accuracy can be improved. According to the present invention, since the defect is detected by the differential processing signal, the defective portion can be prevented. [Embodiment]
如第1圖所示，相位差膜製造線10具 置11、液晶層形成裝置12、缺陷檢測裝B 定向膜形成裝置11對從膜捲筒18送出之透明樹脂膜 14 200946899 15之表面塗布包含定向膜形成用樹脂之塗布液，然後進行 乾燥處理。利用此種方式，在透明樹脂膜15之表面形成樹 脂層。然後，定向膜形成裝置11對樹脂層進行摩擦處理以 形成定向膜。 液晶層形成裝置12對定向膜上塗布包含液晶化合物之 塗布液，塗布後進行加熱乾燥以形成液晶層。然後，對液 晶層照射紫外線，以使定向膜和液晶層接合。利用此種方 式，獲得在透明樹脂膜1 5上形成有定向膜和液晶層之相位 差膜16(以下簡稱爲「膜」）。膜16在經過缺陷檢測裝置13 之後以捲取裝置14捲取。在此處從膜捲筒18退出至捲取 在捲取裝置14之膜16之搬運方向爲X方向。 缺陷檢測裝置1 3用來檢測在膜1 6上發生之缺陷。該 缺陷有如刮傷，厚度不均，塗布不均，分子定向不均等。 另外，缺陷檢測裝置之檢査對象除了相位差膜外，亦可爲 防止反射膜等》 缺陷檢測裝置13具備導引滾子20，21、投光器22、 光量調整部23、受光器24。第1和第2偏極光板25，26、 除去光學系27和控制器28。導引滾子20，21沿著膜16之 搬運路徑配置，隨著膜16之搬運進行旋轉。編碼器30連 接於導引滾子21，編碼器30在膜16每搬運一定長度產生 編碼脈波信號。編碼脈波信號發信到控制器28，以指定膜 16之長度方向之缺陷位置。 投光器22設在膜16之搬運路徑之下方，進行在膜16 之寬度方向延伸之帶狀之照明。該投光器22使用鹵素燈 泡。如第2圖所示，該鹵素燈泡之波長在600nm和800nm 之間，只具有一所謂亮線之放射強度之尖峰。由於從鹵素 200946899 燈泡發出之光只有一亮線，所以在膜16上幾乎不會產生干 涉條紋。另外，即使在暫時產生干涉條紋情況下，光相互 增強明亮部份之亮度只比從鹵素燈泡發出之光之亮度稍大 之程度。因此，干涉條紋不會對缺陷檢測之精度造成影響。 另外，只要是未具有或只具有一亮線之投光器，亦可以使 用鹵素燈泡以外者。另外，亦可以在鹵素燈泡和第1偏極 光板之間設置除去波長600nm以上之光之除去光學系，從 鹵素燈泡所發出之光中除去亮線。 光量調整部23根據設在投光器22附近之感測器（圖示 省略）所檢測到之光量檢測信號，控制投光器22。由於藉此 可對膜16照射光量均一之光，因此，可用經常相同之敏感 度進行缺陷檢測。 第1和第2偏極光板25，26由碘素系偏極光板構成， 在投光器22和膜16之間設置第1偏極光板25，在膜16 和受光器24之間設置第2偏極光板26。另外，第1和第2 偏極光板25, 26配置成使偏極光方向25a，26a互相正交（正 交尼科耳（crossed Nichol))。另外，雖然從性能和價格方面As shown in Fig. 1, the retardation film manufacturing line 10 is provided with a liquid crystal layer forming device 12, and the defect detecting device B is oriented to form a surface of the transparent resin film 14 200946899 15 which is fed from the film roll 18. The coating liquid for the resin for forming an oriented film is then subjected to a drying treatment. In this manner, a resin layer is formed on the surface of the transparent resin film 15. Then, the oriented film forming device 11 rubs the resin layer to form an oriented film. The liquid crystal layer forming apparatus 12 applies a coating liquid containing a liquid crystal compound to the alignment film, and after coating, heat-drying to form a liquid crystal layer. Then, the liquid crystal layer is irradiated with ultraviolet rays to bond the alignment film and the liquid crystal layer. In this manner, a retardation film 16 (hereinafter simply referred to as "film") having an alignment film and a liquid crystal layer formed on the transparent resin film 15 is obtained. The film 16 is taken up by the take-up device 14 after passing through the defect detecting device 13. Here, the film is ejected from the film roll 18 to the take-up. The conveyance direction of the film 16 of the take-up device 14 is the X direction. The defect detecting device 13 is used to detect defects occurring on the film 16. The defects are such as scratches, uneven thickness, uneven coating, and uneven molecular orientation. In addition to the retardation film, the defect detecting device 13 may include the guide rollers 20, 21, the light projector 22, the light amount adjusting portion 23, and the light receiver 24 in addition to the retardation film. The first and second polarizing plates 25, 26 remove the optical system 27 and the controller 28. The guide rollers 20, 21 are disposed along the transport path of the film 16, and are rotated as the film 16 is transported. The encoder 30 is coupled to a guide roller 21 which produces a coded pulse wave signal for each length of transport of the film 16. The encoded pulse signal is signaled to controller 28 to specify the location of the defect in the length direction of film 16. The light projector 22 is disposed below the conveyance path of the film 16, and performs strip-shaped illumination extending in the width direction of the film 16. The light projector 22 uses a halogen bulb. As shown in Fig. 2, the halogen bulb has a wavelength between 600 nm and 800 nm and has only a spike of the intensity of the so-called bright line. Since the light emitted from the halogen 200946899 bulb has only one bright line, there is almost no interference fringe on the film 16. Further, even in the case where interference fringes are temporarily generated, the brightness of the bright portions of the light enhancement is only slightly greater than the brightness of the light emitted from the halogen bulb. Therefore, interference fringes do not affect the accuracy of defect detection. Further, as long as it is a light projector that does not have or has only one bright line, a halogen bulb can be used. Further, a removal optical system for removing light having a wavelength of 600 nm or more may be provided between the halogen bulb and the first polarizing plate, and the bright line may be removed from the light emitted from the halogen bulb. The light amount adjustment unit 23 controls the light projector 22 based on the light amount detection signal detected by a sensor (not shown) provided in the vicinity of the light projector 22. Since the film 16 can be irradiated with light of uniform light amount, defect detection can be performed with often the same sensitivity. The first and second polarizing plates 25 and 26 are made of an iodine-based polarizing plate, and a first polarizing plate 25 is provided between the light projector 22 and the film 16, and a second polarized light is provided between the film 16 and the light receiver 24. Board 26. Further, the first and second polarizing plates 25, 26 are arranged such that the polarization directions 25a, 26a are orthogonal to each other (crossed Nichol). Also, although in terms of performance and price
W 考慮，使第1和第2偏極光板爲碘素系偏極光板，但是亦 可以使用其他染料系偏極光板，金屬偏極光器，或由方解 石等構成之偏極光板。 藉由第1和第2偏極光板25，26配置成正交尼科耳， 在沒有缺陷之膜16情況下，從投光器22發出之光以第1 偏極光板25偏光到偏極光方向25 a之後，直接透過膜16, 被另外一方之第2偏極光板26遮蔽。因此，透過正常部份 之光幾乎不會進入到受光器24。相對於此，在有異物混入， 刮傷等之定向缺陷之膜16情況下，從投光器22發出之光 -10- Ο Ο 200946899 以第1偏極光板25偏光到偏極光方向25a 上散射、彌散。散射、彌散之光因爲偏極光 以透過另外一方之第2偏極光板26進入受为 膜上之刮傷，異物混入等之定向缺陷產 以利用第1和第2偏極光板25, 26兩片之正5 可用只使散射光透過之受光器24檢測。但是 度不均等之缺陷（相位差缺陷）不產生散射光 和第2偏極光板25, 26兩片成爲正交尼科耳 中，不能檢測該缺陷。但是，該塗布厚度不 使第1和第2偏極光板25, 26兩片成爲正交 狀態下，可藉由使膜16之遲相軸和第1偏極 方向25a交叉，予以檢測。 第4圖表示使第1偏極光板之偏極光方 膜16之遲相軸之角度0產生變化時，透過第 光板25, 26兩片之光之透射光量比。在角度 消光模態，透射光量比（％)爲「0.00」。角g 爲藍模態。在該藍模態透射光量比（％)約爲 角度0成爲45°時，透射光量變成最大。在此 (%)表示第2偏極光板26之射出光相對於 25之射入光之比例。 另外，如第5圖所示，即使角度0相同 比（％)仍因膜16之缺陷部份相對於正常部份 光而變化。在該第5圖中，「〇」是角度0j 角度0爲15°，「Z\j是角度0爲30°，「〇」 °。相位差越大，透射光量變成越大。 如以上，在使第1和第2偏極光板25, 之後，在膜16 方向變化，所 3 器 24。 生散射光，所 交尼科耳配置， :，因爲塗布厚 ’所以在使第1 配置之橫構造 均等之缺陷在 尼科耳配置之 光板之偏極光 向25a相對於 1和第2偏極 0爲0°時成爲 E 0爲45°時成 「0.80」。當使 ,處透射光量比 第1偏極光板 丨，，透射光量 ‘之相位差透射 _ 0 ’「□」是 是角度0爲45 26兩片成爲正 -11- 200946899 交尼科耳配置之狀態下，第1偏極光板之偏極光方向25a 相對於膜16之遲相軸之角度Θ成爲45°，可以從正常部份 之透射光量和成爲塗布厚度不均之缺陷部份之透射光量之 差，檢測塗布厚度不均（相位差缺陷）。 因此，在本實施形態中，如第3圖所示，第1偏極光 板之偏極光方向25 a相對於膜16之遲相軸之角度0設定成 爲45°，第2偏極光板之偏極光方向26a相對於膜16之遲 相軸之角度（90° — 0 )成爲45°。另外，亦可以使膜之遲相軸 φ 方向與膜搬運方向X相同。 第6圖表示正交尼科耳配置之一對之偏極光板之光透 過特性。當在膜上沒有缺陷時，從第1和第2偏極光板25, 26只發出波長不滿7 00nm之光，但是由於該等之偏極光板 之特性，因此，波長超過700nm之光會以高透射率洩漏。 除去光學系27由紅外線切斷濾光器構成，配置在受光器24 之前。該紅外線切斷濾光器將斜線所示區域35內之光（紅 外線）切斷。因此，除去從第2偏極光板26發出之光中波 ©長在700nm以上之光（紅外線），只有波長不滿700nm之光 射入到受光器24。 受光器24由CCD攝像機構成，設置在膜16之搬運路 徑之上方。CCD攝像機具有攝像裝置，配置多個受光元件， 使其在膜16之寬度方向成爲線狀，在膜16之寬度方向進 行掃描，產生作爲輸出信號之掃描信號（攝像信號）。該輸 出信號如第7(A)圖所示，包含由於膜16上之缺陷所產生 之缺陷信號，和從第1和第2偏極光板25，26稍微洩漏之 雜訊信號。另外，受光器不只一台，較佳係設置多台》另 外，攝像裝置亦可以使用二次元配置受光元件之影像區域 -12- 200946899 感測器 受光器24 —般亦對波長超過700nm之光進行光電變 換，但是利用除去光學系27除去700nm以上之光。由於僅 依照膜16上之缺陷所產生之光光電變換，所以可將雜訊信 號抑制至最小限度，可獲得很大的S/ N(缺陷信號/雜訊 信號）。例如，在檢測到由於塗布厚度不均產生之塗布條紋 情況下，於未利用除去光學系27除去700nm以上之光時， S/N爲「1.5」，相對於此，在除去700nm以上之光時，S β / N提高爲「2.1」。 如第3圖所示，控制器28具備：缺陷信號檢測部32， 從來自受光器24之輸出信號（攝像信號）中檢測缺陷信號： 和缺陷位置指定部33,根據缺陷信號和來自編碼器30之編 碼脈波信號，指定膜16之長度方向之缺陷之位置。缺陷信 號檢測部32由微分處理部32a，極値信號檢測部32b，亮 度値演算部32c，和缺陷信號指定部32d構成。 微分處理部32a對第7(A)圖所示之輸出信號40施以微 分處理。利用此種方式，如第7(B)圖所示，獲得將亮度値 變化大之部份強調之信號4 2 (以下稱爲「微分處理信號」）。 其次，極値信號檢測部32b，如第8圖所示，從微分處理信 號42之中，檢測在一定像素範圍內之亮度値成爲最大之極 大信號5 0a，51a，52a，和在一定像素範圍內之亮度値成爲 最小之極小信號50b，51b，52b。 亮度値演算部3 2c求得預先設定之基準信號55之亮度 値和極大信號50a，51a，52a之亮度値之差分LAI，LB1， LC1 (以下稱爲「第1差分値」），和求得基準信號55之亮 度値和極小信號5 0b，51b，5 2b之亮度値之差分LA2, LB2， -13- 200946899 LC2(以下稱爲「第2差分値」）。在此處，基準信號55是 使雜訊信號平均化者’每次進行缺陷檢測時依序更新。然 後，亮度値演算部32c求得使第1差分値LAI，LB1，LC1 和第2差分値LA2，LB2 ’ LC2相加後之加算値LA，LB， LC ° 缺陷信號指定部32d判定亮度値演算部32c所求得之 加算値LA，LB，LC是否在一定値以上。判定之結果於加 算値LA，LB，LC在一定値以上情況下，指定極大信號50a， 赢 5 1a，52a和極小信號50b，51b，52b爲缺陷信號。在第8 圖中，因爲加算値LA，LB，LC全部在一定値以上，所以 將檢測到之信號全部指定爲缺陷信號。 在先前技術中，如第7(B)圖所示，對微分處理信號42 設定一定之臨限値Th，由於超過該臨限値Th之信號成爲 缺陷信號，因此，於亮度變化大之信號爲臨限値Th以下之 信號（第7(B)圖中之右側之信號）情況下，不檢測爲缺陷信 號。相對於此，在本發明中因爲以亮度値之變化量代替臨 ©限値Th來指定缺陷信號，所以可確實檢測膜1 6上之缺陷。 ‘ 其次說明本發明之缺陷檢測裝置之作用。經由定向膜 形成裝置11和液晶層形成裝置12，將膜16送入到缺陷檢 測裝置13。投光器22對在第1偏極光板25和第2偏極光 板26之間行走之膜16投光。由於投光器22使用鹵素燈泡 而只有一光亮線，因此，在膜16上不會產生使雜訊變大之 原因之一之干涉條紋。 來自投光器22之光利用第1偏極光板25偏光在偏極 光方向25a。在膜16上有異物混入等之定向缺陷情況下， 來自第1偏極光板25之光因在缺陷之散射、彌散變化其偏 -14 - 200946899 極光方向，所以從第2偏極光板26發出光。另外，有塗布 不均等之相位差缺陷之情況亦相同。由於第1和第2偏極 光板25，26之特性，波長700nm以上之光直接透過第1和 第2偏極光板25，26。 從第2偏極光板26發出之光利用除去光學系27除去 波長700nm以上之光，射入到受光器24 »受光器24將射 入之光變換成爲電信號，依序——讀出每一像素。該受光 器24之輸出信號40發送到控制器28。 控制器28內之微分處理部3 2a對來自受光器24之輸 出信號40施加微分處理。從施加過微分處理之微分處理信 號42，利用極値信號檢測部檢測極大信號50a，51a，52a 和極小信號50b，51b，52b。然後，利用亮度値演算部32c 求得預先設定之基準信號55之亮度値和極大信號50a， 51a，52a之亮度値之差分作爲第1差分値LA卜LB卜LC1， 和求得基準信號55之亮度値和極小信號50b，51b，52b之 亮度値之差分作爲第2差分値LA2，LB2，LC2。然後，利 用亮度値演算部32c求得使第1差分値LAI，LB1，LC1和 第2差分値LA2，LB2，LC2相加後之加算値LA，LB，LC。 在求得之加算値LA，LB，LC爲一定之値以上之情況時， 缺陷信號指定部32d指定利用極値信號檢測部32b檢測到 之極大信號50a，51a，52a和極小信號5 0b，51b，5 2b成爲 缺陷信號。然後，缺陷位置指定部33根據缺陷信號和編碼 脈波信號，指定膜16上之缺陷位置。控制器28之控制結 果顯示在顯示器（未圖示）。 另外，在上述實施形態中，除去光學系27是使用紅外 線切斷濾光器，但是亦可以使用其他使用有介電質多層膜 -15-W Considering that the first and second polarizing plates are iodine-based polarizing plates, but other dye-based polarizing plates, metal polarizing plates, or polarized plates made of calcite may be used. The first and second polarizing plates 25, 26 are arranged in crossed Nicols, and in the case of the film 16 without defects, the light emitted from the light projector 22 is polarized by the first polarizing plate 25 to the direction of the polarized light 25a. Thereafter, the film 16 is directly transmitted through the second polarizing plate 26 of the other one. Therefore, light passing through the normal portion hardly enters the light receiver 24. On the other hand, in the case of the film 16 having the orientation defect such as foreign matter mixed in, scratching, etc., the light emitted from the light projector 22 - 10 Ο Ο 200946899 is scattered and dispersed by the polarization of the first polarizing plate 25 to the polarization direction 25a. . The scattered and diffused light enters the film as a result of the polarized light passing through the other second polarizing plate 26, and the orientation defect such as foreign matter is mixed to produce the first and second polarizing plates 25 and 26 The positive 5 can be detected by the photodetector 24 that transmits only the scattered light. However, the unevenness (phase difference defect) does not generate scattered light and the second polarizing plates 25 and 26 are crossed, and the defect cannot be detected. However, the thickness of the coating is not orthogonal to the first and second polarizing plates 25, 26, and can be detected by intersecting the retardation axis of the film 16 with the first polarization direction 25a. Fig. 4 is a view showing the ratio of the transmitted light amount of the light transmitted through the two sheets of the second light plates 25 and 26 when the angle 0 of the slow axis of the polarizing film 16 of the first polarizing plate is changed. In the angle extinction mode, the transmitted light amount ratio (%) is "0.00". The angle g is the blue mode. When the blue mode transmitted light amount ratio (%) is about 45 degrees from the angle 0, the amount of transmitted light becomes maximum. Here, (%) indicates the ratio of the light emitted from the second polarizing plate 26 to the incident light of 25. Further, as shown in Fig. 5, even if the angle 0 is the same as the ratio (%), the defective portion of the film 16 changes with respect to the normal portion of light. In Fig. 5, "〇" is the angle 0j, the angle 0 is 15°, and "Z\j is the angle 0 is 30°, and "〇" °. The larger the phase difference, the larger the amount of transmitted light becomes. As described above, the first and second polarizing plates 25 are changed, and then the direction of the film 16 is changed. Raw scattered light, Nichol configuration, : Because the coating thickness is 'so that the horizontal configuration of the first configuration is equal, the polarization of the light plate of the Nicols is 25a relative to 1 and the second polarization 0 When it is 0°, it becomes "0.80" when E 0 is 45°. When the amount of transmitted light is higher than that of the first polarizing plate 丨, the phase difference transmitted by the transmitted light amount ′ 0′′ “□” is that the angle 0 is 45 26 and the two pieces become the positive -11-200946899 state of the Nikoel configuration. Then, the angle Θ of the polarization direction 25a of the first partial polarizer with respect to the slow axis of the film 16 is 45°, and the difference between the amount of transmitted light from the normal portion and the amount of transmitted light which becomes a defective portion of the coating thickness unevenness , detecting coating thickness unevenness (phase difference defect). Therefore, in the present embodiment, as shown in Fig. 3, the polarization direction 25a of the first polarizing plate is set to 45° with respect to the angle 0 of the slow axis of the film 16, and the polarized light of the second polarizing plate is polarized. The angle (90° - 0) of the direction 26a with respect to the slow axis of the film 16 becomes 45°. Further, the retardation axis φ direction of the film may be made the same as the film transport direction X. Figure 6 shows the light transmission characteristics of a polarized plate of one of the crossed Nicols configurations. When there is no defect on the film, only light having a wavelength of less than 700 nm is emitted from the first and second polarizing plates 25, 26, but due to the characteristics of the polarizing plates, light having a wavelength exceeding 700 nm is high. Transmittance leaks. The optical system 27 is composed of an infrared cut filter and placed before the light receiver 24. The infrared cut filter cuts light (infrared rays) in the region 35 indicated by oblique lines. Therefore, the light (infrared light) having a wavelength of 700 nm or more is emitted from the light emitted from the second polarizing plate 26, and only light having a wavelength of less than 700 nm is incident on the light receiver 24. The photodetector 24 is constituted by a CCD camera and is disposed above the transport path of the film 16. The CCD camera has an image pickup device, and a plurality of light-receiving elements are arranged in a line shape in the width direction of the film 16, and are scanned in the width direction of the film 16, and a scanning signal (image pickup signal) as an output signal is generated. As shown in Fig. 7(A), the output signal includes a defect signal due to a defect on the film 16, and a noise signal slightly leaking from the first and second polarizing plates 25, 26. In addition, there are more than one light receiver, preferably multiple sets. In addition, the image pickup device can also use the secondary element to configure the image area of the light receiving element -12-200946899. The sensor light receiver 24 generally also performs light with a wavelength exceeding 700 nm. Photoelectric conversion, but removing light of 700 nm or more by the removal optical system 27. Since the photo-electrical conversion is performed only in accordance with the defect on the film 16, the noise signal can be suppressed to a minimum, and a large S/N (defect signal/noise signal) can be obtained. For example, when a coating stripe due to uneven coating thickness is detected, S/N is "1.5" when light of 700 nm or more is not removed by the removal optical system 27, whereas when 700 nm or more is removed, light is removed. , S β / N is increased to "2.1". As shown in FIG. 3, the controller 28 includes a defect signal detecting unit 32 that detects a defect signal from an output signal (image pickup signal) from the photodetector 24: and a defect position specifying portion 33 based on the defect signal and from the encoder 30. The encoded pulse wave signal specifies the position of the defect in the longitudinal direction of the film 16. The defect signal detecting unit 32 includes a differential processing unit 32a, a pole signal detecting unit 32b, a luminance 値 calculating unit 32c, and a defect signal specifying unit 32d. The differential processing unit 32a applies a differential process to the output signal 40 shown in Fig. 7(A). In this manner, as shown in Fig. 7(B), a signal 4 2 (hereinafter referred to as "differential processing signal") which emphasizes a part where the luminance 値 is largely changed is obtained. Next, as shown in Fig. 8, the pole signal detecting unit 32b detects the maximum luminance signals 50a, 51a, 52a which are the maximum in the range of a certain pixel from the differential processing signal 42, and in a certain pixel range. The brightness 内 within the 値 becomes the smallest minimum signal 50b, 51b, 52b. The luminance 値 calculating unit 3 2c obtains the luminance 値 of the reference signal 55 set in advance and the difference LAI, LB1, LC1 (hereinafter referred to as "first differential 値") of the luminance 値 of the maximum signals 50a, 51a, 52a, and obtains The luminance 値 of the reference signal 55 and the luminance 値 of the minimum signal 50b, 51b, and 52b are LA2, LB2, -13-200946899 LC2 (hereinafter referred to as "second difference 値"). Here, the reference signal 55 is an average of the noise signals, and is updated sequentially each time the defect detection is performed. Then, the luminance 値 calculation unit 32c obtains the addition 値LA, LB, LC ° defect signal specifying unit 32d that adds the first difference 値LAI, LB1, LC1 and the second difference 値LA2, LB2 'LC2, and determines the luminance 値 calculation Whether the addition 値LA, LB, and LC obtained by the unit 32c is equal to or greater than a certain value. As a result of the determination, in the case where the addition 値LA, LB, LC is equal to or greater than a certain value, the maximum signal 50a is designated, and the wins 5 1a, 52a and the minimum signals 50b, 51b, 52b are defect signals. In Fig. 8, since the addition 値LA, LB, and LC are all above a certain threshold, all the detected signals are designated as defect signals. In the prior art, as shown in Fig. 7(B), a certain threshold 値Th is set for the differential processing signal 42, and since the signal exceeding the threshold 値Th becomes a defect signal, the signal having a large change in luminance is In the case of a signal below the threshold (the signal on the right side in the 7th (B) diagram), it is not detected as a defect signal. On the other hand, in the present invention, since the defect signal is specified by the amount of change in the luminance 値 instead of the limit 値Th, the defect on the film 16 can be surely detected. ‘ Next, the function of the defect detecting device of the present invention will be explained. The film 16 is fed to the defect detecting device 13 via the alignment film forming device 11 and the liquid crystal layer forming device 12. The light projector 22 projects the film 16 that travels between the first polarizing plate 25 and the second polarizing plate 26. Since the light projector 22 uses a halogen bulb and has only one bright line, interference fringes which are one of the causes of the noise increase are not generated on the film 16. The light from the light projector 22 is polarized in the polarization direction 25a by the first polarizing plate 25. When there is an orientation defect such as foreign matter in the film 16, the light from the first polarizing plate 25 is deflected by the scattering and dispersion of the defect, and the light is emitted from the second polarizing plate 26. . In addition, the case of the phase difference defect in which the coating is uneven is also the same. Due to the characteristics of the first and second polarizing plates 25, 26, light having a wavelength of 700 nm or more is directly transmitted through the first and second polarizing plates 25, 26. The light emitted from the second polarizing plate 26 is removed by the optical system 27 to remove light having a wavelength of 700 nm or more, and is incident on the light receiver 24. The light received by the light receiver 24 is converted into an electrical signal, and sequentially read out each. Pixel. The output signal 40 of the photoreceiver 24 is sent to the controller 28. The differentiation processing unit 32a in the controller 28 applies a differentiation process to the output signal 40 from the photodetector 24. From the differential processing signal 42 to which the differential processing is applied, the maximum signal 50a, 51a, 52a and the minimum signals 50b, 51b, 52b are detected by the pole signal detecting portion. Then, the brightness 値 calculating unit 32c obtains the difference between the luminance 値 of the reference signal 55 set in advance and the luminance 値 of the maximum signals 50a, 51a, 52a as the first difference 値LA LB, LC1, and the reference signal 55. The difference between the luminance 値 and the luminance 値 of the minimum signals 50b, 51b, 52b is the second difference 値LA2, LB2, LC2. Then, the luminance 値 calculation unit 32c obtains the addition 値LA, LB, LC obtained by adding the first difference 値LAI, LB1, LC1 and the second difference 値LA2, LB2, and LC2. When the calculated addition 値LA, LB, and LC are equal to or greater than a certain value, the defect signal specifying unit 32d specifies the maximum signals 50a, 51a, 52a and the minimum signals 50b, 51b detected by the pole signal detecting unit 32b. , 5 2b becomes a defect signal. Then, the defect position specifying portion 33 specifies the position of the defect on the film 16 based on the defect signal and the encoded pulse wave signal. The control result of the controller 28 is displayed on a display (not shown). Further, in the above embodiment, the optical line 27 is removed using an infrared ray cut filter, but other dielectric multilayer films may be used.
❹ 200946899 之帶通濾光器，單色儀，波長切斷濾光器，色玻璃濾 繞射格子等。另外，以除去光學系27除去之光之波 不一定要限定在700nm以上，亦可以依照缺陷檢測 之偏極光板之種類適當地變更。 其次，利用實施例1和比較例1，2來具體地說 明。 [實施例1] 使用第3圖所示之缺陷檢測裝置13,進行膜16上 陷檢測。在檢査對象之膜16之下方設置鹵素燈泡之投 22，在其上方設置受光器24。在投光器22和膜16之 置第1偏極光板25，在膜16和受光器24之間設置第 極光板26。另外，使第1偏極光板25之偏極光方|ΐ 相對於膜16之遲相軸成爲45°，第2偏極光板26之偏 方向26a相對於X方向成爲45°。第1和第2偏極光相 26使用碘素系偏極光板，受光器24使用CCD線攝像 在受光器24之前設置由紅外線切斷濾光器構成 去光學系27。利用除去光學系27除去來自第2偏桓 26之光中波長700nm以上之光。受光器24檢測除5 系27所發出之光，將檢測到之信號發信到控制器28 控制器28藉由對利用受光器24獲得之信號施以搜 理，獲得微分處理信號42。從微分處理信號42，利月 信號檢測部32b檢測在一定像素範圍內成爲最大之卷 號50a，51a，52a，和成爲最小之極小信號50b，51b， 然後，求得預先設定之基準信號55之亮度値和極 50a，51a，52a之亮度値之差分作爲第1差分値LA1， LC1，和求得基準信號55之亮度値和極小信號50b， 器， 區域 使用 本發 之缺 光器 間設 2偏 I 25a 極光 :25， 機。 之除 :光板 :光學 〇 :分處 I極値 i大信 52b ° :信號 LB1， 51b， -16- 200946899 5 2b之亮度値之差分作爲第2差分値LA2，LB2’ LC2’然 後，求得使第1差分値LAI，LB1，LC1和第2差分値LA2， LB2，LC2相加後之加算値LA，LB，LC。在求得之加算値 LA，LB，LC爲一定値以上情況下，指定利用極値信號檢測 部32b所檢測到之極大信號50a，51a，52a和極小信號50b， 51b，52b成爲缺陷信號。 [比較例1 ] 代替鹵素燈泡，使用從金屬鹵化物燈泡發出之光，除 p 此之外進行與實施例1相同之缺陷檢測。 [比較例2] 代替鹵素燈泡者，使用從螢光燈發出之光，除此之外 進行與實施例1相同之缺陷檢測。 [評估] 對上述實施例1和比較例1，2所進行之缺陷檢查，評 估在膜上產生何種程度之干涉條紋，S/N爲何種程度，綜 合評估各個實施例和比較例。另外，使用習知之分光測定 _ 器，測定各個實施例和比較例之投光器之分光放射強度， 從其測定結果檢査有多少亮線。 第9圖顯示各個實施例和比較例所測定到之分光放射 強度，圖形60表示鹵素燈泡之分光放射強度，圖形61表 示金屬鹵化物燈泡之分光放射強度，圖形62表示螢光燈之 分光放射強度， 第10圖顯示實施例和比較例之評估結果。第10圖之 投光器表示各個實施例和比較例所使用之投光器之種類， 亮線表示從投光器發出之光之亮線之數目。另外，干涉條 紋之「0」表示在膜上未發生有干涉條紋，「X」表示在膜 -17- 200946899 上發生有干涉條紋。另外，「S/Ν」之「0」表示S/Ν爲2 以上，「△」表示S/N爲1.5以上。另外，評估之「0」表 示檢測到製品上成爲問題之缺陷，「X」表示不能確實地檢 測製品上成爲問題之缺陷。另外，對於S/Ν假如爲「2」 以上時可以穩定地進行缺陷檢測。 在實施形態1中只有一條從鹵素燈泡發出之光之亮 線，所以不會在膜上發生干涉條紋。相對於此，在比較例 1和2中，因爲有兩條以上之亮線，所以在膜上產生亮部 0 份和暗部份明確之干涉條紋。由於該干涉條紋之影響使S / N降低，不能確實地進行製品上成爲問題之缺陷之檢測。 【圖式簡單說明】 第1圖是相位差膜製造線之槪略圖。 第2圖之圖形表示鹵素燈泡之分光放射強度。 第3圖是缺陷檢測裝置之立體圖。 第4圖之圖形表示第1偏極光板之偏極光方向對膜之 遲相軸之角度0，和透射光量比之關係》 © 第5圖之圖形表示膜之相位差和透射光量比之關係。 第6圖之圖形表示配置成爲正交尼科耳之一對偏極光 板之分光透射率。 第7(A)圖之圖形表示輸p信號，第7(B)圖之圖形表示 對輸出信號施加微分處理之微分處理信號。 第8圖是說明圖，用來說明本發明之缺陷檢測方法。 第9圖之圖形表示鹵素燈泡，金屬鹵化物燈泡，螢光 燈之分光放射強度。 第10圖表示實施例1和比較例1，2之結果。 -18- 200946899 【主要元件符號說明】❹ 200946899 bandpass filter, monochromator, wavelength cut filter, color glass filter diffraction grating, etc. Further, the wave of the light removed by the removal of the optical system 27 is not necessarily limited to 700 nm or more, and may be appropriately changed in accordance with the type of the polarizing plate for detecting the defect. Next, the first embodiment and the comparative examples 1, 2 will be specifically described. [Embodiment 1] The film 16 is detected by the defect detecting device 13 shown in Fig. 3. A halogen bulb projection 22 is disposed below the film 16 of the inspection object, and a photodetector 24 is disposed above it. The first polarizer 26 is disposed between the film 16 and the light receiver 24 in the first polarizer 25 of the light projector 22 and the film 16. Further, the polarization axis | 第 of the first polarization plate 25 is set to 45° with respect to the slow phase axis of the film 16, and the deflection direction 26a of the second polarization plate 26 is 45° with respect to the X direction. The first and second polarized light phases 26 are made of an iodine-based polarizing plate, and the photodetector 24 is imaged by a CCD line. Before the light receiver 24, an optical cut-off filter is provided by an infrared cut filter. Light having a wavelength of 700 nm or more among the light from the second bias 26 is removed by the removal optical system 27. The light receiver 24 detects the light emitted by the 5 series 27, and transmits the detected signal to the controller 28. The controller 28 obtains the differential processing signal 42 by applying a signal to the signal obtained by the light receiver 24. From the differential processing signal 42, the lunar signal detecting unit 32b detects the volume numbers 50a, 51a, 52a which are the largest in a certain pixel range, and the minimum mini signals 50b, 51b, and then obtains the preset reference signal 55. The difference between the brightness 値 and the brightness 极 of the poles 50a, 51a, 52a is taken as the first difference 値LA1, LC1, and the brightness 値 and the minimum signal 50b of the reference signal 55 are obtained, and the area is set between the apertures of the present invention. Partial I 25a Aurora: 25, machine. Divide: Light board: Optical 〇: Division I 値 i big letter 52b °: Signal LB1, 51b, -16- 200946899 5 2b brightness 値 difference as the second difference 値LA2, LB2' LC2' then, get The first difference 値LAI, LB1, LC1 and the second difference 値LA2, LB2, and LC2 are added and added 値LA, LB, LC. When the obtained addition 値 LA, LB, LC is equal to or greater than 値, the maximum signals 50a, 51a, 52a and the minimum signals 50b, 51b, 52b detected by the pole signal detecting unit 32b are designated as defective signals. [Comparative Example 1] The same defect detection as in Example 1 was carried out except that the light emitted from the metal halide bulb was used instead of the halogen bulb. [Comparative Example 2] The same defect detection as in Example 1 was carried out except that the halogen lamp was used instead of the light emitted from the fluorescent lamp. [Evaluation] With respect to the defect inspection performed in the above Example 1 and Comparative Examples 1, 2, it was evaluated to what extent the interference fringes were generated on the film, and to what extent S/N, and various examples and comparative examples were comprehensively evaluated. Further, the spectral emission intensity of the light projectors of the respective examples and comparative examples was measured using a conventional spectrophotometer, and the number of bright lines was examined from the measurement results. Fig. 9 shows the spectral emission intensity measured by each of the examples and the comparative examples, the figure 60 shows the spectral emission intensity of the halogen bulb, the figure 61 shows the spectral emission intensity of the metal halide bulb, and the figure 62 shows the spectral emission intensity of the fluorescent lamp. Fig. 10 shows the evaluation results of the examples and comparative examples. The light projector of Fig. 10 indicates the kind of the light projector used in each of the examples and the comparative examples, and the bright line indicates the number of bright lines of light emitted from the light projector. Further, "0" of the interference pattern indicates that no interference fringes occurred on the film, and "X" indicates that interference fringes occurred on the film -17-200946899. In addition, "0" of "S/Ν" indicates that S/Ν is 2 or more, and "△" indicates that S/N is 1.5 or more. In addition, the evaluation of "0" indicates that the product has been found to be a problem, and "X" indicates that the defect in the product cannot be reliably detected. In addition, when the S/Ν is "2" or more, the defect detection can be performed stably. In the first embodiment, there is only one bright line of light emitted from the halogen bulb, so that interference fringes do not occur on the film. On the other hand, in Comparative Examples 1 and 2, since there are two or more bright lines, a bright portion of the film and a clear interference fringe of the dark portion are formed on the film. Since the influence of the interference fringes lowers S / N, the detection of defects which are a problem on the product cannot be reliably performed. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a schematic diagram of a phase difference film manufacturing line. The graph of Fig. 2 shows the spectral emission intensity of the halogen bulb. Figure 3 is a perspective view of the defect detecting device. The graph of Fig. 4 shows the relationship between the angle of the polarization direction of the first polarizing plate to the slow axis of the film and the ratio of the transmitted light amount. © The graph of Fig. 5 shows the relationship between the phase difference of the film and the ratio of the transmitted light. The graph of Fig. 6 shows the spectral transmittance of the polarized plate which is configured as one of the crossed Nicols. The graph of Fig. 7(A) shows the p signal, and the graph of Fig. 7(B) shows the differential processing signal for the differential processing of the output signal. Fig. 8 is an explanatory view for explaining the defect detecting method of the present invention. The graph in Fig. 9 shows the spectral emission intensity of a halogen bulb, a metal halide bulb, and a fluorescent lamp. Fig. 10 shows the results of Example 1 and Comparative Examples 1, 2. -18- 200946899 [Description of main component symbols]
10 相位差膜製造線 11 定向膜形成裝置 12 液晶層形成裝置 13 缺陷檢測裝置 14 捲取裝置 15 透明樹脂膜 16 膜 18 膜捲筒 20，21 導引滾子 22 投光器 23 光量調整部 24 受光器 25 第1偏極光板 25a 偏極光方向 26 第2偏極光板 26a 偏極光方向 27 除去光學系 28 控制器 30 編碼器 32 缺陷信號檢測部 32a 微分處理部 32b 極値信號檢測部 32c 亮度値演算測部 32d 缺陷信號指定部 33 缺陷位置指定部 -19- 200946899 40 42 5 0 a > 51a，5 2a 50b ， 51b， 52b 55 LA > LB > LC LAI，LB1，LC1 LA2，LB2，LC2 輸出信號 微分處理信號 極大信號 極小信號 基準信號 加算値 第1差分値 第2差分値10 retardation film manufacturing line 11 Orientation film forming device 12 Liquid crystal layer forming device 13 Defect detecting device 14 Winding device 15 Transparent resin film 16 Film 18 Film reel 20, 21 Guide roller 22 Emitter 23 Light amount adjusting portion 24 Photoreceiver 25 First polarizing plate 25a Polarized light direction 26 Second polarizing plate 26a Polarized light direction 27 Removal of optical system 28 Controller 30 Encoder 32 Defect signal detecting unit 32a Differential processing unit 32b Extreme signal detecting unit 32c Brightness measurement Part 32d defect signal specifying unit 33 defect position specifying unit-19-200946899 40 42 5 0 a > 51a, 5 2a 50b , 51b, 52b 55 LA > LB > LC LAI, LB1, LC1 LA2, LB2, LC2 output Signal differential processing signal maximal signal minimum signal reference signal addition 値 first difference 値 second difference 値
-20--20-