As we discussed before, camera resolution has a significant impact on image quality. So, how does the resolution of industrial lenses affect images? And how should we choose a lens that is suitable for our camera?
Currently, megapixel high-definition cameras are increasingly used in various industrial inspection fields, including 1-megapixel, 2-megapixel, 5-megapixel, 14-megapixel, and even 29-megapixel cameras. High pixel count is one aspect of image clarity; however, high-pixel sensors also require high-resolution lenses to fully realize the capabilities of a high-definition camera.
Let's follow Microvision and explore this issue. First, let's look at the images obtained by using lenses of different resolutions with high-definition cameras of the same resolution, under exactly the same shooting conditions:
The two images above were taken using the same high-resolution CCD camera developed and manufactured by Vision Image Technology Co., Ltd. The camera settings and external shooting conditions (such as lighting conditions and field of view) were completely identical; the only difference was the lens resolution. The image above used a 1-megapixel lens, while the image below used a 5-megapixel lens.
We cropped the same field of view for both images. The left side of the image corresponds to the left edge of the entire image (the imaging effect of the right edge of the entire image is the same as the left side, so we won't repeat it). The right side of the image corresponds to the center image of the entire image. From the comparison of the two images, we can see that the images in the central area, that is, the images of areas A and A', are not significantly different and can both be clearly imaged. However, the image quality of the images at the left edge, that is, the images of areas B and B', is very different. In the image above, the image of area B is very blurry, as if it is not in focus. However, area A clearly shows that the lens is in perfect focus. In the image below, the images of areas B' and A' are almost identical and are both very clear. From this, we can see that the difference in lens resolution has a fatal impact on the sharpness of the image, especially the sharpness of the image edges. Lenses that are not well matched with high-definition cameras have a very large resolution difference between the center field of view and the edge field of view. Although the image in the center field of view can be barely clear, the blurry image in the edge field of view will definitely affect the image quality of the entire image, which will bring unpredictable troubles to post-processing.
So, how should you choose the right lens for a high-definition camera?
We know that a lens has a maximum resolution of Nlp/mm. According to the Nalquist sampling theorem, it needs at least 2N/mm of spatial sampling points. This can be understood as follows: if there are N pairs of black and white lines within 1mm, then there are N white lines and N black lines, totaling 2N lines. Assuming one photosensitive element in a camera corresponds to one white or black line, then the camera needs 2N photosensitive elements within 1mm to correspond to the N white lines and N black lines. The camera's photosensitive element density is 2N/mm. At this point, the resolution of the camera's photosensitive element and the resolution of the lens are perfectly matched, with no wasted resources. Similarly, if a camera has a pixel density of M points per millimeter (pixels/mm), then a lens with a resolution of M/2lp/mm should be chosen.
Let's take an example: A 2-megapixel camera has a pixel count of 1600 × 1200 = 1920000 and a sensor size of 1/2 inch. We know that a 1/2-inch sensor has a horizontal dimension of 6.4mm and a vertical dimension of 4.8mm . Its horizontal pixel density is 1600/6.4 = 250 pixels/mm, and its vertical pixel density is 1200 / 4.8 = 250 pixels/mm. Here, both the horizontal and vertical pixel densities are 250 pixels/mm (if the horizontal and vertical pixel densities are the same, the pixels are square; if the pixels are not square, the lens resolution should refer to the one with the higher pixel density). Therefore, the lens resolution should be selected as 125 lp/mm. If a 2-megapixel camera has a 1/3-inch sensor, its horizontal dimensions are 4.8mm and its vertical dimensions are 3.6mm . Its horizontal pixel density is 1600 / 4.8 = 333.3 pixels/mm, and its vertical pixel density is 1200 / 3.6 = 333.3 pixels/mm. Therefore, the lens resolution should be 167 lp/mm.
The example above also shows that a 2-megapixel lens labeled as 1/2 inch is not suitable for a 2-megapixel sensor labeled as 1/3 inch. This distinction is important to understand. A 2-megapixel lens with a 1/2 inch sensor has a resolution of 125 lp/mm . Using it on a 1/3 inch sensor (4.8mm x 3.6mm ) , there are 125 x 4.8 = 600 line pairs horizontally, corresponding to 1200 pixels, and 125 x 3.6 = 450 line pairs vertically, corresponding to 900 pixels. 1200 x 900 = 1080000, which is approximately 1.1 million pixels. Therefore, using a 2-megapixel lens labeled as 1/2 inch on a 1/3 inch sensor is only suitable for a 1-megapixel sensor, or in other words, it can only be used as a 1-megapixel lens. Therefore, simply talking about a megapixel lens without mentioning the size of the applicable image sensor may not provide complete information. Thus, lp/mm is a more accurate description of lens resolution.
Sometimes, camera specifications will state the sensor size L (mm). The corresponding lens resolution's black and white line width should also be L (mm). Therefore, the width of a pair of black and white lines is 2L (mm), and its reciprocal, 1/(2L) (lp/mm), is the lens resolution. For example, if a camera specifies a sensor size of 4µm, then the maximum resolution line width (white or black line) of the selected lens should also be 4µm . The width of a pair of black and white lines is 2 × 0.004 mm, so the lens resolution is 1/(2 × 0.004 ) = 125 lp/mm.
