Q2 You have been given a task to develop an automatic inspection system to inspect objects carried on a conveyor. The captured grayscale image is underexposed and overexposed in some of the area. (a) The initial task is to improve the image quality. Image comprises of high frequency component (HFC) and low frequency components (LFC). (1) Design a filter to remove the HFC while retaining the LFC. In the designed filter, please state clearly on how to differentiate between HFC and LFC in each local window. (4 marks) (i) Illustrate a histogram of a region in the image when it is underexposed and overexposed. (2 marks) (b) (iii) Develop an algorithm to improve the underexposed/overexposed area. State clearly each step taken including the utilised equation and how to detect such regions in an image. (4 marks) The second task is to segmentize the object from the image by using an adaptive threshold technique, (i) Explain how the adaptive threshold technique can be invariant under various lighting condition. (2 marks) (ii) Construct an algorithm for extracting the object region by using mean-C and empirical methods. Please state clearly on how the range of the empirical value is selected. (5 marks) (iii) The segmented image is contaminated with salt & pepper noise and some small holes inside the object. Formulate a procedure to diminish such problem without affecting the object dimension. Q2 You have been given a task to develop an automatic inspection system to inspect objects carried on a conveyor. The captured grayscale image is underexposed and overexposed in some of the area. (a) The initial task is to improve the image quality. Image comprises of high frequency component (HFC) and low frequency components (LFC). (1) Design a filter to remove the HFC while retaining the LFC. In the designed filter, please state clearly on how to differentiate between HFC and LFC in each local window. (4 marks) (i) Illustrate a histogram of a region in the image when it is underexposed and overexposed. (2 marks) (b) (iii) Develop an algorithm to improve the underexposed/overexposed area. State clearly each step taken including the utilised equation and how to detect such regions in an image. (4 marks) The second task is to segmentize the object from the image by using an adaptive threshold technique, (i) Explain how the adaptive threshold technique can be invariant under various lighting condition. (2 marks) (ii) Construct an algorithm for extracting the object region by using mean-C and empirical methods. Please state clearly on how the range of the empirical value is selected. (5 marks) (iii) The segmented image is contaminated with salt & pepper noise and some small holes inside the object. Formulate a procedure to diminish such problem without affecting the object dimension