This chapter explains how machine vision extracts target objects from complex images, a task humans perform instinctively using colour, shape, and accumulated perceptual knowledge. Since computers lack such innate understanding, machine vision must encode this knowledge through algorithms. The chapter introduces threshold-based segmentation, including histogram analysis and the Otsu method; colour-based extraction using RGB and HSI transformations; and motion-based extraction through inter-frame differencing and background subtraction. These techniques allow machines to isolate meaningful objects—such as fruits, crops, vehicles, or people—under diverse environmental and lighting conditions, forming the foundation for automated detection and measurement.