Picture categorization is a fundamental task in vision recognition that aims to understand and label an image in its entirety. While object detection works with the categorization and placement of many elements inside an image, image classification often pertains to photographs containing a single object. The development of sophisticated parallel computers in tandem with the introduction of contemporary remote sensors has fundamentally changed the picture categorization theory. Various algorithms have been created to recognise objects of interest in pictures and then categorise them and practise. In recent years, a number of authors have offered a range of classification strategies. However, there are not many studies or comparisons of classification techniques in soft computing settings. These days, the use of soft computing techniques has improved the performance of classification methods. This work explores the use of soft computing for image classification for various applications. The study explores further details regarding new applications and various classification technique types. To promote greater study in this field, important problems and viable fixes for applications based on soft computing are also covered. As a result, researchers will find this survey study useful in implementing an optimal categorization method for multiple applications.